More pdf : www.motivationbank.

in Design and Analysis of Algorithm (KCS503)

Course Outcome ( CO) Bloom’s Knowledge Level (KL)

At the end of course , the student will be able to:

Design new algorithms, prove them correct, and analyze their asymptotic and absolute runtime K4, K6
CO 1
and memory demands.
Find an algorithm to solve the problem (create) and prove that the algorithm solves the problem K5, K6
CO 2
correctly (validate).
Understand the mathematical criterion for deciding whether an algorithm is efficient, and know K2, K5
CO 3
many practically important problems that do not admit any efficient algorithms.
CO 4 Apply classical sorting, searching, optimization and graph algorithms. K2, K4
Understand basic techniques for designing algorithms, including the techniques of recursion, K2, K3
CO 5
divide-and-conquer, and greedy.
DETAILED SYLLABUS 3-1-0
Unit Topic Proposed
Lecture
Introduction: Algorithms, Analyzing Algorithms, Complexity of Algorithms, Growth of
I Functions, Performance Measurements, Sorting and Order Statistics - Shell Sort, Quick Sort, Merge 08
Sort, Heap Sort, Comparison of Sorting Algorithms, Sorting in Linear Time.
Advanced Data Structures: Red-Black Trees, B – Trees, Binomial Heaps, Fibonacci Heaps,
II 08
Tries, Skip List
Divide and Conquer with Examples Such as Sorting, Matrix Multiplication, Convex Hull and
Searching.
III Greedy Methods with Examples Such as Optimal Reliability Allocation, Knapsack, Minimum 08
Spanning Trees – Prim’s and Kruskal’s Algorithms, Single Source Shortest Paths - Dijkstra’s and
Bellman Ford Algorithms.
Dynamic Programming with Examples Such as Knapsack. All Pair Shortest Paths – Warshal’s
and Floyd’s Algorithms, Resource Allocation Problem.
IV 08
Backtracking, Branch and Bound with Examples Such as Travelling Salesman Problem, Graph
Coloring, n-Queen Problem, Hamiltonian Cycles and Sum of Subsets.
Selected Topics: Algebraic Computation, Fast Fourier Transform, String Matching, Theory of NP-
V 08
Completeness, Approximation Algorithms and Randomized Algorithms
Text books:
1. Thomas H. Coreman, Charles E. Leiserson and Ronald L. Rivest, “Introduction to Algorithms”, Printice Hall of
India.
2. E. Horowitz & S Sahni, "Fundamentals of Computer Algorithms",
3. Aho, Hopcraft, Ullman, “The Design and Analysis of Computer Algorithms” Pearson Education, 2008.
4. LEE "Design & Analysis of Algorithms (POD)",McGraw Hill
5. Richard E.Neapolitan "Foundations of Algorithms" Jones & Bartlett Learning
6. Jon Kleinberg and Éva Tardos, Algorithm Design, Pearson, 2005.
7. Michael T Goodrich and Roberto Tamassia, Algorithm Design: Foundations, Analysis, and Internet Examples,
Second Edition, Wiley, 2006.
8. Harry R. Lewis and Larry Denenberg, Data Structures and Their Algorithms, Harper Collins, 1997
9. Robert Sedgewick and Kevin Wayne, Algorithms, fourth edition, Addison Wesley, 2011.
10. Harsh Bhasin,”Algorithm Design and Analysis”,First Edition,Oxford University Press.
11. Gilles Brassard and Paul Bratley,Algorithmics:Theory and Practice,Prentice Hall,1995.

Curriculum & Evaluation Scheme CS & CSE (V & VI semester) 6

Design and Analysis of Algorithms 1–1 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in

1 Introduction

CONTENTS
Part-1 : Algorithms, Analyzing ............................... 1–2B to 1–3B
Algorithms, Complexity
of Algorithms

Part-2 : Growth of Functions, ............................... 1–3B to 1–13B

Performance
Measurements

Part-3 : Sorting and Order Statistic : ................ 1–14B to 1–20B

Shell Sort, Quick Sort

Part-4 : Merge Sort ............................................... 1–20B to 1–23B

Part-5 : Heap Sort .................................................. 1–23B to 1–30B

Part-6 : Comparison of Sorting ........................... 1–30B to 1–35B

Algorithms, Sorting in
Linear Time
Introduction 1–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
PART-1
Introduction : Algorithms, Analyzing Algorithms, Complexity of
Algorithms.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 1.1. What do you mean by algorithm ? Write the

characteristics of algorithm.

Answer
1. An algorithm is a set of rules for carrying out calculation either by hand
or on machine.
2. It is a finite step-by-step procedure to achieve a required result.
3. It is a sequence of computational steps that transform the input into the
output.
4. An algorithm is a sequence of operations performed on data that have to
be organized in data structures.
Characteristics of algorithm are :
1. Input and output : The algorithm must accept zero or more inputs
and must produce at least one output.
2. Definiteness : Each step of algorithm must be clear and unambiguous.
3. Effectiveness : Every step must be basic and essential.
4. Finiteness : Total number of steps used in algorithm should be finite.
Que 1.2. What do you mean by analysis or complexity of an
algorithm ? Give its types and cases.

Answer
Analysis/complexity of an algorithm :
The complexity of an algorithm is a function g(n) that gives the upper
bound of the number of operation (or running time) performed by an
algorithm when the input size is n.
Types of complexity :
1. Space complexity : The space complexity of an algorithm is the amount
of memory it needs to run to completion.
2. Time complexity : The time complexity of an algorithm is the amount
of time it needs to run to completion.
Cases of complexity :
1. Worst case complexity : The running time for any given size input
will be lower than the upper bound except possibly for some values of
the input where the maximum is reached.
Design and Analysis of Algorithms 1–3 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. Average case complexity : The running time for any given size input
will be the average number of operations over all problem instances
for a given size.
3. Best case complexity : The best case complexity of the algorithm is
the function defined by the minimum number of steps taken on any
instance of size n.

PART-2
Growth of Functions, Performance Measurements.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 1.3. What do you understand by asymptotic notations ?

Describe important types of asymptotic notations.
OR
Discuss asymptotic notations in brief.

Answer
1. Asymptotic notation is a shorthand way to represent the fastest possible
and slowest possible running times for an algorithm.
2. It is a line that stays within bounds.
3. These are also referred to as ‘best case’ and ‘worst case’ scenarios and
are used to find complexities of functions.
Notations used for analyzing complexity are :
1. -Notation (Same order) :
c2g(n)
f(n)
c1g(n)

n0
n
f(n) =  (g(n))
Fig. 1.3.1.
a. This notation bounds a function within constant factors.
b. We say f(n) = g(n) if there exist positive constants n0, c1 and c2
such that to the right of n0 the value of f(n) always lies between
c1 g(n) and c2 g(n) inclusive.
2. O-Notation (Upper bound) :
a. Big-oh is formal method of expressing the upper bound of an
algorithm’s running time.
Introduction 1–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. It is the measure of the longest amount of time it could possibly
take for the algorithm to complete.
c. More formally, for non-negative functions, f (n) and g(n), if there
exists an integer n0 and a constant c > 0 such that for all integers
n  n0.
f (n)  cg(n)
d. Then, f (n) is big-oh of g(n). This is denoted as :
f (n)  O(g(n))
i.e., the set of functions which, as n gets large, grow faster than a
constant time f (n).

cg(n)

f(n)

n
n0 f(n) = O(g(n))
Fig. 1.3.2.
3. -Notation (Lower bound) :
a. This notation gives a lower bound for a function within a constant
factor.
b. We write f(n) = g(n)) if there are positive constants n0 and c such
that to the right of n0, the value of f(n) always lies on or above cg(n).
f(n)

cg(n)

n
n0
f(n) =  (g(n))
Fig. 1.3.3.
4. Little-oh notation (o) : It is used to denote an upper bound that is
asymptotically tight because upper bound provided by O-notation is
not tight.
o(g(n)) = {f(n) : for any positive constant c > 0, if a constant n0 > 0 such
that 0 < f(n) < cg(n) V n > n0}
5. Little omega notation () : It is used to denote lower bound that is
asymptotically tight.
(g(n)) = {f(n) : For any positive constant c > 0, if a constant n0 > 0 such
that 0 < cg(n) < f(n) V n > n0}
Design and Analysis of Algorithms 1–5 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 1.4. If f(n) = 100 * 2n + n5 + n, then show that f(n) = O(2n).

Answer
If f(n) = 100 * 2n + n5 + n
For n5  n
100 * 2n + n5 + n  100 * 2n + n5 + n5
 100 * 2n + 2n5
For 2n  n5
100 * 2n + n5 + n  100 * 2n + 2.2n
 102 * 2n [ n  1, n0 = 23]
Thus, f(n) = O(2n)

Que 1.5. Write Master’s theorem and explain with suitable

examples.

Answer
Master’s theorem :
Let T(n) be defined on the non-negative integers by the recurrence.
 n
T(n) = aT   + f(n) where a  1, b > 1 are constants
 b
a = Number of sub-problems in the recursion
1/b = Portion of the original problem represented by each sub-
problem
f(n) = Cost of dividing the problem and the cost of merging the
solution
Then T(n) can be bounded asymptotically as follows :
Case 1 :
If it is true that : f(n) = O(nlogb a –E) for E> 0
It follows that : T(n) =  (nlogb a)
 n
Example : T(n) = 8T   + 1000n2
 b
In the given formula, the variables get the following values :
a = 8, b = 2, f(n) = 1000n2, logba = log28 = 3
nlogb a = nlog2 8 = n3
f(n) = O(nlogb a – E) = O(n3 – E)
For E = 1, we get
f(n) = O(n3 – 1) = O(n2)
Since this equation holds, the first case of the Master’s theorem applies to the
given recurrence relation, thus resulting solution is
T(n) =  (nlogb a) = (n3)
Case 2 :
If it is true that : f(n) =  (nlogb a)
It follows that : T(n) =  (nlogb a log(n))
Introduction 1–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Example :
 n
T(n) = 2T   + n
 2
In the given formula, the variables get the following values :
a = 2, b = 2, f(n) = n, logba = log22 = 1
nlogb a = nlog2 2 = n
f(n) = (nlogb a) = (n)
Since this equation holds, the second case of the Master’s theorem applies to
the given recurrence relation, thus resulting solution is :
T(n) = (nlogb a log(n)) = (n log n)
Case 3 :
If it is true that : f(n) = (nlogb a + E) for E> 0
and if it is also true that :
 n
if af    cf(n) for a, c < 1 and all sufficiently large n
 b
It follows that : T(n) = (f(n))
 n
Example : T(n) = 2T   + n2
2
In the given formula, the variables get the following values :
a = 2, b = 2, f(n) = n2, logb a = log2 2 = 1
nlogb a = nlog2 2 = n

f(n) = (nlogb a + E)
For E = 1 we get
f(n) = (n1 + 1) = (n2)
Since the equation holds, third case of Master’s theorem is applied.
Now, we have to check for the second condition of third case, if it is true
that :
 n
af    c f(n)
 b
If we insert once more the values, we get :
2
 n 1
2    cn2  n2  cn2
2 2
1
If we choose c = , it is true that :
2
1 2 1 2
n  n  n 1
2 2
So, it follows that : T(n) =  (f(n))
If we insert once more the necessary values, we get :
T(n)   (n2)
Thus, the given recurrence relation T(n) was in (n2).

Que 1.6. The recurrence T(n) = 7T(n/2) + n2 describe the running

time of an algorithm A. A competing algorithm A has a running
Design and Analysis of Algorithms 1–7 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
time T(n) = aT (n/4) + n2. What is the largest integer value for a A is
asymptotically faster than A ? AKTU 2017-18, Marks 10

Answer
Given that :
 n
T(n) = 7T    n2 ...(1.6.1)
 2
 n
T(n) = aT     n2 ...(1.6.2)
 4
Here, eq. (1.6.1) defines the running time for algorithm A and eq. (1.6.2)
defines the running time for algorithm A. Then for finding value of a for
which A is asymptotically faster than A we find asymptotic notation for the
recurrence by using Master’s method.
 n
Now, compare eq. (1.6.1) by T(n) = aT    f (n)
 b
we get, a= 7
b=2
f(n) = n2
log 7
n log b a = n 2 = n2.81
Now, apply cases of Master’s, theorem as :
Case 1 : f(n) = O (n log2 7  E )
 f(n) = O (n2.81 – E)
 f(n) = O (n2.81 – 0.81)
 f(n) = O (n2)
Hence, case 1 of Master’s theorem is satisfied.
Thus, T(n) =  ( nlog b a )
 T(n) =  (n2.81)
Since recurrence given by eq. (1.6.1) is asymptotically bounded by
-notation by which is used to show optimum time we have to show that
recurrence given by eq. (1.6.2) is bounded by -notation which shows
minimum time (best case).
For the use satisfy the case 3 of Master theorem, let a = 16
 n
T(n) = 16T     n2
 4
 a = 16
b=4
f(n) = n2
(nlog b a E ) = (n2 + E)
Hence, case 3 of Master’s theorem is satisfied.
 T(n) =  (f(n))
 T(n) =  (n2)
Therefore, this shows that A is asymptotically faster than A when
a = 16.
Introduction 1–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 1.7. The recurrence T(n) = 7T(n/3) + n2 describes the running
time of an algorithm A. Another competing algorithm B has a
running time of S(n) = a S(n/9) + n2. What is the smallest value of a
such that B is asymptotically faster than A ?
AKTU 2018-19, Marks 10

Answer
Given that :
 n
T(n) = 7T    n2 ...(1.7.1)
 3
 n 2
S(n) = aS    n ...(1.7.2)
9
Here, eq. (1.7.1) defines the running time for algorithm A and eq. (1.7.2)
defines the running time for algorithm B. Then for finding value of a for which
B is asymptotically faster than A we find asymptotic notation for the recurrence
by using Master’s method.
 n
Now, compare eq. (1.7.1) with T(n) = aT    f (n)
 b
we get, a = 7, b = 3
f(n) = n2
a 7
log
nlog b = n 3 = n2.81
Now, apply cases of Master’s, theorem as :
7
Case 3 : f(n) = O (nlog 3   )
 f(n) = O (n1.77 + )
 f(n) = O (n1.77 + 0.23)
 f(n) = O (n2)
Hence, case 3 of Master’s theorem is satisfied.
Thus, T(n) =  f(n)
 T(n) =  (n2)
Since recurrence (1) is asymptotically bounded by -notation which is
used to show optimum time we have to show that recurrence given by
eq. (1.7.2) is bounded by -notation which shows minimum time (best
case).
For the use satisfy the case 2 of Master theorem, Guess a = 81
 n 2
S(n) = f(n) = 81 S    n
9
 a = 81, b = 9
f(n) = nlog 9 81
f(n) =  (nlog b a ) = (n2)
Hence, case 2 of Master’s theorem is satisfied.
 T(n) =  ( nlog 9 81 log n)
Design and Analysis of Algorithms 1–9 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
 T(n) =  (n2 log n)
Therefore, this shows that B is asymptotically faster than A when
a = 81.

Que 1.8. Solve the following recurrences :

T(n) = T( n ) + O(log n)

Answer
T(n) = T( n ) + O(log n) ...(1.8.1)
m= log n
Let n= 2m
n1/2 = 2m/2 ...(1.8.2)
Put value of n in eq. (1.8.1) we get
T(2m) = T (2m / 2 )  O(log 2m ) ...(1.8.3)
x(m) = T(2m) ...(1.8.4)
Putting the value of x(m) in eq. (1.8.3)
 m
x(m) = x    O(m) ...(1.8.5)
2
Solution of eq. (1.8.5) is given as
a = 1, b = 2, f(n) = O(m)
mlogba = mlog21 + E where E = 1
x(m) = (log m)
T(n) = (log log n)

Que 1.9.
i. Solve the recurrence T (n) = 2T(n/2) + n2 + 2n + 1
ii. Prove that worst case running time of any comparison sort
is  (n log n). AKTU 2019-20, Marks 07

Answer
 n
i. T (n) = 2T(n/2) + n2 + 2n + 1  2T    n2
 2
 n
Compare it with T(n) = aT    f (n)
 b
we have, a = 2, b = 2, f(n) = n2
Now, we apply cases for Master’s theorem.
nlog b a = nlog 2 2 = n
This satisfies case 3 of Master’s theorem.

 f(n) =   nlog b a  E     n1  E 
=  (n1 + 1) where E = 1
=  (n2)
Introduction 1–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
 n2 
Again 2 f    c f(n2) ...(1.9.1)
 2
eq. (1.9.1) is true for c = 2
 T(n) =  (f(n))
 T(n) =  f(n2)
ii. Let T(n) be the time taken by merge sort to sort any array of n elements.

  n    n 
Therefore, T(n) = T         g(n)
  2    2 
where g(n)  (n)
This recurrence, which becomes :
 n
T(n) = 2T    g (n)
 2
when n is even is a special case of our general analysis for divide-and
conquer algorithms.
 n
Compare the above given recurrence with T(n) = aT    f (n)
 b
we get a= 2
b=2
f(n) = g(n)
log a log 2
Now we find, n b = n 2  n1  n
 f(n) = (n)
i.e., case 2 of Master’s theorem applied then

T(n) =  nlog b a log n 
 T(n) =  (n log n)
Hence, the worst case running time of merge sort is (n log n).

Que 1.10. What do you mean by recursion ? Explain your answer

with an example.

Answer
1. Recursion is a process of expressing a function that calls itself to perform
specific operation.
2. Indirect recursion occurs when one function calls another function that
then calls the first function.
3. Suppose P is a procedure containing either a call statement to itself or
a call statement to a second procedure that may eventually result in a
call statement back to the original procedure P. Then P is called recursive
procedure.
4. A recursive procedure must have the following two properties :
a. There must be certain criteria, called base criteria, for which the
procedure does not call itself.
Design and Analysis of Algorithms 1–11 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Each time the procedure does call itself, it must be closer to the
criteria.
5. A recursive procedure with these two properties is said to be well-
defined.
For example :
The factorial function may also be defined as follows :
a. If n = 0, then n! = 1.
Here, the value of n! is explicitly given when n = 0 (thus 0 is the
base value).
b. If n > 0, then n! = n. (n – 1)!
Here, the value of n! for arbitrary n is defined in terms of a smaller
value of n which is closer to the base value 0.
Observe that this definition of n! is recursive, since it refers to itself when it
uses (n – 1)!

Que 1.11. What is recursion tree ? Describe.

Answer
1. Recursion tree is a pictorial representation of an iteration method,
which is in the form of a tree, where at each level nodes are expanded.
2. In a recursion tree, each node represents the cost of a single subproblem.
3. Recursion trees are particularly useful when the recurrence describes
the running time of a divide and conquer algorithm.
4. A recursion tree is best used to generate a good guess, which is then
verified by the substitution method.
5. It is a method to analyze the complexity of an algorithm by diagramming
the recursive function calls in the form of tree.

Que 1.12. Solve the recurrence :

T(n) = T(n – 1) + T(n – 2) + 1, when T(0) = 0 and
T(1) = 1.

Answer
T(n) = T(n – 1) + T(n – 2) + 1
At kth level, T(1) will be equal to 1
when, n–k= 1
k= n–1
= 20 + 21 + 22 + ...... 2k
= 20 + 21 + 22 + ...... 2n–1
Introduction 1–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
T(n)

T(n – 1) T(n – 2) 1 20 = 1

T(n – 2) T(n – 3) 1 T(n – 3) T(n – 4) 1 21 = 2

T(n – 3) T(n – 4) 1 T(n – 4) T(n – 5) 1 T(n – 4) T(n – 5) 1 ...... 22 = 4

.............

T(n – k) 2k
n 0 n 1
 a(r  1)  2 (2  1)
Sum of n terms of geometric progression  =
 r 1  21
=2 n – 1 n
– 1 = O(2 )

Que 1.13. Solve the following recurrences :

T(n) = T(n/2) + T(n/4) + T(n/8) + n
AKTU 2019-20, Marks 07

Answer
7n 72 n
T(n) = n   2 + ........+ log n times
8 8
=  (n log n)
n n

n n n 7n
2 4 8 8
log n

2
n n n n n n n n n 7n
2
4 8 16 8 16 32 16 32 64 8

Que 1.14. Use a recursion tree to give an asymptotically tight

solution to the recurrence T(n) = T(n) + T((1 – )n) + cn, where  is
a constant in the range 0 <  <1 and c > 0 is also a constant.
AKTU 2018-19, Marks 07

Answer
T(n) = T(n) + T((1 – )n) + cn
Design and Analysis of Algorithms 1–13 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Recursion tree :
Cost
cn cn

c n (1 –  )cn cn

c 2 n c(1 –  ) n c(1 –  ) n c(1 –  ) n cn

cn
• • • • • •
• • • • • •
• • • • • •
•
T(1) T(1) T(1) T(1) (k + 1) times
Assuming T(1) = 1
So c(1 – )k n = 1
k
1  1 
cn = 
(1  )k  1   
 1 
log (n) = k log 
 1   
k= log cn
 log 1 (cn)
 1 
log  
1 

1   
So, Total cost = cn + cn + .... (k + 1) times = cn(k + 1)
log 1 (cn)
= cn ×
1 

Time complexity = O n log

 1
1 
(cn) = O n log
 1
1
n

Que 1.15. Solve the following by recursion tree method
T(n) = n + T(n/5) + T(4n/5)
AKTU 2017-18, Marks 10

Answer
T(n) = n + n + n + ................+ log n times =  (n log n)
n n

n 4n n
5 5
log n
n 4n 4n 16n
25 25 25 25 n

n 4n 4n 16n 4n 16n 16n 64n n

125 125 125 125 125 125 125 125
Introduction 1–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
PART-3
Sorting and Order Statistic : Shell Sort, Quick Sort.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 1.16. Explain shell sort with example.

Answer
1. Shell sort is a highly efficient sorting algorithm and is based on insertion
sort algorithm and we can code it easily.
2. It roughly sorts the data first, moving large elements towards one end
and small elements towards the other.
3. In shell sort several passes over the data is performed.
4. After the final pass, the data is fully sorted.
5. The shell sort does not sort the data itself; it increases the efficiency of
other sorting algorithms.
Algorithm :
Input : An array a of length n with array elements numbered 0 to n – 1.
1. inc  round (n/2)
2. while inc > 0
3. for i = inc to n – 1
temp  a[i]
ji
while j  inc and a [j – inc] > temp
a[j]  a[j – inc]
j  j – inc
a[j]  temp
4. inc  round (inc/2.2)
For example :

45 36 75 20 05 90 80 65 30 50 10 75 85
The distance between the elements to be compared is 3. The subfiles generated
with the distance of 3 are as follows :
Subfile 1 a[0] a[3] a[6] a[9] a[12]
Subfile 2 a[1] a[4] a[7] a[10]
Subfile 3 a[2] a[5] a[8] a[11]
Input to pass 1 with distance = 3
Design and Analysis of Algorithms 1–15 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
45 36 75 20 05 90 80 65 30 50 10 75 85

(a)
Output of pass 1 is input to pass 2 and distance = 2

20 05 30 45 10 75 50 36 75 80 65 90 85

(b)
Output of pass 2 is input to pass 3 and distance = 1
10 05 20 36 30 45 50 75 65 80 75 90 85

(c)
Output of pass 3

05 10 20 30 36 45 50 65 75 75 80 85 90

(d)
Fig. 1.16.1.

Que 1.17. Describe any one of the following sorting techniques :

i. Selection sort
ii. Insertion sort

Answer
i. Selection sort (A) :
1. n  length [A]
2. for j  1 to n–1
3. smallest  j
4. for i  j + 1 to n
5. if A [i] < A [smallest]
6. then smallest  i
7. exchange (A [j], A [smallest])
ii. Insertion_Sort(A) :
1. for j  2 to length[A]
2. do key  A[j]
3. Insert A[j] into the sorted sequence A[1 ..... j – 1]
4. i  j – 1
5. while i > 0 and A[i] > key
6. do A[i + 1]  A[i]
7. i  i – 1
8. A[i + 1]  key
Introduction 1–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 1.18. Write non-deterministic algorithm for sorting.

AKTU 2016-17, Marks 10

Answer
Non-deterministic algorithms are algorithm that, even for the same input,
can exhibit different behaviours on different runs, iterations and executions.
N SORT(A, B) :
1. for i = 1 to n do
2. j = choice(1 . . . n)
3. if B[j] != 0 then failure
4. B[j] = A[i]
5. endfor
6. for i = 1 to n – 1 do
7. if B[i] < B[i + 1] then failure
8. endfor
9. print(B)
10. success

Que 1.19. Explain the concepts of quick sort method and analyze

its complexity with suitable example. AKTU 2016-17, Marks 10

Answer
Quick sort :
Quick sort works by partitioning a given array A[p ... r] into two non-empty
subarray A[p ... q – 1] and A [q + 1 ... r] such that every key in A[p ... q – 1] is
less than or equal to every key in A[q + 1 ... r]. Then the two subarrays are
sorted by recursive calls to quick sort.
Quick_Sort (A, p, r)
1. If p < r then
2. q  Partition (A, p, r)
3. Recursive call to Quick_Sort (A, p, q – 1)
4. Recursive call to Quick_Sort (A, q + 1, r)
As a first step, Quick sort chooses as pivot one of the items in the array to be
sorted. Then array is partitioned on either side of the pivot. Elements that
are less than or equal to pivot will move toward the left and elements that are
greater than or equal to pivot will move toward the right.
Partition (A, p, r)
1. x  A[r]
2. i  p – 1
3. for j  p to r – 1
4. do if A[j]  x
5. then i  i + 1
6. then exchange A[i]  A[j]
Design and Analysis of Algorithms 1–17 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
7. exchange A[i + 1]  A[r]
8. return i + 1
Example : Given array to be sorted
3 1 4 1 5 9 2 6 5 3 5 8 9
Sort the array A using quick sort algorithm.
Step 1 : The array is Pivoted about it first element i.e., Pivot (P) = 3

3 1 4 1 5 9 2 6 5 3 5 8 9
P
Step 2 : Find first element larger then pivot (make underline) and find
element not larger than pivot from end make over line.

3 1 4 1 5 9 2 6 5 3 5 8 9
P
Underline Overline
Step 3 : Swap these element and scan again.
3 1 3 1 5 9 2 6 5 4 5 8 9
P
Array after swapping

3 1 3 1 5 9 2 6 5 4 5 8 9
P
Underline Overline
Apply swapping,

3 1 3 1 2 9 5 6 5 4 5 8 9
Again apply scanning,

3 1 3 1 2 9 5 6 5 4 5 8 9

Overline Underline
The pointers have crossed
i.e., overline on left of underlined
Then, in this situation swap pivot with overline.

2 1 3 1 3 9 5 6 5 4 5 8 9
P
Now, pivoting process is complete.
Step 4 : Recursively sort subarrays on each side of pivot.
Subarray 1 : 2 1 3 1

Subarray 2 : 9 5 6 5 1 5 8 9
First apply Quick sort for subarray 1.
Introduction 1–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2 1 3 1
P
Underline Overline

2 1 1 3
P
Underline Overline
The pointers have crossed.
i.e., overline on left of underlined.
Swap pivot with overline
1 1 2 3
Sorted array
Now, for subarray 2 we apply Quick sort procedure.

9 5 6 5 4 5 8 9
P
Overline Underline
The pointer has crossed. Then swap pivot with overline.

8 5 6 5 4 5 9 9

Subarray 4
Subarray 3

8 5 6 5 4 5
P Overline
Swap overline with pivot.

5 5 6 5 4 8
P
5 5 6 5 4
Underline Overline
5 5 4 5 6
Overline on left of underlined.
Swap pivot with overline.

4 5 5 5 6
P
Now combine all the subarrays
Sorted array 1 1 2 3 3 4 5 5 5 6 8 9 9
Pivot
Design and Analysis of Algorithms 1–19 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Analysis of complexity :
i. Worst case :
1. Let T(n) be the worst case time for quick sort on input size n. We
have a recurrence
T(n) = max (T(q) + T(n – q – 1)) + (n) ...(1.19.1)
o q n 1
where q ranges from 0 to n – 1, since the partition produces two
regions, each having size n – 1.
2. Now we assume that T(n)  cn2 for some constant c.
Substituting our assumption in eq. (1.19.1) we get
T(n)  max (cq2 + c(n – q – 1)2) + (n)
o q n 1

= max (q2 + (n – q – 1)2) + (n)

o q n 1
3. Since the second derivative of expression q2 + (n – q – 1)2 with
respect to q is positive. Therefore, expression achieves a maximum
over the range 0  q  n – 1 at one of the endpoints.
4. This gives the bound
max0  q  n – 1 (q2 + (n – q – 1)2) (n – 1)2 = n2 – 2n + 1
5. Continuing with the bounding of T(n) we get
T(n)  cn2 – c(2n – 1) + (n) cn2
6. Since we can pick the constant c large enough so that the c(2n – 1)
term dominates the (n) term. We have
T(n) = O(n2 )
7. Thus, the worst case running time of quick sort is (n2).
ii. Average case :
1. If the split induced of RANDOMIZED_PARTITION puts constant
fraction of elements on one side of the partition, then the
recurrence tree has depth (log n) and (n) work is performed at
each level.
2. This is an intuitive argument why the average case running time
of RANDOMIZED_QUICKSORT is (n log n).
3. Let T(n) denotes the average time required to sort an array of n
elements. A call to RANDOMIZED_ QUICKSORT with a 1 element
array takes a constant time, so we have T(1) = (1).
4. After the split RANDOMIZED_QUICKSORT calls itself to sort
two subarrays.
5. The average time to sort an array A[1 .. q] is T[q] and the average
time to sort an array A[q+1 .. n] is T[n – q]. We have
T(n) = 1/n (T(1) + T(n – 1) + n–1Sq=1 T(q)
T(n – q))) + (n) ...(1.19.1)
We know from worst-case analysis
T(1) =  (1) and T(n – 1) = O(n2)
T(n) = 1/n ( (1) + O(n2)) + 1/n n–1q=1 (r(q)
+ T(n – q)) + Q(n)
= 1/n n–1Sq=1 (T(q) + T(n – q)) + Q(n) ...(1.19.2)
= 1/n [2 n–1Sk=1(T(k))] +  (n)
= 2/n n–1Sk=1 (T(k)) +  (n) ...(1.19.3)
Introduction 1–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6. Solve the above recurrence using substitution method. Assume
that T(n)  an n log n + b for some constants a > 0 and b > 0.
If we can pick ‘a’ and ‘b’ large enough so that n log n + b > T(1). Then for
n > 1, we have
T(n)  n–1k=1 2/n (ak log k + b) + (n)
= 2a/n n–1Sk=1 k log k – 1/8(n2) + 2b/n
(n – 1) + n ...(1.19.4)
At this point we are claiming that
n–1 2 2
k=1 k log k  1/2 n log n – 1/8(n )
Substituting this claim in the eq. (1.19.4), we get
T(n)  2a/n [1/2n2 log n – 1/8(n2)] + 2/n b(n – 1) +  (n)
 an log n – an/4 + 2b + (n) ... (1.19.5)
In the eq. (1.19.5), (n) + b and an/4 are polynomials and we can choose ‘a’
large enough so that an/4 dominates (n) + b.
We conclude that QUICKSORT’s average running time is (n log n).

Que 1.20. Discuss the best case and worst case complexities of
quick sort algorithm in detail.

Answer
Best case :
1. The best thing that could happen in quick sort would be that each
partitioning stage divides the array exactly in half.
2. In other words, the best to be a median of the keys in A[p .. r] every
time procedure ‘Partition’ is called.
3. The procedure ‘Partition’ always split the array to be sorted into two
equal sized arrays.
4. If the procedure ‘Partition’ produces two regions of size n/2, then the
recurrence relation is :
T(n)  T(n/2) + T(n/2) + (n) 2T(n/2) + (n)
And from case (2) of master theorem
T(n) = (n log n)
Worst case : Refer Q. 1.19, Page 1–16B, Unit-1.

PART-4
Merge Sort.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 1.21. Explain the concept of merge sort with example.

AKTU 2016-17, Marks 10

Design and Analysis of Algorithms 1–21 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer
1. Merge sort is a sorting algorithm that uses the idea of divide and conquer.
2. This algorithm divides the array into two halves, sorts them separately
and then merges them.
3. This procedure is recursive, with the base criteria that the number of
elements in the array is not more than 1.
Algorithm :
MERGE_SORT (a, p, r)
1. if p < r
2. then q (p + r)/2
3. MERGE-SORT (A, p, q)
4. MERGE-SORT (A, q + 1, r)
5. MERGE (A, p, q, r)
MERGE (A, p, q, r)
1. n1 = q – p + 1
2. n2 = r – q
3. Create arrays L [1 .........n1 + 1] and
R [1......n2 + 1]
4. for i = 1 to n1
do
L[i] = A [p + i – 1]
endfor
5. for j = 1 to n2
do
R[j] = A[q + j]
endfor
6. L[n1 + 1] = , R[n2 + 1] = 
7. i = 1, j = 1
8. for k = p to r
do
if L[i]  R[j]
then A[k]  L[i]
i=i+1
else A[k] = R[j]
j=j+1
endif
endfor
9. exit
Example :
10, 25, 16, 5, 35, 48, 8
1. Divide into two halves : 10, 25, 16, 5 35, 48, 8
2. Consider the first part : 10, 25, 16, 5 again divide into two sub-
arrays
Introduction 1–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
10 , 25 16 , 5

10, 25 5, 16

5, 10, 16, 25
3. Consider the second half : 35, 48, 5 again divide into two sub-arrays
35 , 48 8

35, 48 8

8, 35, 48
4. Merge these two sorted sub-arrays,
5, 10, 16, 25 8, 35, 48

5, 8, 10, 16, 25, 35, 45

This is the sorted array.

Que 1.22. Determine the best case time complexity of merge sort
algorithm.

Answer
1. The best case of merge sort occurs when the largest element of one
array is smaller than any element in the other array.
2. For this case only n/2 comparisons of array elements are made.
3. Merge sort comparisons are obtained by the recurrence equation of
the recursive calls used in merge sort.
4. As it divides the array into half so the recurrence function is defined
as :
 n  n  n
T(n) = T    T   + n = 2T   + n ...(1.22.1)
 2  2  2
5. By using variable k to indicate depth of the recursion, we get
 n
T(n) = 2kT  k  + kn ...(1.22.2)
2 
6. For the best case there are only n/2 comparisons hence equation (1.22.2)
can be written as
 n n
T(n) = 2k  k   k
2  2
7. At the last level of recursion tree
2k = n
Design and Analysis of Algorithms 1–23 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
k = log2 n
8. So the recurrence function is defined as :
 n  n
T(n) = 2log2 n T  log n   log 2 n
2 2  2
n n
= nT(1) + log2 n = log2 n + n
2 2
T(n) = O(n log2 n)
Hence, the best case complexity of merge sort is O(n log2 n).

PART-5
Heap Sort.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 1.23. Explain heap sort algorithm with its complexity.

OR
Discuss Max-Heapify and Build-Max-Heap procedures.

Answer
1. Heap sort is a comparison based sorting technique based on binary heap
data structure.
2. Heap sort finds the largest element and puts it at the end of array, then
the second largest item is found and this process is repeated for all other
elements.
3. The general approach of heap sort is as follows :
a. From the given array, build the initial max heap.
b. Interchange the root (maximum) element with the last element.
c. Use repetitive downward operation from root node to rebuild the
heap of size one less than the starting.
d. Repeat step a and b until there are no more elements.
Analysis of heap sort :
Complexity of heap sort for all cases is O(n log2 n).
MAX-HEAPIFY (A, i) :
1. i  left [i]
2. r  right [i]
3. if l  heap-size [A] and A[l] > A[i]
4. then largest  l
5. else largest  i
6. if r  heap-size [A] and A[r] > A [largest]
7. then largest  r
Introduction 1–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
8. if largest  i
9. then exchange A[i]  A[largest]
10. MAX-HEAPIFY [A, largest]
HEAP-SORT(A) :
1. BUILD-MAX-HEAP (A)
2. for i  length [A] down to 2
3. do exchange A[1]  A[i]
4. heap-size [A]  heap-size [A] – 1
5. MAX-HEAPIFY (A, 1)
BUILD-MAX-HEAP (A)
1. heap-size (A)  length [A]
2. for i  (length [A]/2) down to 1 do
3. MAX-HEAPIFY (A, i)
We can build a heap from an unordered array in linear time.
Average case and worst case complexity :
1. We have seen that the running time of BUILD-HEAP is O(n).
2. The heap sort algorithm makes a call to BUILD-HEAP for creating a
(max) heap, which will take O(n) time and each of the (n – 1) calls to
MAX-HEAPIFY to fix up the new heap (which is created after
exchanging the root and by decreasing the heap size).
3. We know ‘MAX-HEAPIFY’ takes time O(log n).
4. Thus the total running time for the heap sort is O(n log n).

Que 1.24. How will you sort following array A of element using
heap sort : A = (23, 9, 18, 45, 5, 9, 1, 17, 6).
AKTU 2018-19, Marks 10

Answer
Given array : 23 9 18 45 5 9 1 17 6
First we call Build-Max heap
heap size [A] = 9
23

9 18
i=4
45 5 9 1

l=8 r=9
17 6
so i = 4 to 1 call MAX HEAPIFY (A, i)
i.e., first we call MAX HEAPIFY (A, 4)
A[l] = 7, A[i] = A[4] = 45, A[r] = 6
l  left [4] = 2 × 4 = 8
r  right[4] = 2 × 4 + 1 = 9
8  9 and A[8] = 17 < 45 (False)
Then, largest  4.
Design and Analysis of Algorithms 1–25 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Similarly for i = 3, 2, 1 we get the following heap tree :

i 23
23 9 18
i=3 l r
9 18
l=6 r=7 45 5 9 1
45 5 9 1
17 6
17 6

45
i
23 23 18
r
i
45 l 18
9 1
5 9
l
9 5 9 1 r
17 6
17 6

So, final tree after Build-Max heap is

45
23 18

17 5 9 1

9 6

Now i = 9 down to 2 and size = 10 – 1 = 9 and call MAX HEAPIFY

(A, 1) each time
Exchanging A[1]  A[9]

45 6
23 18 23 18

17 5 9 1 17 5 9 1

9 6 9 45

45

Now call MAX HEAPIFY (A, 1) and

Exchange A[1] and A[8], size = 9 – 1 = 8
Introduction 1–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
23 6
17 18 17 18

9 5 9 1 9 5 9 1

6 23
23 45
Now, call MAX HEAPIFY (A, 1), exchange A[1] and A[4] and size = 5 – 1
=4

9 9

5 6 5 6

1 1

9 9 17 18 23 45
Now, call MAX HEAPIFY (A, 1), exchange A[1] and A[3] and size = 4 –
1=3

9 9

1 6 1 6

6 9 9 17 18 23 45
Exchange A[1] and A[2] size 3 – 1 = 2
5 5
1 1 1

5 6 9 9 17 18 23 45
The sorted array
1 5 6 9 9 17 18 23 45
Now call MAX HEAPIFY (A, 1) and exchange A[1] and A[7]

18 1
9 17 9 17
6 5 9 1 6 5 9 18
Now call MAX HEAPIFY (A, 1) and size = 7 – 1 = 6 exchange A[1] and
A[6]

17 17
6 9 6 9

1 5 9 1 5 9
Design and Analysis of Algorithms 1–27 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Exchange A[1] and A[5] and size = 6 – 1 = 5
9 5
6 9 6 9

1 5 1 9

Que 1.25. What is heap sort ? Apply heap sort algorithm for
sorting 1, 2, 3, 4, 5, 6, 7, 8, 9, 10. Also deduce time complexity of heap
sort. AKTU 2015-16, Marks 10

Answer
Heap sort and its time complexity : Refer Q. 1.23, Page 1–23B, Unit-1.
Numerical : Since the given problem is already in sorted form. So, there is
no need to apply any procedure on given problem.

Que 1.26. Explain HEAP SORT on the array. Illustrate the

operation HEAP SORT on the array A = {6, 14, 3, 25, 2, 10, 20, 7, 6}
AKTU 2017-18, Marks 10

Answer
Heap sort : Refer Q. 1.23, Page 1–23B, Unit-1.
Numerical :
Originally the given array is : [6, 14, 3, 25, 2, 10, 20, 7, 6]
First we call BUILD-MAX-HEAP
heap size [A] = 9
6
14 3

25 2 10 20

7 6
so, i = 4 to 1, call MAX-HEAPIFY (A, i)
i.e., first we call MAX-HEAPIFY (A, 4)
Az [l] = 7, A [i] = A [4] = 25, A [r] = 6
l  left [4] = 8
r  right [4] = 9
8  9 and 7 > 25 (False)
Then, largest  4
9  9 and 6 > 25 (False)
Then, largest = 4
A [i]  A [4]
Now call MAX-HEAPIFY (A, 2)
Introduction 1–28 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6
14 3
i=4
25 2 10 20
l=8
7 6 r=9
(i)
Similarly for i = 3, 2, 1 we get the following heap tree.
6 i 6
i
14 3 14 r 20
l
25 2 10 20 25 2 10 3
l r
7 6 7 6
(ii) (iii)
i
6 i 25
r
l 25 20 6 r 20
l
14 2 10 3 14 2 10 3

7 6 7 6
(iv) (v)

25 25
14 20 14 20
i
6 2 10 3 7 2 10 3

l 7 6 r 6 6
(vi) (vii)
So final tree after BUILD-MAX-HEAP is

25
14 20

7 2 10 3

6 6
(viii)

Now i = 9 down to 2, and size = size – 1 and call MAX-HEAPIFY

(A, 1) each time.
exchanging A [1]  A [9]
Design and Analysis of Algorithms 1–29 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
25 6
14 20 14 20

7 2 10 3 7 2 10 3

6 6 6 25

6 14 20 7 2 10 3 6 25
Now call MAX-HEAPIFY (A, 1) we get
Now exchange A [1] and A [8] and size = 8 – 1 = 7

20 6
14 10 14 10
7 2 6 3 7 2 3
6
6 20

6 14 10 7 2 6 3 20
Again call MAX-HEAPIFY (A, 1), we get
exchange A [1] and A [7] and size = 7 – 1 = 6

14 3
7 10 7 10

6 2 6 3 6 2 6 14

3 7 10 6 2 6 14
Again call MAX-HEAPIFY (A, 1), we get
exchange A [1] and A [6] and now size = 6 – 1 = 5

10 3
7 6 7 6

6 2 3 6 2 10

3 7 6 6 2 10
Again call MAX-HEAPIFY (A, 1)
exchange A [1] and A [5] and now size = 5 – 1 = 4

7 2
6 6 6 6

3 2 3 7

2 6 6 3 7
Introduction 1–30 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Again, call MAX-HEAPIFY (A, 1)
exchange A [1] and A [4] and size = 4 – 1 = 3

6 3
6 2 6 2

3 6

3 6 2 6

call MAX-HEAPIFY (A, 1)

exchange A [1] and A [3], size = 3 – 1 = 2

6 2
3 2 3 6

2 3 6

call MAX-HEAPIFY (A, 1)

exchange A [1] and A [2] and size = 2 – 1 = 1

3 2
2 3
2 3

Thus, sorted array :

2 3 6 6 7 10 14 20 25

PART-6
Comparison of Sorting Algorithms, Sorting in Linear Time.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 1.27. How will you compare various sorting algorithms ?

Design and Analysis of Algorithms 1–31 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer

Name Average Worst Stable Method Other notes

case case

Selection O(n2) O(n2) No Selection Can be

sort implemented as
a stable sort

Insertion O(n2) O(n2) Yes Insertion Average case is

sort also O(n + d),
where d is the
number of
inversion

Shell – O(n log2 n) No Insertion No extra

sort memory required

Merge O(n log n) O(n log n) Yes Merging Recursive, extra

sort memory required

Heap O(n log n) O(n log n) No Selection Recursive, extra

sort memory required

Quick O(n log n) O(n2) No Partitioning Recursive, based

sort on divide conquer
technique

Que 1.28. Explain the counting sort algorithm.

Answer
Counting sort is a linear time sorting algorithm used to sort items when they
belong to a fixed and finite set.
Algorithm :
Counting_Sort(A, B, k)
1. let C[0..k] be a new array
2. for i  0 to k
3. do C[i]  0
4. for j  1 to length[A]
5. do C[A[j]]  C[A[j]] + 1
// C[i] now contains the number of elements equal to i.
6. for i  1 to k
7. do C[i]  C[i] + C[i – 1]
// C[i] now contains the number of elements less than or equal to i.
8. for j  length[A] down to 1
9. do B[C[A[j]]]  A[j]
10. C[A[j]]  C[A[j]] – 1
Introduction 1–32 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 1.29. What is the time complexity of counting sort ? Illustrate
the operation of counting sort on array A = {1, 6, 3, 3, 4, 5, 6, 3, 4, 5}.

Answer
Given array : Time complexity of counting sort is O(n).
1 2 3 4 5 6 7 8 9 10
A 1 6 3 3 4 5 6 3 4 5
Step 1 : i = 0 to 6 k = 6 (largest element in array A)
C[i] 0
0 1 2 3 4 5 6
C 0 0 0 0 0 0 0
Step 2 : j = 1 to 10 ( length [A] = 10)
C[A[j]]  C[A[j]] + 1
For j = 1
0 1 2 3 4 5 6
C[A[1]]  C[1] + 1 = 0 + 1 = 1 C 0 1 0 0 0 0 0
C[1]  1
For j = 2
0 1 2 3 4 5 6
C[A[2]]  C[6] + 1 = 0 + 1 = 1 C 0 1 0 0 0 0 1
C[6]  1
0 1 2 3 4 5 6
Similarly for j = 5, 6, 7, 8, 9, 10 C 0 1 0 3 2 2 2
Step 3 :
For i = 1 to 6
C[i] C[i] + C[i – 1]
For i = 1
0 1 2 3 4 5 6
C[1]  C[1] + C[0] C 0 1 0 3 2 2 2
C[1]  1 + 0 = 1
For i = 2
0 1 2 3 4 5 6
C[2]  C[2] + C[1] C 0 1 1 3 2 2 2
C[1]  1 + 0 = 1
0 1 2 3 4 5 6
Similarly for i = 4, 5, 6 C 0 1 1 4 6 8 10
Step 4 :
For j = 10 to 1
B[C[A[j]]] A [j]
C[A[j ]] C[A[j] – 1
Design and Analysis of Algorithms 1–33 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
j A[j] C[A[j ]] B[C[A[j]]]  A[j] C[A[j]]  C[A[j]] – 1
10 5 8 B [8]  5 C[5]  7
9 4 6 B [6]  4 C[4]  5
8 3 4 B [4]  3 C[3]  3
7 6 10 B [10]  6 C[6]  9
6 5 7 B [7]  5 C[5]  6
5 4 5 B [5]  4 C[4]  4
4 3 3 B [3]  3 C[3]  2
3 3 2 B [2]  3 C[3]  1
2 6 9 B [9]  6 C[6]  8
1 1 1 B [1]  1 C[1]  0
1 2 3 4 5 6 7 8 9 10
B 1 3 3 3 4 4 5 5 6 6

Que 1.30. Write the bucket sort algorithm.

Answer
1. The bucket sort is used to divide the interval [0, 1] into n equal-sized
sub-intervals, or bucket, and then distribute the n-input numbers into
the bucket.
2. Since the inputs are uniformly distributed over [0, 1], we do not except
many numbers to fall into each bucket.
3. To produce the output, simply sort the numbers in each bucket and then
go through the bucket in order, listing the elements in each.
4. The code assumes that input is in n-element array A and each element
in A satisfies 0  A[i]  1. We also need an auxiliary array B[0 ... n – 1] for
linked-list (buckets).
BUCKET_SORT (A)
1. n  length [A]
2. for i  1 to n
3. do Insert A[i] into list B   nA[ i] 
4. for i  0 to n – 1
5. do Sort list B[i] with insertion sort
6. Concatenate the lists B[0], B[1], …. B[n – 1] together in order.

Que 1.31. What do you mean by stable sort algorithms ? Explain it

with suitable example.
Introduction 1–34 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer
1. A sorting algorithm is said to be stable if two objects with equal keys
appear in the same order in sorted output as they appear in the input
sorted array.
2. A stable sort is one where the initial order of equal items is preserved.
3. Some sorting algorithms are stable by nature, such as bubble sort,
insertion sort, merge sort, counting sort etc.
4. Let A be an array, and let < be a strict weak ordering on the elements of
A. Sorting algorithm is stable if :
i < j and A[i]  A[j] i.e., A[i] comes before A[j].
5. Stability means that equivalent elements retain their relative positions,
after sorting.
For example :
10 20 20 30 10

Sorting is stable because the

order of balls is maintained when
values are same. The bal with gray
color and value 10 appears before the
white color ball with value 10.
Similarly order is maintained for 20.

10 10 20 20 30

Que 1.32. Write a short note on radix sort.

Answer
1. Radix sort is a sorting algorithm which consists of list of integers or
words and each has d-digit.
2. We can start sorting on the least significant digit or on the most significant
digit.
3. On the first pass entire numbers sort on the least significant digit (or
most significant digit) and combine in a array.
4. Then on the second pass, the entire numbers are sorted again on the
second least significant digits and combine in an array and so on.
RADIX_SORT (A, d)
1. for i  1 to d do
2. use a stable sort to sort array A on digit i
// counting sort will do the job
The code for radix sort assumes that each element in the n-element
array A has d-digits, where digit 1 is the lowest-order digit and d is the
highest-order digit.
Analysis :
1. The running time depends on the table used as an intermediate sorting
algorithm.
Design and Analysis of Algorithms 1–35 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. When each digit is in the range 1 to k, and k is not too large,
COUNTING_SORT is the obvious choice.
3. In case of counting sort, each pass over n d-digit numbers takes
 (n + k) time.
4. There are d passes, so the total time for radix sort is (n + k) time. There
are d passes, so the total time for radix sort is (dn + kd). When d is
constant and k = (n), the radix sort runs in linear time.
For example : This example shows how radix sort operates on seven 3-
digit number.
Table 1.32.1.
Input 1st pass 2nd pass 3rd pass
329 720 720 329
457 355 329 355
657 436 436 436
839 457 839 457
436 657 355 657
720 329 457 720
355 839 657 839
In the table 1.32.1, the first column is the input and the remaining
shows the list after successive sorts on increasingly significant digits
position.

Que 1.33. Among Merge sort, Insertion sort and quick sort which
sorting technique is the best in worst case. Apply the best one among
these algorithms to sort the list E, X, A, M, P, L, E in alphabetic order.
AKTU 2019-20, Marks 07

Answer
Merge sort technique is best in worst case because of its time complexity
O(n log n).
Numerical :
Given : E, X, A, M, P, L, E
Pass 1 : Merge each pair of element to obtain sorted list :
E X A M P L E
After sorting each pair, we get
E X A M L P E
Pass 2 : Merge each pair to obtain the list :
A E M X E L P
Pass 3 : Again merge the two sub arrays to obtain the list :
A E E L M P X
Introduction 1–36 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
VERY IMPORTANT QUESTIONS
Following questions are very important. These questions
may be asked in your SESSIONALS as well as
UNIVERSITY EXAMINATION.

Q. 1. What do you mean by algorithm ? Write its characteristics.

Ans. Refer Q. 1.1.

Q. 2. Write short note on asymptotic notations.

Ans. Refer Q. 1.3.

Q. 3. i. Solve the recurrence T (n) = 2T(n/2) + n2 + 2n + 1

ii. Prove that worst case running time of any comparison sort
is  (n log n).
Ans. Refer Q. 1.9.

Q. 4. Explain shell sort with example.

Ans. Refer Q. 1.16.

Q. 5. Discuss quick sort method and analyze its complexity.

Ans. Refer Q. 1.19.

Q. 6. Explain the concept of merge sort with example.

Ans. Refer Q. 1.21.

Q. 7. Write short note on heap sort algorithm with its analysis.

Ans. Refer Q. 1.23.

Q. 8. Explain HEAP SORT on the array. Illustrate the operation

HEAP SORT on the array A = {6, 14, 3, 25, 2, 10, 20, 7, 6}
Ans. Refer Q. 1.26.


 More pdf : www.motivationbank.in
2–1 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

2 Advanced Data
Structure

CONTENTS
Part-1 : Red-Black Trees ........................................ 2–2B to 2–19B

Part-2 : B-Trees ...................................................... 2–19B to 2–33B

Part-3 : Binomial Heaps ....................................... 2–33B to 2–44B

Part-4 : Fibonacci Heaps ...................................... 2–44B to 2–48B

Part-5 : Tries, Skip List ........................................ 2–48B to 2–51B

Advanced Data Structure 2–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
PART-1
Red-Black Trees.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 2.1. Define a red-black tree with its properties. Explain

the insertion operation in a red-black tree.

Answer
Red-black tree :
A red-black tree is a binary tree where each node has colour as an extra
attribute, either red or black. It is a self-balancing Binary Search Tree (BST)
where every node follows following properties :
1. Every node is either red or black.
2. The root is black.
3. Every leaf (NIL) is black.
4. If a node is red, then both its children are black.
5. For each node, all paths from the node to descendent leave contain the
same number of black nodes.
Insertion :
i. We begin by adding the node as we do in a simple binary search tree
and colouring it red.
RB-INSERT(T, z)
1. y  nil [T]
2. x  root [T]
3. while x  nil [T]
4. do y  x
5. if key[z] < key [x]
6. then x  left [x]
7. else x  right [x]
8. p[z]  y
9. if y = nil [T]
10. then root [T]  z
11. else if key [z] < key[y]
12. then left [y] z
13. else right [y]  z
14. left [z]  nil[T]
15. right [z]  nil[T]
16. colour [z]  RED
17. RB-INSERT-FIXUP(T, z)
 More pdf : www.motivationbank.in
2–3 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

ii. Now, for any colour violation, RB-INSERT-FIXUP procedure is used.

RB-INSERT-FIXUP(T, z)
1. while colour [p[z]] = RED
2. do if p[z] = left[p [p [z]]]
3. then y  right[p [p [z]]]
4. if colour[y] = RED
5. then colour[p [z]]  BLACK  case 1
6. colour[y]  BLACK  case 1
7. colour[p [p [z]]] RED  case 1
8. z  p [p [ z]]  case 1
9. else if z = right[p[z]]
10. then z  p[z]  case 2
11. LEFT-ROTATE(T, z)  case 2
12. colour[p[z]]  BLACK  case 3
13. colour[p[p[z]]]  RED  case 3
14. RIGHT-ROTATE(T, p[p[z]])  case 3
15. else (same as then clause with “right” and “left” exchanged)
16. colour[root[T]]  BLACK
Cases of RB-tree for insertion :
Case 1 : z’s uncle is red :
P[z] = left[p[p[z]]]
then uncle  right[p[p[z]]]
a. change z’s grandparent to red.
b. change z’s uncle and parent to black.
c. change z to z’s grandparent.

B B
11 11
R B R B
14 2 14
B B B R y
1 7 Case 1 1 7 New z

R R R R
5 8 y 5 8
R R
4 4
z ( a) Fig. 2.1.1. (b )

Now, in this case violation of property 4 occurs, because z’s uncle y is red,
then case 1 is applied.
Case 2 : z’s uncle is black, z is the right of its parent :
a. Change z to z’s parent.
b. Rotate z’s parent left to make case 3.
Advanced Data Structure 2–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B B
11 11
B R B
R
2 14 7 14
B R y R B R
1 7 z 2 8 15
Case 2 New z
B B
B B
5 8 5
R
4 R 4
( a) Fig. 2.1.2. (b )
Case 3 : z’s uncle is black, z is the left child of its parent :
a. Set z’s parent black.
b. Set z’s grandparent to red.
c. Rotate z’s grandparent right.
B
11 B
7
R B z R R
11
7 14 y 2
B B B B
R B 8 14
z 2 8 Case 3 1 5
15
B B R
4 R
1 5
R
4 (a) Fig. 2.1.3. (b )

Que 2.2. What are the advantages of red-black tree over binary
search tree ? Write algorithms to insert a key in a red-black tree
insert the following sequence of information in an empty red-black
tree 1, 2, 3, 4, 5, 5.

Answer
Advantages of RB-tree over binary search tree :
1. The main advantage of red-black trees over AVL trees is that a single
top-down pass may be used in both insertion and deletion operations.
2. Red-black trees are self-balancing while on the other hand, simple binary
search trees are unbalanced.
3. It is particularly useful when inserts and/or deletes are relatively frequent.
4. Time complexity of red-black tree is O(log n) while on the other hand, a
simple BST has time complexity of O(n).
Algorithm to insert a key in a red-black tree : Refer Q. 2.1, Page 2–2B,
Unit-2.
 More pdf : www.motivationbank.in
2–5 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Numerical :

Insert 1 : B
1
B
Insert 2 : 1
R
2

B
Insert 3 : B
1 2
R Left
2 R R
rotate 1 3
R
B 3
B
Insert 4 : 2 2
R B B
1 3 1 B
3
R R
4 4
B
B
Insert 5 : 2 Left 2
B B
B
1 3 1 4 B
R R
4 5
3 R
B 5 R
Insert 5 : 2
B
1 B
4

B B
3 5

5
R
Fig. 2.2.1.

Que 2.3. Explain red-black tree. Show steps of inserting the keys
41, 38, 31, 12, 19, 8 into initially empty red-black tree.
OR
What is red-black tree ? Write an algorithm to insert a node in an
empty red-black tree explain with suitable example.
AKTU 2017-18, Marks 10
Advanced Data Structure 2–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer
Red-black tree and insertion algorithm : Refer Q. 2.1, Page 2–2B, Unit-2.
Numerical :
B
Insert 41 : 41
B
41
Insert 38 : R
38
Insert 31 :
B B
41
R 38
Case 3 R R
38
R 31 41
31
Insert 12 :
B
38 B
B R 38
Case 1 R R
31 41
R 31 41
12 R
12
Insert 19 :
B B
38
B B 38
Case 2, 3 B B
31 41
R 19 41
12 R R
R 12 31
19
Insert 8 :
B B
38 38
B B R B
Case 1
19 41 19 41
R R B B
12 31 12 31
R R
8 8
Thus final tree is

B
38
R B
19 41
B
B
12 31
R
8
 More pdf : www.motivationbank.in
2–7 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Que 2.4. Explain insertion in red-black tree. Show steps for

inserting 1, 2, 3, 4, 5, 6, 7, 8 and 9 into empty RB-tree.
AKTU 2015-16, Marks 10

Answer
Insertion in red-black tree : Refer Q. 2.1, Page 2–2B, Unit-2.
B
Insert 1 : 1

B
1
Insert 2 : R
2
Insert 3 :
B
B
1 Case 2 2
R
2 1 3
R R R
3

Insert 4 :

B B
2 2

1 3 1 3
R R B B 4
4
R R

Insert 5 :
B B
2 Case 2 2 B
B B
1 3 1 4
R B R
4 3 5
R
5 R

Insert 6 :

B B
2 2
B B B R
1 4 1 4
R
3 5 3 5
R
R 6 B B 6
R
Advanced Data Structure 2–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 7 :
B B
2 Case 2 2
B B B R
1 4 1 4
R B
3 5 B 3 6
R R
R 6 R 5 7
R
7

Insert 8 :
B B
2 2
B R B R
1 4 1 4
B R
B 3 6 B 3 6
R B
R 5 7 B 5 7
R R
8 8

Insert 9 :
B
2 Case 2 2
B R
1 4 1 4
R
B 3 6 3 6
B
R 5 7 5 8
R
8 7 9
R
9

Que 2.5. How to remove a node from RB-tree ? Discuss all cases
and write down the algorithm.

Answer
To remove a node from RB-tree RB-DELETE procedure is used. In
RB-DELETE procedure, after splitting out a node, it calls an auxiliary
procedure RB-DELETE-FIXUP that changes colours and performs rotations
to restore the red-black properties.
RB-DELETE(T, z)
1. if left[z] = nil[T] or right[z] = nil[T]
2. then y  z
3. else y  TREE-SUCCESSOR(z)
4. if left[y]  nil[T]
5. then x  left[y]
6. else x  right[y]
7. p[x]  p[y]
8. if p[y] = nil[T]
 More pdf : www.motivationbank.in
2–9 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

9. then root[T]  x
10. else if y = left[p[y]]
11. then left[p[y]]  x
12. else right[p[y]]  x
13. if y  z
14. then key[z]  key[y]
15. copy y’s sibling data into z
16. if colour[y] = BLACK
17. then RB-DELETE-FIXUP(T, x)
18. return y
RB-DELETE-FIXUP(T, x)
1. while x  root[T] and colour[x] = BLACK
2. do if x = left[p[x]]
3. then w  right[p[x]]
4. if colour[w] = RED
5. then colour[w]  BLACK  case 1
6. colour[p[x]]  RED  case 1
7. LEFT-ROTATE(T, p[x])  case 1
8. w  right[p[x]]  case 1
9. if colour[left[w]] = BLACK and colour[right[w]] = BLACK
10. then colour[w]  RED  case 2
11. x  p[x]  case 2
12. else if colour[right[w]] = BLACK
13. then colour[left[w]]  BLACK  case 3
14. colour[w]  RED  case 3
15. RIGHT-ROTATE(T, w)  case 3
16. w  right[p[x]]  case 3
17. colour[w]  colour[p[x]]  case 4
18. colour[p[x]]  BLACK  case 4
19. colour[right[w]]  BLACK  case 4
20. LEFT-ROTATE(T, p[x])  case 4
21. x  root[T]  case 4
22. else (same as then clause with “right” and “left” exchanged).
23. colour[x]  BLACK
Cases of RB-tree for deletion :
Case 1 : x’s sibling w is red :
1. It occurs when node w the sibling of node x, is red.
2. Since w must have black children, we can switch the colours of w
and p[x] and then perform a left-rotation on p[x] without violating
any of the red-black properties.
3. The new sibling of x, which is one of w’s children prior to the
rotation, is now black, thus we have converted case 1 into case 2, 3
or 4.
4. Case 2, 3 and 4 occur when node w is black. They are distinguished
by colours of w’s children.
Advanced Data Structure 2–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B B
B D
B R
x Case 1
B
A D B R E

  B B ^ 
x AB
C E New C B
w
  ^     
(a) (b )
Fig. 2.5.1.

Case 2 : x’s sibling w is black, and both of w’s children are black :
1. Both of w’s children are black. Since w is also black, we take one
black of both x and w, leaving x with only one black and leaving w
red.
2. For removing one black from x and w, we add an extra black to
p[x], which was originally either red or black.
3. We do so by repeating the while loop with p[x] as the new node x.
4. If we enter in case 2 through case 1, the new node x is red and
black, the original p[x] was red.
5. The value c of the colour attribute of the new node x is red, and
the loop terminates when it tests the loop condition. The new
node x is then coloured black.

B New x B
B B
B Case 2 R
x x
A D w A D

  B B   B B
C E C E

     
^ ^
( a) (b )
Fig. 2.5.2.
Case 3 : x’s sibling w is black, w’s left child is red, and w’s right child
is black :
1. Case 3 occurs when w is black, its left child is red and its right child
is black.
2. We can switch the colours of w and its left child left[w] and then
perform a right rotation on w without violating any of the red-
black properties, the new sibling w of x is a black node with a red
right child and thus we have transformed case 3 into case 4.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 2–11 B (CS/IT-Sem-5)

B B
B
R Case 3
x B
A D w A B C New x

  B B    R
D
C E
B
    E
^
(a) (b )
Fig. 2.5.3. 
^
Case 4 : x’s sibling w is black, and w’s right child is red :
1. When node x’s sibling w is black and w’s right child is red.
2. By making some colour changes and performing a left rotation on
p[x], we can remove the extra black on x, making it singly black,
without violating any of the red-black properties.

B D
B B
B Case 4 B
x
A D w B E

  B ^ 
R
B C
C C E A C

  ^     
( a) Fig. 2.5.4. (b )

Que 2.6. Insert the nodes 15, 13, 12, 16, 19, 23, 5, 8 in empty
red-black tree and delete in the reverse order of insertion.
AKTU 2016-17, Marks 10

Answer
Insertion :
B
Insert 15 : 15
B
R 15
Insert 13 : 13
Insert 12 :
B
B
15 13
R
R 13 12 15
12 R R
Advanced Data Structure 2–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 16 :
B B
13 Case 3 13

12 15 12 15
R R B B 16
16
R R

Insert 19 :

B B
13 Case 2 13
B B B
12 15 12 16
R B R
16 15 19
R
19 R

Insert 23 :

B B
13 13
B B B R
12 16 12 16
R
15 19 15 19
R
R 23 B B 23
R
Insert 5 :

B
B 13
R
12 16
B
5 15 19
R B
23
R
Insert 8 :
B B
13 13
B R
B
B 12 16 8 16
R R
2nd 15 19 B
R 5 5 12 15 19
B B
23 B R
1st 8 23
R
R
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 2–13 B (CS/IT-Sem-5)

Deletions :
Delete 8 :
B B
13 13
B R B R
8 16 12 16
R
5 12 15 19 B 5 R B 15 19 B
R B
R 23 23 R
Delete 5 :
B
B
13 13
R R B
R
12 16 12 16
R B B
5 15 19 15 19 B
B
23 R 23 R
Delete 23 :
B B
13 13
B 12 R B 12 R
16 16

B 15 19 B B 15 19 B

23 R
Delete 19 :
B B B
13 13 13
R Case 2 B
B 12 16 B 12 16 12 16 B
R
B 15 19 B B 15 15 R
Delete 16 :
B B
13 13
B B
B 12 16 B 12 15

15 R
Delete 12 :
B B
13 13
B 12 15 B 15 B
Delete 13 :
B B
13 15
15 B
Advanced Data Structure 2–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Delete 15 :
No tree

Que 2.7. Insert the following element in an initially empty

RB-Tree.
12, 9, 81, 76, 23, 43, 65, 88, 76, 32, 54. Now delete 23 and 81.
AKTU 2019-20, Marks 07

Answer
B
Insert 12 : 12

B
Insert 9 : 12

R
9
B
Insert 81 : 12
R
R
9 81
B B
12 12
Insert 76 : R B
R Case 1 B
9 81 9 81
R
76 76 R
Insert 23 : B B
12 12
B B
B Case 3 B
9 81 9 76
R
76 23 81
R R R
23

Insert 43 : B B
12 12
B B
B Case 1 R
9 76 9 76
R R B B
23 81 23 81

R 43 43 R
 More pdf : www.motivationbank.in
2–15 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Insert 65 : B B
12 12
B B
R Case 3 R
9 76 9 76
R B B
23 81 B 43 81
R
R
43 23 65 R

R
65

Insert 88 : B
12
B
R
9 76
B
43 81 B
R
R R
23 65 88

Insert 76 : B
12
B
R
9 76 Case 2
B B
43 81
R
R R
23 65 76 R 88

76 B

R
12 81 B

B 9 B 43 76 R 88 R

R 23 R 65
Advanced Data Structure 2–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 32 : B
76
R B
12 81
B
B R
9 43 76 R 88

R
23 R 65

R
32

76 B

R
12 81 B

B 9 R 43 76 R 88 R

B 23 B 65

32 R

Insert 54 : B
76

R 12 81 B

R R
B 9 R 43 76 88

B 23 65 B

R R
32 54
 More pdf : www.motivationbank.in
2–17 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Delete 23 : B B
76 76

R 12 81 B R 12 81 B

R R R R
B 9 R 43 76 88 B 9 R 43 76 88

B 23 65 B B 32 65 B

R R R
32 54 54

Delete 81 : B B
76 76

R 12 81 B R 12 88 B

R R R
B 9 R 43 76 88 B 9 R 43 76

B 32 65 B B 32 65 B

R R
54 54

Que 2.8. Describe the properties of red-black tree. Show the

red-black tree with n internal nodes has height at most 2 log (n + 1).
OR
Prove the height h of a red-black tree with n internal nodes is not
greater than 2 log (n + 1).

Answer
Properties of red-black tree : Refer Q. 2.1, Page 2–2B, Unit-2.
1. By property 5 of RB-tree, every root-to-leaf path in the tree has the
same number of black nodes, let this number be B.
2. So there are no leaves in this tree at depth less than B, which means the
tree has at least as many internal nodes as a complete binary tree of
height B.
3. Therefore, n  2B – 1. This implies B  log (n + 1).
4. By property 4 of RB-tree, at most every other node on a root-to-leaf
path is red, therefore, h  2B.
Putting these together, we have
h  2 log (n + 1).
Advanced Data Structure 2–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 2.9. Insert the elements 8, 20, 11, 14, 9, 4, 12 in a Red-Black
tree and delete 12, 4, 9, 14 respectively.
AKTU 2018-19, Marks 10

Answer
Insert 8 :
B
8
Insert 20 :
B
8 R
20
Insert 11 : Since, parent of node 11 is red. Check the colour of uncle of
node 11. Since uncle of node 11 is nil than do rotation and recolouring.
B B
8 8 Left B
R rotation 11
20 R R
R 11 R 8 20
11
R
20
Insert 14 : Uncle of node 14 is red. Recolour the parent of node 14 i.e.,
20 and uncle of node 14 i.e., 8. No rotation is required.
B B
11 11
B R B B
8 20 8 20

R 14 R 14
Insert 9 : Parent of node 9 is black. So no rotation and no recolouring.
B B
11 11
B B B B
8 20 8 20
R R
R 9 14 R 9 14
Insert 4 : Parent of node 4 is black. So no rotation and no recolouring.
B B
11 11
B B B B
8 20 8 20
R R
R 4 R 9 14 R 4 R 9 14
 More pdf : www.motivationbank.in
2–19 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Insert 12 : Parent of node 12 is red. Check the colour of uncle of node

12, which is nil. So do rotation and recolouring.

B B
11 11
B Right B
B B
8 20 rotation 8 14
R
R 4 R 9 14 R 4 R 9 12 R 20 R

12 R

Delete 12 : Node 12 is red and leaf node. So simply delete node 12.
B B
11 11
B B B B
8 14 8 14

R 4 R 9 12 R 20 R R 4 R 9 20 R
Delete 4 : Node 4 is red and leaf node. So simply delete node 4.
B B
11 11
B B B B
8 14 8 14

R 4 9 R 20 R R 9 20 R

Delete 9 : Node 9 is red and leaf node. So simply delete node 9.

B B
11 11
B B B B
8 14 Case 1 8 14

9 R 20 R 20 R
Delete 14 : Node 14 is internal node replace node 14 with node 20 and do
not change the colour.
B B
11 11
B B B B
8 14 8 20
R
20

PART-2
B-Trees.
Advanced Data Structure 2–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 2.10. Define a B-tree of order m. Explain the searching and

insertion algorithm in a B-tree.

Answer
A B-tree of order m is an m-ary search tree with the following properties :
1. The root is either leaf or has atleast two children.
2. Each node, except for the root and the leaves, has between m/2 and m
children.
3. Each path from the root to a leaf has the same length.
4. The root, each internal node and each leaf is typically a disk block.
5. Each internal node has upto (m – 1) key values and upto m children.
SEARCH(x, k)
1. i  1
2. while i  n[x] and k > keyi[x]
3. do i  i + 1
4. if i  n[x] and k = keyi[x]
5. then return(x, i)
6. if leaf[x]
7. then return NIL
8. else DISK-READ(ci[x])
9. return B-TREE-SEARCH (ci[x], k)
B-TREE-INSERT(T, k)
1. r  root[T]
2. if n[r] = 2t – 1
3. then s  ALLOCATE-NODE ( )
4. root[T]  S
5. leaf[s]  FALSE
6. n[s]  0
7. c1[s]  r
8. B-TREE SPLIT CHILD(S, l, r)
9. B-TREE-INSERT-NONFULL(s, k)
10. else B-TREE-INSERT-NONFULL(r, k)
B-TREE SPLIT CHILD(x, i, y)
1. z  ALLOCATE-NODE ( )
2. leaf[z]  leaf[y]
3. n[z]  t – 1
 More pdf : www.motivationbank.in
2–21 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

4. for j  1 to t – 1
5. do keyj[z] ← keyj+t[y]
6. if not leaf[y]
7. then for j  1 to t
8. do cj[z]  cj+t [y]
9. n[y]  t – 1
10. for j  n[x] + 1 down to i + 1
11. do cj+1[x]  cj [x]
12. ci+1[x]  z
13. for j  n[x] down to i
14. do keyj+1[x]  keyj[x]
15. keyi[x]  keyt[y]
16. n[x]  n[x] + 1
17. DISK-WRITE[y]
18. DISK-WRITE[z]
19. DISK-WRITE[x]
The CPU time used by B-TREE SPLIT CHILD is (t). The procedure performs
(1) disk operations.
B-TREE-INSERT-NONFULL(x, k)
1. i  n[x]
2. if leaf[x]
3. then while i  1 and k < keyi[x]
4. do keyi+1[x]  keyi[x]
5. i  i – 1
6. keyi+1[x]  k
7. n[x]  n[x] + 1
8. DISK-WRITE(x)
9. else while i  1 and k < keyi[x]
10. do i  i – 1
11. i  i + 1
12. DISK-READ(c i[x])
13. if n[ci[x]] = 2t – 1
14. then B-TREE-SPLIT-CHILD(x, i, ci[x])
15. if k > keyi[x]
16. then i  i + 1
17. B-TREE INSERT NONFULL(ci [x], k)
The total CPU time use is O(th) = O(t logt n)

Que 2.11. What are the characteristics of B-tree ? Write down the
steps for insertion operation in B-tree.
Advanced Data Structure 2–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer
Characteristic of B-tree :
1. Each node of the tree, except the root node and leaves has at least m/2
subtrees and no more than m subtrees.
2. Root of tree has at least two subtree unless it is a leaf node.
3. All leaves of the tree are at same level.
Insertion operation in B-tree :
In a B-tree, the new element must be added only at leaf node. The insertion
operation is performed as follows :
Step 1 : Check whether tree is empty.
Step 2 : If tree is empty, then create a new node with new key value and
insert into the tree as a root node.
Step 3 : If tree is not empty, then find a leaf node to which the new key value
can be added using binary search tree logic.
Step 4 : If that leaf node has an empty position, then add the new key value
to that leaf node by maintaining ascending order of key value within the
node.
Step 5 : If that leaf node is already full, then split that leaf node by sending
middle value to its parent node. Repeat the same until sending value is fixed
into a node.
Step 6 : If the splitting is occurring to the root node, then the middle value
becomes new root node for the tree and the height of the tree is increased by
one.

Que 2.12. Describe a method to delete an item from B-tree.

Answer
There are three possible cases for deletion in B-tree as follows :
Let k be the key to be deleted, x be the node containing the key.
Case 1 : If the key is already in a leaf node, and removing it does not cause
that leaf node to have too few keys, then simply remove the key to be
deleted. Key k is in node x and x is a leaf, simply delete k from x.
Case 2 : If key k is in node x and x is an internal node, there are three cases
to consider :
a. If the child y that precedes k in node x has at least t keys (more than the
minimum), then find the predecessor key k in the subtree rooted at y.
Recursively delete k and replace k with k in x.
b. Symmetrically, if the child z that follows k in node x has at least t keys,
find the successor k and delete and replace as before.
c. Otherwise, if both y and z have only t – 1 (minimum number) keys,
merge k and all of z into y, so that both k and the pointer to z are
removed from x, y now contains 2t – 1 keys, and subsequently k is
deleted.
Case 3 : If key k is not present in an internal node x, determine the root of
the appropriate subtree that must contain k. If the root has only t – 1 keys,
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms2–23 B (CS/IT-Sem-5)

execute either of the following two cases to ensure that we descend to a node
containing at least t keys. Finally, recurse to the appropriate child of x.
a. If the root has only t – 1 keys but has a sibling with t keys, give the root
an extra key by moving a key from x to the root, moving a key from the
roots immediate left or right sibling up into x, and moving the appropriate
child from the sibling to x.
b. If the root and all of its siblings have t – 1 keys, merge the root with one
sibling. This involves moving a key down from x into the new merged
node to become the median key for that node.
Que 2.13. How B-tree differs with other tree structures ?

Answer
1. In B-tree, the maximum number of child nodes a non-terminal node
can have is m where m is the order of the B-tree. On the other hand,
other tree can have at most two subtrees or child nodes.
2. B-tree is used when data is stored in disk whereas other tree is used
when data is stored in fast memory like RAM.
3. B-tree is employed in code indexing data structure in DBMS, while,
other tree is employed in code optimization, Huffman coding, etc.
4. The maximum height of a B-tree is log mn (m is the order of tree and
n is the number of nodes) and maximum height of other tree is log2 n
(base is 2 because it is for binary).
5. A binary tree is allowed to have zero nodes whereas any other tree
must have atleast one node. Thus binary tree is really a different kind
of object than any other tree.

Que 2.14. Insert the following key in a 2-3-4 B-tree :

40, 35, 22, 90, 12, 45, 58, 78, 67, 60 and then delete key 35 and 22 one after
other. AKTU 2018-19, Marks 07

Answer
In 2-3-4 B-trees, non-leaf node can have minimum 2 keys and maximum 4
keys so the order of tree is 5.
Insert 40, 35, 22, 90 :
22 35 40 90
Insert 12 :
35

12 22 40 90
Insert 45, 58 :
35

12 22 40 45 58 90
Advanced Data Structure 2–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 78 :
35 58

12 22 40 45 78 90
Insert 67, 60 :
35 58

12 22 40 45 60 67 78 90
Delete 35 :
40 58

12 22 45 60 67 78 90
Delete 22 :
40 58

12 25 60 67 78 90

Que 2.15. Explain B-tree and insert elements B, Q, L, F into

B-tree Fig. 2.15.1 then apply deletion of elements F, M, G, D, B on
resulting B-tree.
G M P X

A C D E I K N O R S T U V Y Z

Fig. 2.15.1.

AKTU 2015-16, Marks 10

Answer
B-tree : Refer Q. 2.10, Page 2–20B, Unit-2.
Numerical :
Insertion :
G M P X

A C D E I K N O R S T U V Y Z
Assuming, order of B-tree = 5

G M P T X

A C D E I K N O R S U V Y Z
 More pdf : www.motivationbank.in
2–25 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

G M T X

A C D E I K N O R S U V Y Z

Insert B :

G M T X

A B C D E I K N O R S U V Y Z

C G M T X

A B D E I K N O R S U V Y Z

Insert Q :

C G M T X

A B D E I K N O Q R S U V Y Z

Insert L :
P

C G M T X

A B D E I K L N O Q R S U V Y Z
Advanced Data Structure 2–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert F :
P

C G M T X

A B D E F I K L N O Q R S U V Y Z
Deletion :
Delete F :

C G M T X

A B D E I K L N O Q R S U V Y Z
Delete M :

C G L T X

A B D E I K N O Q R S U V Y Z
Delete G :
P

C L T X

A B D E I K N O Q R S U V Y Z
Delete D :

C L P T X

A B E I K N O Q R S U V Y Z
Delete B :
E L P T X

A C I K NO Q R S UV Y Z
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms2–27 B (CS/IT-Sem-5)

Que 2.16. Insert the following information, F, S, Q, K, C, L, H, T, V,

W, M, R, N, P, A, B, X, Y, D, Z, E, G, I into an empty B-tree with degree
t = 3. AKTU 2017-18, Marks 10

Answer
Assume that t=3
2t – 1 = 2 × 3 – 1 = 6 – 1 = 5
and t–1= 3–1=2
So, maximum of 5 keys and minimum of 2 keys can be inserted in a node.
Now, apply insertion process as :
Insert F : F

Insert S : F S

Insert Q : F Q S

Insert K : F K Q S

Insert C : C F K Q S

Insert L : C F K L Q S
As, there are more than 5 keys in this node.
 Find median, n[x] = 6 (even)
n[ x ] 6
Median = = 3
2 2
Now, median = 3,
So, we split the node by 3rd key.

K
C F K L Q S
Median
(splitting point) C F L Q S
Move up
Insert H, T :
K

C F H L Q S T
Insert V :
K

C F H L Q S T V
Advanced Data Structure 2–28 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert W :
K

C F H L Q S T V W
More than 5 keys split node from Median.
n[x] = 6 [even]
n[ x] 6
Median =  3 (i.e., 3rd key move up)
2 2
K S

C F H L Q T V W
Insert M, R, N :
K S

CF H L MNQR T V W
Insert P :
K S

C F H LM N P QR T V W
More than 5 key
split the node
n[x] = 6
n[ x] 6
Median =   3 (i.e., 3rd key move up)
2 2

K N S

C F H LM P QR T V W
Insert A, B :

KN S

A B CF H LM PQ R T V W
 More pdf : www.motivationbank.in
2–29 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Insert X, Y :
K N S

A B CF H LM PQR TV WXY
Insert D :
K N S

A B CD FH LM PQ R TV WXY
More than 5 key
split the node
n[x] = 6 (even)
n[ x] 6
Median =  3 (i.e., 3rd key move up)
2 2

C K NS

AB D FH LM PQR TV WXY
Insert Z, E :

CK NS

A B DEFH LM PQR TVWXYZ

More than 5 key
split the node
n[x] = 6
n[ x] 6
Median =  3 (i.e., 3rd key move up)
2 2
CK N SW

A B D E FH L M PQR TV X YZ
Insert G, I :
N

CK SW

A B DEFG HI LM PQR TV X YZ
Advanced Data Structure 2–30 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
N

CF K SW

A B DE GHI LM PQR TV X YZ
Fig. 2.16.1. Inserted all given information with degree t = 3.

Que 2.17. Using minimum degree ‘t’ as 3, insert following sequence

of integers 10, 25, 20, 35, 30, 55, 40, 45, 50, 55, 60, 75, 70, 65, 80, 85 and 90
in an initially empty B-Tree. Give the number of nodes splitting
operations that take place. AKTU 2019-20, Marks 07

Answer
Insert 10, 25, 20, 35 :
20

Split
10 20 25 35  10 25 35
Insert 30 :

20

10 25 30 35
Insert 55 :
20 20 30

Split
10 25 30 35 55  10 25 35 55
Insert 40, 45 :

20 30 20 30 40

Split
10 25 35 40 45 55  10 25 35 45 55
 More pdf : www.motivationbank.in
2–31 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Insert 50, 55 :

20 30 40 20 30 40 50
Split

10 25 35 45 50 55 55 10 25 35 45 55 55


Split
30

20 40 50

10 25 35 45 55 55

Insert 60, 75 :

30

20 40 50 Split

10 25 35 45 55 55 60 75

30

20 40 50 55

10 25 35 45 55 60 75
Advanced Data Structure 2–32 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 70, 65 :
30

20 40 50 55 Split


10 25 35 45 55 60 65 70 75

30

40 50 55 65 Split
20


10 25 35 45 55 60 70 75

30 50

20 40 55 65

10 25 35 45 55 60 70 75
Insert 80, 85 :
30 50

20 40 55 65
Split

10 25 35 45 55 60 70 75 80 85

30 50

20 40 55 65 75

10 25 35 45 55 60 70 80 85
Insert 90 :
30 50

20 40 55 65 75

10 25 35 45 55 60 70 80 85 90
 More pdf : www.motivationbank.in
2–33 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Number of nodes splitting operations = 9.

PART-3
Binomial Heaps.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 2.18. Explain binomial heap and properties of binomial tree.

Answer
Binomial heap :
1. Binomial heap is a type of data structure which keeps data sorted and
allows insertion and deletion in amortized time.
2. A binomial heap is implemented as a collection of binomial tree.
Properties of binomial tree :
1. The total number of nodes at order k are 2k.
2. The height of the tree is k.
 k
3. There are exactly   i.e., kCi nodes at depth i for i = 0, 1, …. , k (this is
 i
why the tree is called a “binomial” tree).
4. Root has degree k (children) and its children are Bk-1, Bk-2, …, B0 from
left to right.

Que 2.19. What is a binomial heap ? Describe the union of

binomial heap.
OR
Explain the different conditions of getting union of two existing
binomial heaps. Also write algorithm for union of two binomial
heaps. What is its complexity ? AKTU 2018-19, Marks 10

Answer
Binomial heap : Refer Q. 2.18, Page 2–33B, Unit-2.
Union of binomial heap :
1. The BINOMIAL-HEAP-UNION procedure repeatedly links binomial
trees where roots have the same degree.
2. The following procedure links the Bk-1 tree rooted at node to the
Bk-1 tree rooted at node z, that is, it makes z the parent of y. Node z thus
becomes the root of a Bk tree.
Advanced Data Structure 2–34 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
BINOMIAL-LINK (y, z)
i p[y] ← z
ii. sibling [y]  child[z]
iii. child[z]  y
iv. degree[z]  degree[z] + 1
3. The BINOMIAL-HEAP-UNION procedure has two phases :
a. The first phase, performed by the call of BINOMIAL-HEAP-
MERGE, merges the root lists of binomial heaps H1 and H2 into a
single linked list H that is sorted by degree into monotonically
increasing order.
b. The second phase links root of equal degree until at most one root
remains of each degree. Because the linked list H is sorted by
degree, we can perform all the like operations quickly.
BINOMIAL-HEAP-UNION(H1, H2)
1. H  MAKE-BINOMIAL-HEAP ( )
2. head[H] ← BINOMIAL-HEAP-MERGE(H1, H2)
3. Free the objects H1 and H2 but not the lists they point to
4. if head[H] = NIL
5. then return H
6. prev-x  NIL
7. x  head[H]
8. next-x  sibling[x]
9. while next-x  NIL
10. do if (degree[x]  degree[next-x]) or
(sibling[next-x]  NIL and degree[sibling[next-x]] = degree[x])
11. then prev-x  x  case 1 and 2
12. x  next-x  case 1 and 2
13. else if key[x]  key[next-x]
14. then sibling[x]  sibling[next-x]  case 3
15. BINOMIAL-LINK(next-x, x)  case 3
16. else if prev-x = NIL  case 4
17. then head[H]  next-x  case 4
18. else sibling[prev-x]  next-x  case 4
19. BINOMIAL-LINK(x, next-x)  case 4
20. x  next-x  case 4
21. next-x  sibling[x]
22. return H
BINOMIAL-HEAP-MERGE(H1, H2)
1. a  head[H1]
2. b  head[H2]
3. head[H1]  min-degree (a, b)
4. if head[H1] = NIL
5. return
6. if head[H1] = b
7. then b  a
8. a  head[H1]
9. while b  NIL
 More pdf : www.motivationbank.in
2–35 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

10. do if sibling[a] = NIL

11. then sibling[a]  b
12. return
13. else if degree [sibling[a]] < degree[b]
14. then a  sibling[a]
15. else c  sibling[b]
16. sibling[b]  sibling[a]
17. sibling[a]  b
18. a  sibling[a]
19. b  c
There are four cases/conditions that occur while performing union
on binomial heaps.
Case 1 : When degree[x]  degree[next-x] = degree [sibling[next-x]], then
pointers moves one position further down the root list.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a b c d

Bk BL Case1 Bk BL
(a) (b )
Fig. 2.19.1.
Case 2 : It occurs when x is the first of three roots of equal degree, that is,
degree[x] = degree[next-x] = degree[sibling[next-x]], then again pointer
move one position further down the list, and next iteration executes either
case 3 or case 4.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a b c d

Bk Bk Bk Case 2 Bk Bk Bk
(a) (b )
Fig. 2.19.2.
Case 3 : If degree[x] = degree[next-x]  degree [sibling[next-x]] and key[x]
 key[next-x], we remove next-x from the root list and link it to x, creating
Bk+1 tree.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a b d

Bk Bk BL Case 3
Bk BL
c
key [x]  key [next-x]
(a) Bk
Bk+1
Fig. 2.19.3. (b )
Advanced Data Structure 2–36 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Case 4 : degree[x] = degree[next-x]  degree[sibling[next-x] and key[next-x]
 key x, we remove x from the root list and link it to next-x, again creating
a Bk+1 tree.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a c d

Bk Bk BL Case 4
b Bk BL
key [x]  key [next-x]
Bk
(a) Bk+1
Fig. 2.19.4. (b )
Time complexity of union of two binomial heap is O(log n).

Que 2.20. Explain properties of binomial heap. Write an algorithm

to perform uniting two binomial heaps. And also to find Minimum
key. AKTU 2017-18, Marks 10

Answer
Properties of binomial heap : Refer Q. 2.18, Page 2–33B, Unit-2.
Algorithm for union of binomial heap : Refer Q. 2.19, Page 2–33B,
Unit-2.
Minimum key :
BINOMIAL-HEAP-EXTRACT-MIN (H) :
1. Find the root x with the minimum key in the root list of H, and remove
x from the root list of H.
2. H'  MAKE-BINOMIAL-HEAP( ).
3. Reverse the order of the linked list of x’s children, and set head[H] to
point to the head of the resulting list.
4. H  BINOMIAL-HEAP-UNION(H, H’).
5. Return x
Since each of lines 1-4 takes O(log n) time of H has n nodes, BINOMIAL-
HEAP-EXTRACT-MIN runs in O(log n) time.

Que 2.21. Construct the binomial heap for the following sequence
of number 7, 2, 4, 17, 1, 11, 6, 8, 15.

Answer
Numerical :
Insert 7 :
Head [H]
7
 More pdf : www.motivationbank.in
2–37 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Insert 2 :
x next-x
7 2
Head [H]
prev-x = NIL
degree [x] = 0. So, degree [x]  degree [next-x] is false.
degree [next-x] = 0 and Sibling [next-x] = NIL
So, case 1 and 2 are false here.
Now key [x] = 7 and key [next-x] = 2
Now prev-x = NIL
then Head [H]  next-x and
i.e.,
Head [H]
7 2
and BINOMIAL-LINK (x, next-x)
i.e.,
2

7
Now
Head [H] 2 x

and next-x = NIL

7
So, after inserting 2, binomial heap is
Head [H] 2

7
Insert 4 :
x next-x
4 2

7
degree [x]  degree [next-x]
So, Now next-x makes x and x makes prev-x.
Now next-x = NIL
So, after inserting 4, final binomial heap is :
Advanced Data Structure 2–38 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Head [H] 4 2

7
Insert 17 :
After Binomial-Heap-Merge, we get
x next-x
4 17 2

7
degree [x] = degree [next-x]
degree [Sibling-[next-x]]  degree [x]
key [x]  key [next-x]
4  17 [True]
So,

4 17 2
and call Binomial-Link [next-x, x]

7
We get
x next-x
4 2

7
17
degree [x] = degree [next-x]
Sibling [next-x] = NIL
Key [x]  key [next-x] [False]
prev-x = NIL then
Head [H]  [next-x]
Binomial-Link [x, next-x]
x  next-x
Head [H]
2 x

4
7
next-x = NIL

17
 More pdf : www.motivationbank.in
2–39 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

So, after inserting 17, final binomial heap is :

Head [H] 2 x

4
7

17

Insert 1 :
x next-x
Head [H] 1 2

4
7

17
degree [x]  degree [next-x]
So, next-x makes x and next-x = NIL
and after inserting 1, binomial heap is :
Head [H] 1 2

4
7

17
Insert 11 :
After Binomial-Heap-Merge, we get
x next-x
Head [H] 1 11 2

4
7

17
degree [x] = degree [next-x]
degree [Sibling [next-x]]  degree [x]
key [x]  key [next-x] [True]
Advanced Data Structure 2–40 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
So,
x next-x
1 2

4 7
11

17
degree [x]  degree [next-x]
So, next-x makes x and next-x = NIL
and final binomial heap after inserting 11 is
Head [H] 1 2

4 7
11

17
Insert 6 :
x next-x
Head [H] 6 1 2

4 7
11

17
degree [x]  degree [next-x]
So, next-x becomes x
Sibling [next-x] becomes next-x.
i.e.,
x next-x
6 1 2

4 7
11

17
 More pdf : www.motivationbank.in
2–41 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

degree [x]  degree [next-x]

So, no change and final heap is :
Head [H] 6 1 2

4 7
11

17
Insert 8 :
x next-x
6 8 1 2

4 7
11

17
degree [x] = degree [next-x]
degree [Sibling [next-x]]  degree [x]
key [x]  key [next-x] [True]
So,
x next-x
6 1 2

8 11 4 7

17
degree [x] = degree [next-x]
degree [Sibling p[next-x]]  degree [x]
key [x]  key [next-x] [False]
prev-x = NIL
So,
x next-x
1 2

6 4
11 7

8
17
Advanced Data Structure 2–42 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
degree [x] = degree [next-x]
Sibling [next-x] = NIL
key [x]  key [next-x] [True]
So, Sibling [x] = NIL.
and

Head [H]
1

2
6 11

4 7
8

17
next [x] = NIL
So, this is the final binomial heap after inserting 8.
Insert 15 :
x next-x
Head [H] 15 1

2
6 11

8 4 7

17
degree [x]  degree [next-x]
So, no change and this is the final binomial heap after inserting 15.

Que 2.22. Explain the algorithm to delete a given element in a

binomial heap. Give an example for the same.
AKTU 2019-20, Marks 07

Answer
Deletion of key from binomial heap :
The operation BINOMIAL-HEAP-DECREASE (H, x, k) assigns a new key ‘k’
to a node ‘x’ in a binomial heap H.
BINOMIAL-HEAP-DECREASE-KEY (H, x, k)
1. if k > key [x] then
2. Message “error new key is greater than current key”
3. key [x]  k
 More pdf : www.motivationbank.in
2–43 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

4. y  x
5. z  P [y]
6. While (z  NIL) and key [y] < key [z]
7. do exchange key [y]  key [z]
9. y  z
10. z  P [y]
Deleting a key : The o pe ration BIN OMIAL-HEAP-DELETE
(H, x) is used to delete a node x’s key from the given binomial heap H. The
following implementation assumes that no node currently in the binomial
heap has a key of – .
BINOMIAL-HEAP-DELETE (H, x)
1. BINOMIAL-HEAP-DECREASE-KEY (H, x, – )
2. BINOMIAL-HEAP-EXTRACT-MIN(H)
For example : Operation of Binomial-Heap-Decrease (H, x, k) on the
following given binomial heap :
Suppose a binomial heap H is as follows :

Head[H] 37 10 1

41 28 13 6 16 12 25

77 8 14 29 26 23 18

11 17 38 42

27
The root x with minimum key is 1. x is removed from the root list of H. i.e.,
x
Head[H] 37 10 1

41 28 13 6 16 12 25

77 8 14 29 26 23 18

11 17 38 42

27
Now, the linked list of x’s children is reversed and set head[H] to point to the
head of the resulting list, i.e., another binomial heap H.
Advanced Data Structure 2–44 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Head[H ] 25 12 16 6

18 26 23 8 14 29

42 11 17 38

27

Now, call BINOMIAL-HEAP-UNION (H, H) to uniting the two binomial

heaps H and H. The resulting binomial heap is
Head[H]
25 12 6

37 18 10 8 14 29

41 16 28 13 11 17 38

26 23 77 27

42

PART-4
Fibonacci Heaps.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 2.23. What is a Fibonacci heap ? Discuss the applications of

Fibonacci heaps.

Answer
1. A Fibonacci heap is a set of min-heap-ordered trees.
2. Trees are not ordered binomial trees, because
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 2–45 B (CS/IT-Sem-5)

a. Children of a node are unordered.

b. Deleting nodes may destroy binomial construction.
min(H)

12 7 4 8

14 15 9 10 11 6

Fig. 2.23.1.
20
3. Fibonacci heap H is accessed by a pointer min[H] to the root of a tree
containing a minimum key. This node is called the minimum node.
4. If Fibonacci heap H is empty, then min[H] = NIL.
Applications of Fibonacci heap :
1. Fibonacci heap is used for Dijkstra’s algorithm because it improves the
asymptotic running time of this algorithm.
2. It is used in finding the shortest path. These algorithms run in O(n2)
time if the storage for nodes is maintained as a linear array.

Que 2.24. What is Fibonacci heap ? Explain CONSOLIDATE

operation with suitable example for Fibonacci heap.
AKTU 2015-16, Marks 15

Answer
Fibonacci heap : Refer Q. 2.23, Page 2–44B, Unit-2.
CONSOLIDATE operation :
CONSOLIDATE(H)
1. for i  0 to D(n[H])
2. do A[i]  NIL
3. for each node w in the root list of H
4. do x  w
5. d  degree[x]
6. while A[d]  NIL
7. do y  A[d]  Another node with the same degree as x.
8. if key[x] > key[y]
9. then exchange x  y
10. FIB-HEAP-LINK(H, y, x)
11. A[d]  NIL
12. d  d + 1
13. A[d]  x
14. min[H]  NIL
15. for i  0 to D(n[H])
16. do if A[i]  NIL
17. then add A[i] to the root list of H
Advanced Data Structure 2–46 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
18. if min[H] = NIL or key [A[i]] < key[min[H]]
19. then min[H]  A[i]
FIB-HEAP-LINK(H, y, x)
1. remove y from the root list of H
2. make y a child of x, incrementing degree[x]
3. mark[y]  FALSE

Que 2.25. Define Fibonacci heap. Discuss the structure of a

Fibonacci heap with the help of a diagram. Write a function for
uniting two Fibonacci heaps.

Answer
Fibonacci heap : Refer Q. 2.23, Page 2–44B, Unit-2.
Structure of Fibonacci heap :
1. Node structure :
a. The field “mark” is True if the node has lost a child since the node
became a child of another node.
b. The field “degree” contains the number of children of this node.
The structure contains a doubly-linked list of sibling nodes.
Parent
Key
Mark
Degree
8 Left Right

Child

Fig. 2.25.1. Node structure.

2. Heap structure :
min(H) : Fibonacci heap H is accessed by a pointer min[H] to the root of
a tree containing a minimum key; this node is called the minimum node.
If Fibonacci heap H is empty, then min[H] = NIL.
n(H) : Number of nodes in heap H
min(H)

12 7 4

14 15 9 10 11 6

20
Fig. 2.25.2. Heap structure.
Function for uniting two Fibonacci heap :
Make-Heap :
MAKE-FIB-HEAP( )
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms2–47 B (CS/IT-Sem-5)

allocate(H)
min(H) = NIL
n(H) = 0
FIB-HEAP-UNION(H1, H2)
1. H  MAKE-FIB-HEAP( )
2. min[H]  min[H1]
3. Concatenate the root list of H2 with the root list of H
4. if (min[H1] = NIL) or (min[H2]  NIL and min[H2] < min[H1])
5. then min[H]  min[H2]
6. n[H]  n[H1] + n[H2]
7. Free the objects H1 and H2
8. return H

Que 2.26. Discuss following operations of Fibonacci heap :

i. Make-Heap
ii. Insert
iii. Minimum
iv. Extract-Min

Answer
i. Make-Heap : Refer Q. 2.25, Page 2–46B, Unit-2.
ii. Insert : (H, x)
1. degree[x]  0
2. p[x]  NIL
3. child[x]  NIL
4. left[x]  x
5. right[x]  x
6. mark[x]  FALSE
7. concatenate the root list containing x with root list H
8. if min[H] = NIL or key[x] < key[min[H]]
9. then min[H]  x
10. n[H]  n[H] + 1
To determine the amortized cost of FIB-HEAP-INSERT, Let H be the
input Fibonacci heap and H’ be the resulting Fibonacci heap, then
t(H) = t(H) + 1 and m(H’) = m(H), and the increase in potential is,
(t(H) + 1) + 2m(H) – (t(H) + 2m(H)) = 1
Since the actual cost is O(1), the amortized cost is O(1) + 1 = O(1)
iii. Minimum :
The minimum node of a Fibonacci heap H is always the root node given
by the pointer min[H], so we can find the minimum node in O(1) actual
time. Because the potential of H does not change, the amortized cost of
this operation is equal to its O(1) actual cost.
iv. FIB-HEAP-EXTRACT-MIN(H)
1. z  min[H]
2. if z  NIL
3. then for each child x of z
Advanced Data Structure 2–48 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
4. do add x to the root list of H
5. p[x]  NIL
6. remove z from the root list of H
7. if z = right[z]
8. then min[H]  NIL
9. else min[H]  right[z]
10. CONSOLIDATE (H)
11. n[H]  n[H] – 1
12. return z

PART-5
Tries, Skip List.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 2.27. What is trie ? What are the properties of trie ?

Answer
1. A trie (digital tree / radix tree / prefix free) is a kind of search tree i.e., an
ordered tree data structure that is used to store a dynamic set or
associative array where the keys are usually strings.
2. Unlike a binary search tree, no node in the tree stores the key associated
with that node; instead, its position in the tree defines the key with
which it is associated.
3. All the descendants of a node have a common prefix of the string
associated with that node, and the root is associated with the empty
string.
4. Values are not necessarily associated with every node. Rather, values
tend only to be associated with leaves, and with some inner nodes that
correspond to keys of interest.
Properties of a trie :
1. Trie is a multi-way tree.
2. Each node has from 1 to d children.
3. Each edge of the tree is labeled with a character.
4. Each leaf node corresponds to the stored string, which is a concatenation
of characters on a path from the root to this node.

Que 2.28. Write an algorithm to search and insert a key in trie

data structure.
 More pdf : www.motivationbank.in
2–49 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Answer
Search a key in trie :
Trie-Search(t, P[k..m]) // inserts string P into t
1. if t is leaf then return true
2. else if t.child(P[k]) = nil then return false
3. else return Trie-Search(t.child(P[k]), P[k + 1..m])
Insert a key in trie :
Trie-Insert(t, P[k..m])
1. if t is not leaf then //otherwise P is already present
2. if t.child(P[k]) = nil then
//Create a new child of t and a “branch” starting with that child and
storing P[k..m]
3. else Trie-Insert(t.child(P[k]), P[k + 1..m])

Que 2.29. What is skip list ? What are its properties ?

Answer
1. A skip list is built in layers.
2. The bottom layer is an ordinary ordered linked list.
3. Each higher layer acts as an “express lane”, where an element in layer
i appears in layer (i + 1) with some fixed probability p (two commonly
used values for p are ½ and ¼.).
4. On average, each element appears in 1/(1– p) lists, and the tallest element
(usually a special head element at the front of the skip list) in all the lists.
5. The skip list contains log1/pn (i.e., logarithm base 1/p of n).
Properties of skip list :
1. Some elements, in addition to pointing to the next element, also point to
elements even further down the list.
2. A level k element is a list element that has k forward pointers.
3. The first pointer points to the next element in the list, the second pointer
points to the next level 2 element, and in general, the ith pointer points
to the next level i element.

Que 2.30. Explain insertion, searching and deletion operation in

skip list.

Answer
Insertion in skip list :
1. We will start from highest level in the list and compare key of next node
of the current node with the key to be inserted.
Advanced Data Structure 2–50 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. If key of next node is less than key to be inserted then we keep on
moving forward on the same level.
3. If key of next node is greater than the key to be inserted then we store
the pointer to current node i at update[i] and move one level down and
continue our search.
At the level 0, we will definitely find a position to insert given key.
Insert(list, searchKey)
1. local update[0...MaxLevel+1]
2. x : = list  header
3. for i : = list  level down to 0 do
4. while x  forward[i]  key forward[i]
5. update[i] : = x
6. x : = x  forward[0]
7. lvl : = randomLevel()
8. if lvl > list  level then
9. for i : = list  level + 1 to lvl do
10. update[i] : = list  header
11. list  level : = lvl
12. x : = makeNode(lvl, searchKey, value)
13. for i : = 0 to level do
14. x  forward[i] : = update[i]  forward[i]
15. update[i]  forward[i] : = x
Searching in skip list :
Search(list, searchKey)
1. x : = list  header
2. loop invariant : x  key level down to 0 do
3. while x  forward[i]  key forward[i]
4. x : = x  forward[0]
5. if x  key = searchKey then return x  value
6. else return failure
Deletion in skip list :
Delete(list, searchKey)
1. local update[0..MaxLevel+1]
2. x : = list  header
3. for i : = list  level down to 0 do
4. while x  forward[i]  key forward[i]
5. update[i] : = x
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms2–51 B (CS/IT-Sem-5)

6. x : = x  forward[0]
7. if x  key = searchKey then
8. for i : = 0 to list  level do
9. if update[i]  forward[i]  x then break
10. update[i]  forward[i] : = x  forward[i]
11. free(x)
12. while list  level > 0 and list  header  forward[list  level] = NIL do
13. list  level : = list  level – 1

Que 2.31. Given an integer x and a positive number n, use divide

and conquer approach to write a function that computes xn with
time complexity O(log n). AKTU 2018-19, Marks 10

Answer
Function to calculate xn with time complexity O(log n) :
int power(int x, unsigned int y)
{
int temp;
if(y == 0)
return 1;
temp = power(x, y/2);
if (y%2 == 0)
return temp * temp;
else
return x * temp * temp;
}

VERY IMPORTANT QUESTIONS

Following questions are very important. These questions
may be asked in your SESSIONALS as well as
UNIVERSITY EXAMINATION.

Q. 1. Define red-black tree and give its properties.

Ans. Refer Q. 2.1.

Q. 2. Insert the following element in an initially empty RB-Tree.

12, 9, 81, 76, 23, 43, 65, 88, 76, 32, 54. Now delete 23 and 81.
Ans. Refer Q. 2.7.
Advanced Data Structure 2–52 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Q. 3. Define a B-tree of order m. Explain the searching and
insertion algorithm in a B-tree.
Ans. Refer Q. 2.10.

Q. 4 Explain the insertion and deletion algorithm in a red-black

tree.
Ans. Insertion algorithm : Refer Q. 2.1.
Deletion algorithm : Refer Q. 2.5.

Q. 5. Using minimum degree ‘t’ as 3, insert following sequence of

integers 10, 25, 20, 35, 30, 55, 40, 45, 50, 55, 60, 75, 70, 65, 80, 85
and 90 in an initially empty B-Tree. Give the number of
nodes splitting operations that take place.
Ans. Refer Q. 2.17.

Q. 6. What is binomial heap ? Describe the union of binomial

heap.
Ans. Refer Q. 2.19.

Q. 7. What is a Fibonacci heap ? Discuss the applications of

Fibonacci heaps.
Ans. Refer Q. 2.23.

Q. 8. What is trie ? Give the properties of trie.

Ans. Refer Q. 2.27.

Q. 9. Explain the algorithm to delete a given element in a binomial

heap. Give an example for the same.
Ans. Q. 2.22.

Q. 10. Explain skip list. Explain its operations.

Ans. Skip list : Refer Q. 2.29.
Operations : Refer Q. 2.30.


 More pdf : www.motivationbank.in
3–1 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

3 Graph Algorithms

CONTENTS
Part-1 : Divide and Conquer with ........................ 3–2B to 3–10B
Examples such as Sorting,
Matrix Multiplication, Convex
Hull, Searching

Part-2 : Greedy Methods with Examples .......... 3–11B to 3–15B

such as Optimal Reliability
Allocation

Part-3 : Knapsack .................................................. 3–16B to 3–23B

Part-4 : Minimum Spanning Trees-Prim’s ....... 3–23B to 3–33B

and Kruskal’s Algorithm Single

Part-5 : Source Shortest ...................................... 3–33B to 3–41B

Paths-Dijkstra’s
and Bellman-Ford Algorithm
 More pdf : www.motivationbank.in
Graph Algorithm 3–2 B (CS/IT-Sem-5)

PART-1
Divide and Conquer with Examples such as
Sorting, Matrix Multiplication, Convex Hull, Searching.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 3.1. Write short note on divide and conquer. Write algorithm
for merge sort and quick sort.

Answer
Divide and conquer :
1. Divide and conquer is an algorithm design paradigm based on multi-
branched recursion.
2. A divide-and-conquer algorithm works by recursively breaking down a
problem into two or more sub-problems of the same or related type,
until these become simple enough to be solved directly.
3. The solutions to the sub-problems are then combined to give a solution
to the original problem.
4. Divide and conquer technique can be divided into the following three
parts :
a. Divide : This involves dividing the problem into some sub-problem.
b. Conquer : Recursively calling sub-problem until it is solved.
c. Combine : This involves combination of solved sub-problem so
that we will get the solution of problem.
Merge Sort : Refer Q. 1.21, Page 1–20B, Unit-1.
Quick Sort : Refer Q. 1.19, Page 1–16B, Unit-1.

Que 3.2. What is matrix chain multiplication problem ? Describe

a solution for matrix chain multiplication problem.

Answer
1. Matrix chain multiplication (or Matrix Chain Ordering Problem, MCOP)
is an optimization problem that can be solved using dynamic
programming.
2. MCOP helps to find the most efficient way to multiply given matrices.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms3–3 B (CS/IT-Sem-5)

3. Solution for matrix chain multiplication problem is Strassen’s matrix

multiplication.
Strassen’s matrix multiplication :
1. It is an application of divide and conquer technique.
2. Suppose we wish to compute the product C = AB where each A, B and C
are n × n matrices.
3. Assuming that n is an exact power of 2. We divide each of A, B and C into
four n/2 × n/2 matrices.
Rewriting the equation C = AB as
 r s  a b  e g
 t u =  c d  f h
...(3.2.1)

4. For convenience, the sub-matrices of A are labelled alphabetical from

left to right, whereas those of B are labelled from top to bottom. So that
matrix multiplication is performed.
Equation (3.2.1) corresponds to the four equations :
r = ae + bf ...(3.2.2)
s = ag + bh ...(3.2.3)
t = ce + df ...(3.2.4)
u = cg + dh ...(3.2.5)
5. Each of these four equations specifies two multiplications of
n/2 × n/2 matrices and the addition of their n/2 × n/2 products.
6. Using these equations to define a straight-forward divide and conquer
strategy. We derive the following recurrence for the time T(n) to multiply
two n × n matrices :
T(n) = 8T (n/2) + (n2)
7. Unfortunately, this recurrence has the solution T(n) = (n3) and thus,
this method is no faster than the ordinary one.

Que 3.3. Describe in detail Strassen’s matrix multiplication

algorithms based on divide and conquer strategies with suitable
example.

Answer
Strassen’s matrix multiplication algorithm has four steps :
1. Divide the input matrices A and B into n/2 × n/2 sub-matrices.
2. Using (n 2 ) scalar additio ns and subtractio n compute 14
n/2 × n/2 matrices A1, B1, A2, B2, ...A7, B7.
3. Recursively compute the seven matrix products.
Pi = AiBi for i = 1, 2, 3, ..., 7
 More pdf : www.motivationbank.in
Graph Algorithm 3–4 B (CS/IT-Sem-5)

4. Compute the desired sub-matrices r, s, t, u of the result matrix C by

adding and/or subtracting various combinations of the Pi matrices using
only (n2) scalar additions and subtractions.
A A12   B11 B12   C11 C12 
Suppose, A =  11  B=   and C =  
 A21 A22   B21 B22  C21 C22 

 A11 A12   B11 B12   C11 C12 

   = 
 A21 A22   B21 B22  C21 C22 
In Strassen method first compute the 7 n/2 × n/2 matrices.
P, Q, R, S, T, U, V as follows :
P = (A11 + A22) (B11 + B22)
Q = (A21 + A22) B11
R = A11 (B12 – B22)
S = A22 (B21 – B11)
T = (A11 + A12) B22
U = (A21 – A11) (B11 + B12)
V = (A12 – A22) (B21 + B22)
Then the cij’s are computed using the formulas
c11 = P + S – T + V
c12 = R + T
c21 = Q + S
c22 = P + R – Q + U
As can be seen P, Q, R, S, T, U and V can be computed using 7 matrix
multiplication and 10 matrix addition or subtraction and cij’s require an
additional 8 addition or subtraction.
For example :
Suppose,
2 9   4 11  C11 C12 
A=   B= 8 7  C =  
5 6    C21 C22 
AB = C
 2 9   4 11  C11 C12 
 5 6   8 7  = C 
C22 
    21
So, A11 = 2, B11 = 4
A12 = 9, B12 = 11
A21 = 5, B21 = 8
A22 = 6, B22 = 7
Now calculate,
S1 = B12 – B22 = 5, S6 = B11 + B12 = 15
S2 = A11 + A12 = 11, S7 = A12 – A22 = 3
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 3–5 B (CS/IT-Sem-5)

S3 = A21 + A22 = 11, S8 = B21 + B22 = 15

S4 = B21 – B11 = 4, S9 = A12 – A22 = 3
S5 = A11 + A22 = 8, S10 = B11 + B12 = 15
Now, R= A11 × S1 = 10, P = S5 × S6 = 120
T= S2 × B22 = 77, U = S7 × S8 = 45
Q= S3 × B11 = 44, V = S9 × S10 = 45
S= A22 × S4 = 24
Now, C11 = P + S – T + V = 88 + 24 – 77 + 45 = 80
C12 = R + T = 10 + 77 = 87
C21 = Q + S = 44 + 24 = 68
C22 = P + R – Q + U = 88 + 10 – 44 + 45 = 99
 80 71 
Now matrix =  
 68 99 
Que 3.4. What do you mean by graphs ? Discuss various
representations of graphs.

Answer
1. A graph G consists of a set of vertices V and a set of edges E.
2. Graphs are important because any binary relation is a graph, so graphs
can be used to represent essentially any relationship.
1 2

3 4
Fig. 3.4.1.
Various representations of graphs :
1. Matrix representation : Matrices are commonly used to represent
graphs for computer processing. Advantage of representing the graph
in matrix lies in the fact that many results of matrix algebra can be
readily applied to study the structural properties of graph from an
algebraic point of view.
a. Adjacency matrix :
i. Representation of undirected graph :
The adjacency matrix of a graph G with n vertices and no
parallel edges is a n × n matrix A = [aij] whose elements are
given by
aij = 1, if there is an edge between ith and jth vertices
= 0, if there is no edge between them
ii. Representation of directed graph :
The adjacency matrix of a digraph D, with n vertices is the
matrix
 More pdf : www.motivationbank.in
Graph Algorithm 3–6 B (CS/IT-Sem-5)

A = [aij]n×n in which
aij = 1 if arc (vi, vj) is in D
= 0 otherwise
b. Incidence matrix :
i. Representation of undirected graph :
Consider an undirected graph G = (V, E) which has n vertices
and m edges all labelled. The incidence matrix I(G) = [bij], is
then n × m matrix, where
bij = 1 when edge ej is incident with vi
= 0 otherwise
ii.
Representation of directed graph :
The incidence matrix I(D) = [bij] of digraph D with n vertices
and m edges is the n × m matrix in which.
bij = 1 if arc j is directed away from vertex vi
= – 1 if arc j is directed towards vertex vi
= 0 otherwise.
2. Linked representation :
a. In linked representation, the two nodes structures are used :
i. For non-weighted graph,
INFO Adj-list
ii. For weighted graph,
Weight INFO Adj-list
Where Adj-list is the adjacency list i.e., the list of vertices which are
adjacent for the corresponding node.
b. The header nodes in each list maintain a list of all adjacent vertices
of that node for which the header node is meant.

Que 3.5. Explain DFS. Also give DFS algorithm.

Answer
Depth First Search Algorithm :
1. Algorithm starts at a specific vertex S in G, which becomes current
vertex.
2. Then algorithm traverse graph by any edge (u, v) incident to the current
vertex u.
3. If the edge (u, v) leads to an already visited vertex v, then we backtrack
to current vertex u.
4. If, on other hand, edge (u, v) leads to an unvisited vertex v, then we go
to v and v becomes our current vertex.
5. We proceed in this manner until we reach to “dead end”. At this point we
start backtracking.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 3–7 B (CS/IT-Sem-5)

6. The process terminates when backtracking leads back to the start vertex.
7. Edges leads to new vertex are called discovery or tree edges and edges
lead to already visited vertex are called back edges.
Algorithm :
In DFS, each vertex v has two timestamps : the first timestamp d[v] i.e.,
discovery time records when v is first discovered i.e., grayed, and the second
timestamp f[v] i.e. finishing time records when the search finishes examining
v’s adjacency list i.e., blacked. For every vertex d[u] < f[u].
DFS(G) :
1. for each vertex u  V[G]
2. do color[u]  WHITE
3. [u]  NIL
4. time  0
5. for each vertex u  V[G]
6. do if color[u] = WHITE
7. then DFS-VISIT(u)
DFS-VISIT(u) :
1. color[u]  GRAY // White vertex u has just been discovered.
2. time  time + 1
3. d[u]  time
4. for each v  Adj[u] // Explore edge (u, v)
5. do if color[v] = WHITE
6. then [v]  u
7. DFS-VISIT (v)
8. color[u]  BLACK // Blacken u, it is finished.
9. f[u]  time  time + 1
Que 3.6. Explain Breadth First Search (BFS). Give its algorithm.

Answer
Breadth first search :
1. The general idea behind a breadth first search is as follows :
a. First we examine the starting node A.
b. Then, we examine all the neighbours of A, and so on.
2. Naturally, we need to keep track of the neighbours of a node, and we
need to guarantee that no node is processed more than once.
3. This is accomplished by using a queue to hold nodes that are waiting to
be processed.
Algorithm :
BFS (G, s) :
1. for each vertex u  V[G] – {s}
2. do color[u]  WHITE
 More pdf : www.motivationbank.in
Graph Algorithm 3–8 B (CS/IT-Sem-5)

3. d[u] 
4. [u]  NIL
5. color[s]  GRAY
6. d[s]  0
7. [s]  NIL
8. Q 
9. ENQUEUE(Q, s)
10. while Q 
do c
11. do u  DEQUEUE(Q)
12. for each v Adj[u]
13. do if color [v] = WHITE
14. then color[v]  GRAY
15. d[v] d[u] + 1
16. [v] u
17. ENQUEUE(Q, v)
18. color[u]  BLACK

Que 3.7. Write an algorithm to test whether a given graph is

connected or not.

Answer
Test-connected (G) :
1. Choose a vertex x
2. Make a list L of vertices reachable from x,
and another list K of vertices to be explored.
3. Initially, L = K = x.
4. while K is non-empty
5. Find and remove some vertex y in K
6. for each edge (y, z)
7. if (z is not in L)
8. Add z to both L and K
9. if L has fewer than n items
10. return disconnected
11. else return connected.

Que 3.8. Discuss strongly connected components with its

algorithm.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 3–9 B (CS/IT-Sem-5)

Answer
1. The Strongly Connected Components (SCC) of a directed graph G are
its maximal strongly connected subgraphs.
2. If each strongly connected component is contracted to a single vertex, the
resulting graph is a directed acyclic graph called as condensation of G.
3. Kosaraju’s algorithm is an algorithm to find the strongly connected
components of a directed graph.
Kosaraju’s algorithm :
1. Let G be a directed graph and S be an empty stack.
2. While S does not contain all vertices :
i. Choose an arbitrary vertex v not in S. Perform a depth first search
starting at v.
ii. Each time that depth first search finishes expanding a vertex u,
push u onto S.
3. Reverse the direction of all arcs to obtain the transpose graph.
4. While S is non-empty :
i. Pop the top vertex v from S. Perform a depth first search starting at
v.
ii. The set of visited vertices will give the strongly connected component
containing v; record this and remove all these vertices from the
graph G and the stack S.

Que 3.9. Explain convex hull problem.

AKTU 2017-18, Marks 10

OR
Discuss convex hull. Give Graham-Scan algorithm to compute
convex hull.
OR
What do you mean by convex hull ? Describe an algorithm that
solves the convex hull problem. Find the time complexity of the
algorithm. AKTU 2019-20, Marks 07

Answer
1. The convex hull of a set S of points in the plane is defined as the smallest
convex polygon containing all the points of S.
2. The vertices of the convex hull of a set S of points form a (not necessarily
proper) subset of S.
3. To check whether a particular point p S is extreme, see each possible
triplet of points and check whether p lies in the triangle formed by these
three points.
 More pdf : www.motivationbank.in
Graph Algorithm 3–10 B (CS/IT-Sem-5)

Fig. 3.9.1.
4. If p lies in any triangle then it is not extreme, otherwise it is.
5. We denote the convex hull of S by CH(S). Convex hull is a convex set
because the intersection of convex sets is convex and convex hull is also
a convex closure.
Graham-Scan algorithm :
The procedure GRAHAM-SCAN takes as input a set Q of points, where
|Q|  3. It calls the functions Top(S), which return the point on top of stack
S without changing S, and to NEXT-TO-TOP(S), which returns the point one
entry below the top of stack S without changing S.
GRAHAM-SCAN(Q)
1. Let p0 be the point in Q with the minimum y-coordinate, or the leftmost
such point in case of a tie.
2. Let <p1, p2, ...., pm> be the remaining points in Q, sorted by polar angle in
counter clockwise order around po (if more than one point has the same
angle remove all but the one that is farthest from po).
3. Top [S]  0
4. PUSH (p0, S)
5. PUSH (p1, S)
6. PUSH (p2, S)
7. for i  3 to m
8. do while the angle formed by points NEXT-To-TOP(S), Top(S), and
pi makes a non left turn.
9. do POP(S)
10. PUSH (pi, S)
11. return S
Time complexity :
The worst case running time of GRAHAM-SCAN is
T(n) = O(n) + O(n log n) + O(1) + O(n) = O(n log n)
where n = |Q|
Graham’s scan running time depends only on the size of the input it is
independent of the size of output.
 More pdf : www.motivationbank.in
3–11 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

PART-2
Greedy Methods with Examples such as Optimal
Reliability Allocation.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 3.10. Write note on the greedy algorithm.

Answer
1. Greedy algorithms are simple and straight forward.
2. Greedy algorithms are shortsighted in their approach in the sense that
they take decisions on the basis of information at hand without worrying
about the effect these decisions may have in the future.
3. Greedy algorithms are easy to invent, easy to implement and most of
the time quite efficient.
4. Many problems cannot be solved correctly by greedy approach.
5. Greedy algorithms are used to solve optimization problems.

Que 3.11. What are the four functions included in greedy

algorithm ? Write structure of greedy algorithm.

Answer
The greedy algorithm consists of four functions :
i. A function that checks whether chosen set of items provide a solution.
ii. A function that checks the feasibility of a set.
iii. The selection function tells which of the candidates are most promising.
iv. An objective function, which does not appear explicitly, gives the value
of a solution.
Structure of greedy algorithm :
i. Initially the set of chosen items is empty i.e., solution set.
ii. At each step
a. Item will be added in a solution set by using selection function.
b. If the set would no longer be feasible
Reject items under consideration (and is never consider again).
c. Else if set is still feasible
add the current item.
 More pdf : www.motivationbank.in
Graph Algorithm 3–12 B (CS/IT-Sem-5)

Que 3.12. Define activity selection problem and give its solution
by using greedy approach with its correctness.

Answer
1. An activity selection is the problem of scheduling a resource among
several competing activity. Given a set S = {1, 2, …, n} of n activities.
2. Each activity has si a start time, and fi a finish time.
3. If activity i is selected, the resource is occupied in the intervals (si, fi).We
say i and j are compatible activities if their start and finish time does not
overlap i.e., i and j compatible if si  fj and sj  fi
4. The activity selection problem is, to select a maximal sized subset of
mutually compatible activities.
Here we maximize the number of activities selected, but if the profit were
proportional to si – fi, this will not maximize the profit.

2
i si fi 1
3
1 1 4 1
2 3 5 4
1
3 0 6 5
1 4
4 5 7
6
4 3 8 1 4
7
6 5 9 1 4
7 6 10 8
1 4
8 8 11 9
1 4 8
9 8 12
10
10 2 13 1 4 8
11
11 12 14 1 4 8
12  -
1 4 8 11
time
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Fig. 3.12.1.

Greedy algorithm :
Assume that f1  f2  …  fn
 More pdf : www.motivationbank.in
3–13 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Greedy-Activity-Selector (s, f)
1. n  length [s]
2. A {a1}
3. i1
4. for m  2 to n
5. do if sm  fi
6. then A  A {(m)}
7. im
8. return A
The algorithm starts with {1} and checks to see which can be added after 1,
updating the global “finishing time” and comparing with each start time.
The activity picked is always the first that is compatible. Greedy algorithms
do not always produce optimal solutions.
Correctness : Greedy algorithm does not always produce optimal solutions
but GREEDY-ACTIVITY-SELECTOR does.

Que 3.13. What are greedy algorithms ? Find a solution to the

following activity selection problem using greedy technique. The
starting and finishing times of 11 activities are given as follows :
(2, 3) (8, 12) (12, 14) (3, 5) (0, 6) (1, 4) (6, 10) (5, 7) (3, 8) (5, 9) (8, 11)
OR
What is greedy approach ? Write an algorithm which uses this
approach.

Answer
Greedy algorithm : Refer Q. 3.10, Page 3–11B, Unit-3.
Greedy approach : Greedy approach works by making the decision that
seems most promising at any moment it never reconsiders this decision,
whatever situation may arise later. Activity selection problem uses greedy
approach.
Algorithm for greedy activity selection : Refer Q. 3.12, Page 3–12B,
Unit-3.
Numerical :
Sorted activities a1 a2 a3 a4 a5 a6 a7 a8 a9 a10 a11
Starting time 2 1 3 0 5 3 5 6 8 8 12
Finish time 3 4 5 6 7 8 9 10 11 12 14
We select first activity a1
(2, 3)
Check for activity a2
Starting time of a2  time of a1
 More pdf : www.motivationbank.in
Graph Algorithm 3–14 B (CS/IT-Sem-5)

 a2 is not selected
Check for activity a3
Starting time of a3  finish time of a1
 a3 is selected
Check for activity a4
Starting time of a4  finish time of a3
 a4 is not selected
Check for activity a5
Starting time of a5  finish time of a3
 a5 is selected
Check for activity a6
Starting time of a6  finish time of a5
 a6 is not selected
Check for activity a7
Starting time of a7  finish time of a5
 a7 is not selected
Check for activity a8
Starting time of a8  finish time of a5
 a8 is not selected
Check for activity a9
Starting time of a9  finish time of a5
 a9 is selected
Check for activity a10
Starting time of a10  finish time of a9
 a10 is not selected
Check for activity a11
Starting time of a11  finish time of a9
 a11 is selected.
 Therefore selected activities are :
a1 : (2, 3)
a3 : (3, 5)
a5 : (5, 7)
a9 : (8, 11)
a11 : (12, 14)

Que 3.14. What is “Greedy Algorithm” ? Write its pseudocode for

recursive and iteration process.

Answer
Greedy algorithm : Refer Q. 3.10, Page 3–11B, Unit-3.
Greedy algorithm defined in two different forms :
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 3–15 B (CS/IT-Sem-5)

i. Pseudo code for recursive greedy algorithm :

R_A_S (s, f, i, j)
1. mi+1
2. while m < j and sm < fi
3. do m  m + 1
4. if m < j
5. then return {am} R_A_S(s, f, m, j)
6. else return 
ii. Pseudo code for iterative greedy algorithm :
G_A_S (s, f)
1. n  length [s]
2. A [a1]
3. i1
4. m  2 to n
5. do if sm  fi
6. then A  A {am}
7. im
8. return A

Que 3.15. What is an optimization problem ? How greedy method

can be used to solve the optimization problem ?

Answer
1. An optimization problem is the problem of finding the best solution
from all feasible solutions.
2. Optimization problems can be divided into two categories depending
on whether the variables are continuous or discrete.
3. There is no way in general that one can specify if a greedy algorithm
will solve a particular optimization problem.
4. However if the following properties can be demonstrated, then it is
probable to use greedy algorithm :
a. Greedy choice property : A globally optimal solution can be
arrived at by making a locally optimal greedy choice. That is, when
we are considering which choice to make, we make the choice
that looks best in the current problem, without considering results
from sub-problems. ·
b. Optimal substructure : A problem exhibits optimal substructure
if an optimal solution to the problem contains within it optimal
solutions to sub-problems.
 More pdf : www.motivationbank.in
Graph Algorithm 3–16 B (CS/IT-Sem-5)

PART-3
Knapsack.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 3.16. What is knapsack problem ? Describe an approach

used to solve the problem.

Answer
1. The knapsack problem is a problem in combinatorial optimization.
2. Given a set of items, each with a weight and a value, determine the
number of each item to include in a collection so that the total weight is
less than or equal to a given limit and the total value is as large as
possible.
Approach use to solve the problem :
1. In knapsack problem, we have to fill the knapsack of capacity W, with
a given set of items I1, I2... In having weight w1, w2...wn in such a manner
that the total weight of items cannot exceed the capacity of knapsack
and maximum possible value (v) can be obtained.
2. Using branch and bound approach, we have a bound that none of the
items can have total sum more than the capacity of knapsack and must
give maximum possible value.
3. The implicit tree for this problem is a binary tree in which left branch
implies inclusion and right implies exclusion.
4. Upper bound of node can be calculated as :
ub = v + (W – wi+1) (vi+1 / wi+1)

Que 3.17. Write greedy algorithm for discrete knapsack problem

with example.

Answer
Greedy algorithm for the discrete knapsack problem :
1. Compute value/weight ratio vi /wi for all items.
2. Sort the items in non-increasing order of the ratios vi /wi.
3. Repeat until no item is left in sorted list using following steps :
a. If current item fits, use it.
b. Otherwise skip this item, and proceed to next item.
 More pdf : www.motivationbank.in
3–17 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

For example : Knapsack problem for the following instance using greedy
approach. The item can be selected or skipped completely.

Item Weight Value

1 7 ` 49

2 3 ` 12

3 4 ` 42

4 5 ` 30
Consider W = 10.
Solution : This is also called 0/1 knapsack. Either we can completely select
an item or skip it. First of all we will compute value-to-weight ratio and
arrange them in non-increasing order of the ratio.
Item Weight Value Value/Weight

3 4 ` 42 10.5
1 7 ` 49 7
4 5 ` 30 6
2 3 ` 12 4

To fulfill the capacity W = 10, we will have

1. Add item of weight 4, W = 10 – 4 = 6
2. Skip item of weight 7
3. Add item of weight 5, W = 6 – 5 = 1
4. Skip item of weight 3
Maximum value = 10.5 + 6 = 16.5
But the greedy algorithm does not give optimal solution always rather there
is no upper bound on the accuracy of approximate solution.

Que 3.18. What is 0/1-knapsack problem ? Does greedy method

effective to solve the 0/1-knapsack problem ?

Answer
The 0/1-knapsack problem is defined as follows :
1. Given, a knapsack of capacity c and n items of weights {w1, w2, ..., wn}
and profits {p1, p2, ....., pn}, the objective is to choose a subset of n
objects that fits into the knapsack and that maximizes the total profit.
2. Consider a knapsack (bag) with a capacity of c.
3. We select items from a list of n items.
 More pdf : www.motivationbank.in
Graph Algorithm 3–18 B (CS/IT-Sem-5)

4. Each item has both a weight of wi and profit of pi.

5. In a feasible solution, the sum of the weights must not exceed the
knapsack capacity (c) and an optimal solution is both feasible and
reaches the maximum profit.
6. An optimal packing is a feasible solution one with a maximum profit :
n
p1x1 + p2x2 + p3x3 + .... + pnxn = px
i1
i i

which is subjected to constraints :

n
w1x1 + w2x2 + w3x3 + .... + wnxn = w x i i c
i1
and xi = 1 or 0, 1  i  n
7. We have to find the values of xi where xi = 1 if ith item is packed into
the knapsack and xi = 0 if ith item is not packed.
Greedy strategies for the knapsack problem are :
i. From the remaining items, select the item with maximum profit that
fits into the knapsack.
ii. From the remaining items, select the item that has minimum weight
and also fits into the knapsack.
iii. From the remaining items, select the one with maximum pi/wi that
fits into the knapsack.
Greedy method is not effective to solve the 0/1-knapsack problem. By using
greedy method we do not get optimal solution.
Que 3.19. Given the six items in the table below and a knapsack
with weight 100, what is the solution to the knapsack problem in
all concepts. i.e., explain greedy all approaches and find the optimal
solution.
Item ID Weight Value Value/Weight
A 100 40 0.4
B 50 35 0.7
C 40 20 0.5
D 20 4 0.2
E 10 10 1
F 10 6 0.6

AKTU 2017-18, Marks 10

 More pdf : www.motivationbank.in
3–19 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Answer
We can use 0/1-knapsack problem when the items cannot be divided into
parts and fractional knapsack problem when the items can be divided into
fractions.
First arrange in non-increasing order of value/weight :
Item ID Weight Value Value/Weight
E 10 10 1
B 50 35 0.7
F 10 6 0.6
C 40 20 0.5
A 100 40 0.4
D 20 4 0.2
According to 0/1-knapsack problem, either we select an item or reject. So the
item will be selected according to value per weight.
E is selected W = 10 < 100
B is selected W = 10 + 50
= 60 < 100
F is selected W = 60 + 10
= 70 < 100
C cannot be selected because
W = 70 + 40 = 110 > 100
Hence we select D
W = 70 + 20 = 90 < 100
Total value = 10 + 35 + 6 + 4 = 55
According to fractional knapsack problem, we can select fraction of any item.
E is selected W = 10 < 100
B is selected W = 10 + 50
= 60 < 100
F is selected W = 60 + 10
= 70 < 100
If we select C W = 70 + 40
= 110 > 100
Hence we select the fraction of item C as
100  W 100  70
=
Weight of C 40
Weight of C = 30/40 = 0.75
So, W= 0.75 × 40 = 30
W= 70 + 30 = 100
Total value = 10 + 35 + 6 + 0.75 (20)
= 10 + 35 + 6 + 15 = 66
 More pdf : www.motivationbank.in
Graph Algorithm 3–20 B (CS/IT-Sem-5)

Que 3.20. Consider the weight and values of item listed below.
Note that there is only one unit of each item. The task is to pick a
subset of these items such that their total weight is no more than
11 kgs and their total value is maximized. Moreover, no item may
be split. The total value of items picked by an optimal algorithm is
denoted by Vopt. A greedy algorithm sorts the items by their value-
to-weight rations in descending order and packs them greedily,
starting from the first item in the ordered list. The total value of
items picked by the greedy algorithm is denoted by V greedy. Find
the value of V opt – V greedy.
Item I1 I2 I3 I4
W 10 7 4 2
V 60 28 20 24

AKTU 2018-19, Marks 07

Answer
For Vgreedy :
Item W V
I1 10 60
I2 7 28
I3 4 20
I4 2 24
Arrange the items by V/W ratio in descending order :
Item W V V/W
I4 2 24 12
I1 10 60 6
I3 4 20 5
I2 7 28 4
Total weight W = 11 kg
I4 is picked so W = 11 – 2 = 9 kg
I1 cannot picked 10 > 9
I3 is picked, W = 9 – 4 = 5 kg
I2 cannot be picked 7 > 5
I4 and I3 are picked so
Vgreedy = V(I4) + V(I3) = 24 + 20 = 44
 More pdf : www.motivationbank.in
3–21 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

For Vopt : For calculating Vopt we use 0/1 knapsack problem, so only item 1 is
picked. Hence, Vopt = 60
So, Vopt – Vgreedy = 60 – 44 = 16

Que 3.21. Consider following instance for simple knapsack

problem. Find the solution using greedy method.
N=8
P = {11, 21, 31, 33, 43, 53, 55, 65}
W = {1, 11, 21, 23, 33, 43, 45, 55}
M = 110 AKTU 2016-17, Marks 7.5

Answer
N=8
W = {1, 11, 21, 23, 33, 43, 45, 55}
P = {11, 21, 31, 33, 43, 53, 55, 65}
M = 110
Now, arrange the value of Pi in decreasing order
N Wi Pi V i = Wi × P i
1 1 11 11
2 11 21 231
3 21 31 651
4 23 33 759
5 33 43 1419
6 43 53 2279
7 45 55 2475
8 55 65 3575
Now, fill the knapsack according to decreasing value of Pi. First we choose
item N = 1 whose weight is 1.
Then choose item N = 2 whose weight is 11.
Then choose item N = 3 whose weight is 21.
Now, choose item N = 4 whose weight is 23.
Then choose item N = 5 whose weight is 33.
Total weight in knapsack is = 1 + 11 + 21 + 23 + 33 = 89
Now, the next item is N = 6 and its weight is 43, but we want only 21 because
M = 110.
So, we choose fractional part of it, i.e.,
The value of fractional part of N = 6 is,
2279
× 21 = 1113
43
 More pdf : www.motivationbank.in
Graph Algorithm 3–22 B (CS/IT-Sem-5)

21

33

23

21  110

11

Thus, the maximum value is,

= 11 + 231 + 651 + 759 + 1419 + 1113
= 4184

Que 3.22. Solve the following 0/1-knapsack problem using dynamic

programming P = {11, 21, 31, 33} w = {2, 11, 22, 15} c = 40, n = 4.
AKTU 2019-20, Marks 07

Answer
Numerical :
w = {2, 11, 22, 15}
c = 40
p = {11, 21, 31, 33}
Initially,
Item wi pi
I1 2 11
I2 11 21
I3 22 31
I4 15 33
Taking value per weight ratio, i.e., pi / wi
Item wi vi / wi pi
I1 2 11 22
I2 11 21 232
I3 22 31 682
I4 15 33 495

Now, arrange the value of pi in decreasing order.

 More pdf : www.motivationbank.in
3–23 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Item wi pi pi
I3 22 31 682
I4 15 33 495
I2 11 21 232
I1 2 11 22

Now, fill the knapsack according to decreasing value of pi.

First we choose item I3 whose weight is 22, then choose item I4 whose
weight is 15. Now the total weight in knapsack is 22 + 15 = 37. Now, next
item is I2 and its weight is 11 and then again I1. So, we choose fractional
part of it, i.e.,

22
15
40
3
The value of fractional part of I1 is,
232
= × 3 = 63
11
Thus, the maximum value is,
= 682 + 495 + 63 = 1190

PART-4
Minimum Spanning Trees-Prim’s and Kruskal’s Algorithm, Single.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 3.23. What do you mean by spanning tree and minimum

spanning tree ?

Answer
Spanning tree :
1. A spanning tree of a graph is a subgraph that contains all the vertices
and is a tree.
2. A spanning tree of a connected graph G contains all the vertices and has
the edges which connect all the vertices. So, the number of edges will be
1 less the number of nodes.
 More pdf : www.motivationbank.in
Graph Algorithm 3–24 B (CS/IT-Sem-5)

3. If graph is not connected, i.e., a graph with n vertices has edges less than
n – 1 then no spanning tree is possible.
4. A graph may have many spanning trees.
Minimum spanning tree :
1. Given a connected weighted graph G, it is often desired to create a
spanning tree T for G such that the sum of the weights of the tree edges
in T is as small as possible.
2. Such a tree is called a minimum spanning tree and represents the
‘‘cheapest’’ way of connecting all the nodes in G.
3. There are number of techniques for creating a minimum spanning tree
for a weighted graph but the most famous methods are Prim’s and
Kruskal’s algorithm.

Que 3.24. Write Kruskal’s algorithm to find minimum spanning

tree.

Answer
i. In this algorithm, we choose an edge of G which has smallest weight
among the edges of G which are not loops.
ii. This algorithm gives an acyclic subgraph T of G and the theorem given
below proves that T is minimal spanning tree of G. Following steps are
required :
Step 1 : Choose e1, an edge of G, such that weight of e1, w(e1) is as small
as possible and e1 is not a loop.
Step 2 : If edges e1, e2, ........., ei have been selected then choose an edge
ei+1 not already chosen such that
i. the induced subgraph
G[{e1 ..... ei+1}] is acyclic and
ii. w(ei+1) is as small as possible
Step 3 : If G has n vertices, stop after n – 1 edges have been chosen.
Otherwise repeat step 2.
If G be a weighted connected graph in which the weight of the edges are all
non-negative numbers, let T be a subgraph of G obtained by Kruskal’s
algorithm then, T is minimal spanning tree.

Que 3.25. Des cribe and compare following algorithms to

determine the minimum cost spanning tree :
i. Kruskal’s algorithm
ii. Prim’s algorithm
OR
Define spanning tree. Write Kruskal’s algorithm or finding minimum
cost spanning tree. Describe how Kruskal’s algorithm is different
from Prim’s algorithm for finding minimum cost spanning tree.
AKTU 2019-20, Marks 07
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 3–25 B (CS/IT-Sem-5)

Answer
Spanning tree : Refer Q. 3.23, Page 3–23B, Unit-3.
i. Kruskal’s algorithm : Refer Q. 3.24, Page 3–24B, Unit-3.
ii. Prim’s algorithm :
First it chooses a vertex and then chooses an edge with smallest weight
incident on that vertex. The algorithm involves following steps :
Step 1 : Choose any vertex V1 of G.
Step 2 : Choose an edge e1 =V1V2 of G such that V2  V1 and e1 has
smallest weight among the edge e of G incident with V1.
Step 3 :If edges e1, e2, ......., ei have been chosen involving end points
V1, V2, ............., Vi+1, choose an edge ei+1 = VjVk with Vj = {V1 ....... Vi+1}
and Vk  {V1 ...........Vi+1} such that ei+1 has smallest weight among the
edges of G with precisely one end in {V1 ............... Vi+1}.
Step 4 :Stop after n – 1 edges have been chosen. Otherwise goto
step 3.
Comparison :
S. No. Kruskal’s algorithm Prim’s algorithm
1. Kruskal’s algo rithm Prim’s algorithm initializes with a
initiates with an edge. node.
2. Kruskal’s algorithm selects Prim’s algorithms span from one
the edges in a way that the node to another.
position of the edge is not
based on the last step.
3. Kruskal’s can be used on In Prim’s algorithm, graph must be
disconnected graphs. a connected graph.
4. Kruskal’s time complexity Prim’s algo rithm has a time
in worst case is O(E log E). co mple xity in wo rst case o f
O(E log V).

Que 3.26. What do you mean by minimum spanning tree ? Write

an algorithm for minimum spanning tree that may generate
multiple forest trees and also explain with suitable example.

Answer
Minimum spanning tree : Refer Q. 3.23, Page 3–23B, Unit-3.
Prim’s algorithm : Refer Q. 3.25, Page 3–24B, Unit-3.
For example :
According to algorithm we choose vertex A from the set {A, B, C, D, E, F, G,
H, I, J}.
 More pdf : www.motivationbank.in
Graph Algorithm 3–26 B (CS/IT-Sem-5)

4 C
4 B 2 1
A E
4 F
2
1 10
D 3
G 5
6 3
5 4
H J
2 I
3
Now edge with smallest weight incident on A is e = AD
4
B C
4 2
1
A E
4 F
2 3
1 10
D G
5
5 6 4 3

H J I
2 3
Now we look on weight
W(A, B) = 4
W(D, B) = 4 W(D, H) = 5
W(D, J) = 6
We choose e = AB, since it is minimum.
W(D, B) can also be chosen because it has same value.
4
B C
4 2
1
A 4 E
2 F
1 10 3
D G
5
5 6 4 3

H J I
2 3
Again, W(B, C) = 4
W(B, J) = 10
W(D, H) = 5
W(D, J) = 6
We choose e = BC, since it has minimum value.
 More pdf : www.motivationbank.in
3–27 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

B 4 C
4 2 1
A 4 E F
1 10 2 3
D G
5 6 4 3 5

H J I
2 3
Now, W(B, J) = 10
W(C, E) = 2
W(C, F) = 1
We choose e = CF, since it has minimum value.

B 4 C
4 2 1
A 4 E F
1 10 2 3
D G
5 6 4 3 5

H J I
2 3
Now, W(C, E) = 2
W(F, G) = 3
W(F, I) = 5
We choose e = CE, since it has minimum value.

B 4 C
4 2 1
A 4 E F
1 10 2 3
D G
5 6 4 3 5

H J I
2 3

W(E, G) = 2
W(F, G) = 3
W(F, I) = 5
We choose e = EG, since it has minimum value.
 More pdf : www.motivationbank.in
Graph Algorithm 3–28 B (CS/IT-Sem-5)

B 4 C
4 2 1
A 4 F
E
2 3
1 10
D G
5 6 4 3 5

H J I
2 3

W(G, J) = 4
W(G, I) = 3
W(F, I) = 5
We choose e = GI, since it has minimum value.

B 4 C
4 2 1
A 4 F
E
2 3
1 10
D G
5 6 4 3 5

H J I
2 3

W(I, J) = 3
W(G, J) = 4
We choose e = IJ, since it has minimum value.
W(J, H) = 2
Hence, e = JH will be chosen.
The final minimal spanning tree is given as :

B 4 C
4 2 1
A E F
1 2
D G
4
3
H J I
2 3
Fig. 3.26.1.
 More pdf : www.motivationbank.in
3–29 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Que 3.27. What is minimum cost spanning tree ? Explain Kruskal’s

algorithm and find MST of the graph. Also write its time complexity.

7 1 6
2
2 4
6 6
3 3
4 5
7
5 6
7
Fig. 3.27.1.

AKTU 2017-18, Marks 10

Answer
Minimum spanning tree : Refer Q. 3.23, Page 3–23B, Unit-3.
Kruskal’s algorithm : Refer Q. 3.24, Page 3–24B, Unit-3.
Numerical :
Step 1 : Arrange the edge of graph according to weight in ascending order.
Edges Weight Edge Weight
13 2 32 6
46 3 17 7
25 4 35 7
36 5 56 7
34 6
41 6
Step 2 : Now draw the vertices as given in graph,

2 4
3

5 6

Now draw the edge according to the ascending order of weight. If any edge
forms cycle, leave that edge.
 More pdf : www.motivationbank.in
Graph Algorithm 3–30 B (CS/IT-Sem-5)

Step 3 : Select edge 13

1

2 2 4
3

5 6
Step 4 : Select edge 46
1

2 2 4
3
3

5 6
Step 5 : Select edge 25
1

2 2 4

4 3
3

5 6
Step 6 : Select edge 36
1

2 2 4

4 3
3
5
5 6
Step 7 : Select edge 23

2 2 4
6
4 3
3
5
5 6
All the remaining edges, such as 34, 41, 12, 35, 56 are rejected because they
form cycle.
All the vertices are covered in this tree. So, the final tree with minimum cost
of given graph is
 More pdf : www.motivationbank.in
3–31 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

2 2 4
6
4 3
3
5
5 6

Minimum cost = 2 + 3 + 4 + 5 + 6 = 20
Time complexity : Time complexity is O(|E| log |E|).

Que 3.28. What is minimum spanning tree ? Explain Prim’s

algorithm and find MST of graph Fig. 3.28.1.
6 7
b c d
4 9
2
a 11 i 4 14 e
7 6
5 10
h g f
1 2
Fig. 3.28.1.

AKTU 2015-16, Marks 05

Answer
Minimum spanning tree : Refer Q. 3.23, Page 3–23B, Unit-3.
Prim’s algorithm : Refer Q. 3.25, Page 3–24B, Unit-3.
Numerical :
6 7
b c d
4 2 9
4
a 11 i 14 e
6
5 7 10
h g f
1 2
Fig. 3.28.2.
Let a be the source node. Select edge (a, b) as distance between edge (a, b)
is minimum.
b
4

a
 More pdf : www.motivationbank.in
Graph Algorithm 3–32 B (CS/IT-Sem-5)

Now, select edge (b, c)

6
b c
4

a
Now, select edge (c, i)
6
b c
4
2
a i
Now, select edge (i, g)
6
b c
4
2
a i
6
g
Now, select edge (g, h)
6
b c
4
2
a i
6

h g
1
Now, select edge (g, f)
6
b c
4
2
a i
6
1 2
h g f
Now, select edge (f, e)
6
b c
4
2
a i e
6
10
1 g 2
h f
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 3–33 B (CS/IT-Sem-5)

Now, select edge (e, d)

6
b c
4 2
a i d
6

h 1 g 2 f e

Thus, we obtained MST for Fig. 3.28.1.

PART-5
Source Shortest Paths-Dijkstra’s and Bellman-Ford Algorithm.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 3.29. Prove that the weights on the edge of the connected
undirected graph are distinct then there is a unique minimum
spanning tree. Give an example in this regard. Also discuss prim’s
minimum spanning tree algorithm in detail.
AKTU 2018-19, Marks 07

Answer
Proof :
1. Let we have an algorithm that finds an MST (which we will call A) based
on the structure of the graph and the order of the edges when ordered
by weight.
2. Assume MST A is not unique.
3. There is another spanning tree with equal weight, say MST B.
4. Let e1 be an edge that is in A but not in B.
5. Then, B should include at least one edge e2 that is not in A.
6. Assume the weight of e1 is less than that of e2.
7. As B is a MST, {e1} B must contain a cycle.
8. Replace e2 with e1 in B yields the spanning tree {e1} B – {e2} which has
a smaller weight compared to B.
9. This contradicts that B is not a MST.
So, MST of undirected graph with distinct edge is unique.
 More pdf : www.motivationbank.in
Graph Algorithm 3–34 B (CS/IT-Sem-5)

Example :
1
1 7

2 2 4

4 3 3
8 5

5 6
6
Fig. 3.29.1.
Step 1 : Arrange the edge of graph according to weight in ascending
order.
Edges Weight Edge Weight
12 1 14 7
13 2 35 8
46 3
25 4
36 5
56 6
Step 2 : Now draw the vertices as given in graph,

2 4
3

5 6

Now draw the edge according to the ascending order of weight. If any
edge forms cycle, leave that edge.
Step 3 :

1 1

2 4
3

5 6
 More pdf : www.motivationbank.in
3–35 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Step 4 :

1 1

2 2 4
3

5 6
Step 5 :
1 1

2 2 4
3
3

5 6
Step 6 :

1 1

2 2 4

4 3
3

5 6

Step 7 :

1 1

2 2 4

4 3
3
5
5 6
All the remaining edges, such as : 14, 35, 56 are rejected because they
form cycle.
All the vertices are covered in this tree. So, the final tree with minimum
cost of given graph is
1 1

2 2 4

4 3
3
5
5 6
Prim’s algorithm : Refer Q. 3.25, Page 3–24B, Unit-3.
 More pdf : www.motivationbank.in
Graph Algorithm 3–36 B (CS/IT-Sem-5)

Que 3.30. Write an algorithm to find shortest path between all

pairs of nodes in a given graph.
OR
Explain greedy single source shortest path algorithm with example.
AKTU 2015-16, Marks 10
OR
Write short note on Dijkstra’s algorithm shortest paths problems.
AKTU 2016-17, Marks 10

Answer
1. Dijkstra’s algorithm, is a greedy algorithm that solves the single source
shortest path problem for a directed graph G = (V, E) with non-negative
edge weights, i.e., we assume that w(u, v)  0 each edge (u, v)  E.
2. Dijkstra’s algorithm maintains a set S of vertices whose final shortest
path weights from the source s have already been determined.
3. That is, for all vertices v  S, we have d[v] = (s, v).
4. The algorithm repeatedly selects the vertex u  V – S with the minimum
shortest path estimate, inserts u into S, and relaxes all edges leaving u.
5. We maintain a priority queue Q that contains all the vertices in
v – s, keyed by their d values.
6. Graph G is represented by adjacency list.
7. Dijkstra’s always chooses the “lightest or “closest” vertex in V – S to
insert into set S that it uses as a greedy strategy.
Dijkstra’s algorithm :
DIJKSTRA (G, w, s)
1. INITIALIZE-SINGLE-SOURCE (G, s)
2. s  
3. Q  V[G]
4. while Q  
5. do u  EXTRACT-MIN (Q)
6. S  S  {u}
7. for each vertex v  Adj [u]
8. do RELAX (u, v, w)
RELAX (u, v, w) :
1. If d[u] + w(u, v) < d[v]
2. then d[v]  d[u] + w(u, v)
3. [v]  u
 More pdf : www.motivationbank.in
3–37 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Que 3.31. Find the shortest path in the below graph from the
source vertex 1 to all other vertices by using Dijkstra’s algorithm.
50
2 3
10
10
1 30 20
100
5 4
60

AKTU 2017-18, Marks 10

Answer
Initialize :
 50  S:{}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20

100
5 4
 60 
Extract min (1) :
 50  S : {1}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20

100
5 4
 60 
All edges leaving (1) :
10 
50 S : {1}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20
10  30 100
100
5 4
60
100 30
 More pdf : www.motivationbank.in
Graph Algorithm 3–38 B (CS/IT-Sem-5)

Extract min(2) :
50 60 S : {1, 2}
2 3
10 Q:1 2 3 4 5
10 0    
1 30 20
10 60 30 100
100
5 4
60
All edges leaving (2) :
50 60 S : {1, 2}
2 3
10 Q:1 2 3 4 5
10 0    
1 30 20
10 60 30 100
100
5 4 60 30 100
60 30
100
Extract min(4) :
10 60
50 S : {1, 2, 4}
2 3
10 Q:1 2 3 4 5
10 0    
0 1 30 20
10 60 30 100
100
5 4 60 30 100
60
100 30
All edges leaving (4) :
50 50 S : {1, 2, 4}
2 3 Q:1 2 3 4 5
10
0    
10
1 30 20 10 60 30 100
100 60 30 100
5 4 50
60
100 30
Extract min(3) :
10 S : {1, 2, 4, 3}
50
2 3 50 Q:1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50
100
 More pdf : www.motivationbank.in
3–39 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

All edges leaving (3) :

10 S : {1, 2, 4, 3, }
50
2 3 50 Q:1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50 60
100
Extract min(5) :
10 S : {1, 2, 4, 3, 5}
50
2 3 50 Q: 1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50
60
60

Shortest path
2 3
10

1 30 10 20

5 4

Que 3.32. State Bellman-Ford algorithm.

AKTU 2016-17, Marks 7.5

Answer
1. Bellman-Ford algorithm finds all shortest path length from a source
s  V to all v  V or determines that a negative-weight cycle exists.
2. Bellman-Ford algorithm solves the single source shortest path problem
in the general case in which edges of a given digraph G can have
negative weight as long as G contains no negative cycles.
3. This algorithm, uses the notation of edge relaxation but does not use
with greedy method.
4. The algorithm returns boolean TRUE if the given digraph contains no
negative cycles that are reachable from source vertex otherwise it
returns boolean FALSE.
 More pdf : www.motivationbank.in
Graph Algorithm 3–40 B (CS/IT-Sem-5)

Bellman-Ford (G, w, s) :
1. INITIALIZE-SINGLE-SOURCE (G, s)
2. for each vertex i  1 to V[G] – 1
3. do for each edge (u, v) in E[G]
4. do RELAX (u, v, w)
5. for each edge (u, v) in E[G] do
6. do if d[u] + w(u, v) < d[v] then
7. then return FALSE
8. return TRUE
RELAX (u, v, w) :
1. If d[u] + w(u, v) < d[v]
2. then d[v]  d[u] + w(u, v)
3. [v]  u
If Bellman-Ford returns true, then G forms a shortest path tree, else there
exists a negative weight cycle.

Que 3.33. When do Dijkstra and the Bellman Ford algorithm

both fail to find a shortest path ? Can Bellman Ford detect all
negative weight cycles in a graph ? Apply Bellman Ford algorithm
on the following graph :
Source
vertex 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G
2 C 11 1
D E
Fig. 3.33.1.

AKTU 2018-19, Marks 07

Answer
Dijkstra algorithm fails to find a shortest path when the graph contains
negative edges.
Bellman Ford algorithm fails to find a shortest path when the graph contain
negative weight cycle.
No, Bellman Ford cannot detect all negative weight cycle in a graph.
 More pdf : www.motivationbank.in
3–41 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Numerical :
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 
2 C 11
 D E 1
7 
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 
1
2 C 11
 D E 1
7 
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 
2 C 11
D E 1
7 12
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 13
2 C 11
D E 1
7 12
 15 
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 13
2 C 11
D E 1
7 12
 15 18
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 13
2 C 11
D E 1
7
 More pdf : www.motivationbank.in
Graph Algorithm 3–42 B (CS/IT-Sem-5)

VERY IMPORTANT QUESTIONS

Following questions are very important. These questions
may be asked in your SESSIONALS as well as
UNIVERSITY EXAMINATION.

Q. 1. Discuss matrix chain multiplication problem and its solution.

Ans. Refer Q. 3.2.

Q. 2. Explain graphs with its representations.

Ans. Refer Q. 3.4.

Q. 3. Write short note on convex null problem.

Ans. Refer Q. 3.9.

Q. 4. What is greedy algorithm ? Write its pseudocode for

recursive and iterative process.
Ans. Refer Q. 3.14.

Q. 5. Discuss 0/1-knapsack problem.

Ans. Refer Q. 3.18.

Q. 6. Write short note on the following :

a. Minimum spanning tree
b. Kruskal’s algorithm
c. Prim’s algorithm
Ans.
a. Refer Q. 3.23.
b. Refer Q. 3.24.
c. Refer Q. 3.25.

Q. 7. Write an algorithm to find shortest path between all pairs

of nodes in a given graph.
Ans. Refer Q. 3.30.
Q. 8. State Bellman Ford algorithm.
Ans. Refer Q. 3.32.

Q. 9. Solve the following 0/1-knapsack problem using dynamic

programming P = {11, 21, 31, 33} w = {2, 11, 22, 15} c = 40,
n = 4.
Ans. Refer Q. 3.22.


 More pdf : www.motivationbank.in
4–1 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

4 Dynamic Programming
and Backtracking

CONTENTS
Part-1 : Dynamic Programming with .................... 4–2B to 4–8B
Examples such as Knapsack

Part-2 : All Pair Shortest Paths : .......................... 4–8B to 4–12B

Warshall’s and Floyd’s Algorithm,
Resource Allocation Problem

Part-3 : Backtracking, Branch and .................... 4–12B to 4–21B

Bound with Examples such as
Travelling Salesman Problem

Part-4 : Graph Colouring, .................................... 4–21B to 4–26B

N-Queen Problem

Part-5 : Hamiltonian Cycles ................................ 4–26B to 4–33B

and Sum of Subsets
 4–2 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
PART-1
Dynamic Programming with Examples such as Knapsack.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 4.1. What do you mean by dynamic programming ?

OR
What is dynamic programming ? How is this approach different
from recursion? Explain with example.
AKTU 2019-20, Marks 07

Answer
1. Dynamic programming is a stage-wise search method suitable for
optimization problems whose solutions may be viewed as the result of a
sequence of decisions.
2. It is used when the sub-problems are not independent.
3. Dynamic programming takes advantage of the duplication and arranges
to solve each sub-problem only once, saving the solution (in table or
something) for later use.
4. Dynamic programming can be thought of as being the reverse of
recursion. Recursion is a top-down mechanism i.e., we take a problem,
split it up, and solve the smaller problems that are created. Dynamic
programming is a bottom-up mechanism i.e., we solve all possible small
problems and then combine them to obtain solutions for bigger problems.
Difference :
1. In recursion, sub-problems are solved multiple times but in dynamic
programming sub-problems are solved only one time.
2. Recursion is slower than dynamic programming.
For example :
Consider the example of calculating nth Fibonacci number.
fibo(n) = fibo(n – 1) + fibo(n – 2)
fibo(n – 1) = fibo(n – 2) + fibo(n – 3)
fibo(n – 2) = fibo(n – 3) + fibo(n – 4)
.................................
................................
................................
fibo(2) = fibo(1) + fibo(0)
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 4–3 B (CS/IT-Sem-5)

In the first three steps, it can be clearly seen that fibo(n – 3) is calculated
twice. If we use recursion, we calculate the same sub-problems again
and again but with dynamic programming we calculate the sub-problems
only once.

Que 4.2. What is the principle of optimality ? Also give approaches

in dynamic programming.

Answer
Principle of optimality : Principle of optimality states that whatever the
initial state is, remaining decisions must be optimal with regard the state
following from the first decision.
Approaches in dynamic programming :
There are two approaches of solving dynamic programming problems :
1. Bottom-up approach : Bottom-up approach simply means storing the
results of certain calculations, which are then re-used later because the
same calculation is a sub-problem in a larger calculation.
2. Top-down approach : Top-down approach involves formulating a
complex calculation as a recursive series of simpler calculations.

Que 4.3. Discuss the elements of dynamic programming.

Answer
Following are the elements of dynamic programming :
1. Optimal sub-structure : Optimal sub-structure holds if optimal solution
contains optimal solutions to sub-problems. It is often easy to show the
optimal sub-problem property as follows :
i. Split problem into sub-problems.
ii. Sub-problems must be optimal; otherwise the optimal splitting would
not have been optimal.
There is usually a suitable “space” of sub-problems. Some spaces are
more “natural” than others. For matrix chain multiply we choose sub-
problems as sub-chains.
2. Overlapping sub-problem :
i. Overlapping sub-problem is found in those problems where bigger
problems share the same smaller problems. This means, while
solving larger problems through their sub-problems we find the
same sub-problems more than once. In these cases a sub-problem
is usually found to be solved previously.
ii. Overlapping sub-problems can be found in Matrix Chain
Multiplication (MCM) problem.
Dynamic Programming & Backtracking 4–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1..4

1..1 2..4 1..2 3..4 1..3 4..4

2..2 3..4 2..3 4..4 1..1 2..2 3..3 4..4 1..1 2..3 1..2 3..3

3..3 4..4 2..2 3..3 2..2 3..3 1..1 2..2

Fig. 4.3.1. The recursion tree for the computation
of Recursive-Matrix-Chain (p, 1, 4).
3. Memoization :
i. The memoization technique is the method of storing values of
solutions to previously solved problems.
ii. This generally means storing the values in a data structure that
helps us reach them efficiently when the same problems occur
during the program’s execution.
iii. The data structure can be anything that helps us do that but
generally a table is used.
Que 4.4. Write down an algorithm to compute Longest Common
Subsequence (LCS) of two given strings and analyze its time
complexity. AKTU 2017-18, Marks 10

Answer
LCS-Length (X, Y) :
1. m  length[X]
2. n  length[Y]
3. for i  1 to m
4. do c[i, 0]  0
5. for j  0 to n
6. do c[0, j]  0
7. for i  1 to m
8. do for j  1 to n
9. do if xi = yj
10. then c[i, j]  c[i – 1, j – 1] + 1
11. b[i, j]  “”
12. else if c[i – 1, j]  c[i, j – 1]
13. then c[i, j]  c[i – 1, j]
14. b[i, j] “”
15. else c[i, j]  c[i, j – 1]
16. b[i, j] “”
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 4–5 B (CS/IT-Sem-5)

17. return c and b

Note :
1. “” means both the same.
2. “” means c[i – 1, j]  c[i, j – 1].
3. “” means c[i – 1, j] < c[i, j – 1].
4. The “” diagonal arrows lengthen the LCS.
Since, two for loops are present in LCS algorithm first for loop runs upto m
times and second for loop runs upto n times. So, time complexity of LCS is
O(mn).

Que 4.5. Give the algorithm of dynamic 0/1-knapsack problem.

Answer
Dynamic 0/1-knapsack(v, w, n, W) :
1. for (w = 0 to W) V[0, w] = 0
2. for (i = 1 to n)
3. for (w = 0 to W)
4. if (w[i]  w) then
5. V[i, w] = max{V[i – 1, w], v[i] + V[i – 1, w – w[i]]};
6. else V[i, w] = V[i – 1, w];
7. return V[n, W];
Now, as we know that V [n, W] is the total value of selected items, the can
be placed in the knapsack. Following steps are used repeatedly to select
actual knapsack item.
Let, i = n and k = W then
while (1 > 0 and k > 0)
{
if (V[i, k]  V[i – 1, k]) then
mark ith item as in knapsack
i = i – 1 and k = k – wi // selection of ith item
else
i=i–1 //do not select ith item
}
Que 4.6. Differentiate between dynamic programming and greedy
approach. What is 0/1 knapsack problem ? Solve the following
instance using dynamic programming. Write the algorithm also.
Knapsack Capacity = 10, P = < 1, 6, 18, 22, 28 > and w = < 1, 2, 5, 6, 7>.
Dynamic Programming & Backtracking 4–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer

S. No. Dynamic programming Greedy approach

1. Solves eve ry optimal sub- Do not solve every optimal
problem. problem.
2. We make a choice at each step, We make whatever cho ice
but the choice may depend on seems best at the moment and
the solutions to sub-problem. then solve the problem.
3. It is bottom-up approach. It is top-down approach.
4. Dynamic programming works Greedy algorithm works when
when a problem has following a problem exhibits the following
properties : properties :
a. Optimal sub-structure a. Greedy choice property
b. Overlapping sub-problems b. Optimal sub-structure
0/1-knapsack problem : Refer Q. 3.18, Page 3–17B, Unit-3.
0/1-kanpsack algorithm : Refer Q. 4.5, Page 4–5B, Unit-4.
Numerical :
Item wi pi = vi/wi (Given) vi
I1 1 1 1
I2 2 6 12
I3 5 18 90
I4 6 22 132
I5 7 28 196
Now, fill the knapsack according to given value of pi.
First we choose item I1 whose weight is 1, then choose item I2 whose
weight is 2 and item I3 whose weight is 5.
 Total weight in knapsack : 1 + 2 + 5 = 8
Now, the next item is I4 and its weight is 6, but we want only 2
{ W = 10}. So we choose fractional part of it i.e.,
The value of fractional part of I4 is = 132/6 × 2 = 44
Thus the maximum value is = 1 + 12 + 90 + 44 = 147
2
5
= 10
2
1

Que 4.7. Discuss knapsack problem with respect to dynamic

programming approach. Find the optimal solution for given
problem, w (weight set) = {5, 10, 15, 20} and W (Knapsack size) = 25
and v = {50, 60, 120, 100}. AKTU 2015-16, Marks 10
 More pdf : www.motivationbank.in
4–7 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Answer
Knapsack problem with res pect to dynamic programming
approach : Refer Q. 4.5, Page 4–5B, Unit-4.
Numerical :
w = {5, 10, 15, 20}
W = 25
v = {50, 60, 120, 100}
Initially,
Item wi vi
I1 5 50
I2 10 60
I3 15 120
I4 20 100

Taking value per weight ratio, i.e., pi = vi/wi

Item wi vi pi = vi/wi
I1 5 50 10
I2 10 60 6
I3 15 120 8
I4 20 100 5

Now, arrange the value of pi in decreasing order.

Item wi vi pi = vi/wi
I1 5 50 10
I3 15 120 8
I2 10 60 6
I4 20 100 5
Now, fill the knapsack according to decreasing value of pi.
First we choose item I1 whose weight is 5, then choose item I3 whose
weight is 15. Now the total weight in knapsack is 5 + 15 = 20. Now, next
item is I2 and its weight is 10, but we want only 5. So, we choose fractional
part of it, i.e.,
Dynamic Programming & Backtracking4–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in

5
15 25

5
The value of fractional part of I2 is,
60
= × 5 = 30
10
Thus, the maximum value is,
= 50 + 120 + 3 = 200

Que 4.8. Compare the various programming paradigms such as

divide-and-conquer, dynamic programming and greedy approach.
AKTU 2019-20, Marks 07

Answer

S. No. Divide and Dynamic Greedy

conquer programming approach
approach approach
1. Optimizes by Same as Divide and Optimizes by
breaking down a Conquer, but optimizes making the best
subpro blem into by caching the answers choice at the
simpler versions of to each subproblem as moment.
itself and using multi- no t to repeat the
thre ading and calculation twice.
recursion to solve.
2. Always finds the Always finds the optimal Does not always
optimal solution, but solutio n, but canno t find the optimal
is slower than work on small datasets. solution, but is
Greedy. very fast.
3. Re quires some Requires a lot of memory Requires almost
me mo ry to for tabulation. no memory.
remember recursive
calls.

PART-2
All Pair Shortest Paths : Warshall’s and Floyd’s Algorithm,
Resource Allocation Problem.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms4–9 B (CS/IT-Sem-5)

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 4.9. Describe the Warshall’s and Floyd’s algorithm for

finding all pairs shortest paths.

Answer
1. Floyd-Warshall algorithm is a graph analysis algorithm for finding
shortest paths in a weighted, directed graph.
2. A single execution of the algorithm will find the shortest path between
all pairs of vertices.
3. It does so in (V3) time, where V is the number of vertices in the graph.
4. Negative-weight edges may be present, but we shall assume that there
are no negative-weight cycles.
5. The algorithm considers the “intermediate” vertices of a shortest path,
where an intermediate vertex of a simple path p = (v1, v2, ..., vm) is any
vertex of p other than v1 or vm, that is, any vertex in the set {v2, v3, ...,
vm–1}.
6. Let the vertices of G be V = {1, 2, ..., n}, and consider a subset {1, 2, ...,
k} of vertices for some k.
7. For any pair of vertices i, j  V, consider all paths from i to j whose
intermediate vertices are all drawn from {1, 2, ..., k}, and let p be a
minimum-weight path from among them.
8. Let dij( k) be the weight of a shortest path from vertex i to vertex j with
all intermediate vertices in the set {1, 2, ..., k}.
A recursive definition is given by
 wij if k  0
dij( k) =  ( k –1) ( k –1) ( k –1)
min( dij , dik  dkj ) if k  1
Floyd-Warshall (W) :
1. n  rows [W]
2. D(0)  W
3. for k  1 to n
4. do for i  1 to n
5. do for j  1 to n
( k –1)
6. do dij( k)  min( dij , dik( k–1)  dkj( k –1) )
7. return D(n)
 4–10 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
Que 4.10. Define Floyd Warshall algorithm for all pair shortest
path and apply the same on following graph :
3
1 2
4
6 3 5
1
1 2
3 4
2
Fig. 4.10.1.
AKTU 2019-20, Marks 07

Answer
Floyd Warshall algorithm : Refer Q. 4.9, Page 4–9B, Unit-4.
Numerical :
dij(k) = min[ dij(k–1) , dik
(k –1) ( k –1)
 dkj ]

 (ijk–1) if dij(k –1)  dik

( k –1) ( k 1)
 dkj
 (ijk) = 
( k1) ( k –1) ( k1) ( k1)
 kj if dij  dik  dkj

1 2 3 4 5
1 0  6 3 

2 3 0   
D(0) =
3   0 2 

4  1 1 0 

5  4  2 0

1 2 3 4 5
1 0 4 6 3 

2 3 0 9 6 
D(1) = 
3 6 4 0 2

4 4 1 1 0 

5 7 4 3 2 0
 More pdf : www.motivationbank.in
4–11 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

1 2 3 4 5
1 0 4 6 3 

2 3 0 7 6 
D(2) = 3 6 3 0 2 

4 4 1 1 0 

5 6 3 3 2 0

Now, if we find D(3), D(4) and D(5) there will be no change in the entries.

Que 4.11. Give Floyd-Warshall algorithm to find the shortest

path for all pairs of vertices in a graph. Give the complexity of the
algorithm. Explain with example. AKTU 2018-19, Marks 07

Answer
Floyd-Warshall algorithm : Refer Q. 4.9, Page 4–9B, Unit-4.
Time complexity of Floyd-Warshall algorithm is O(n3).
Example :
2
3 4
8
1 2 3
–4 1 –5
7
5 4
6
Fig. 4.11.1.

dij(k) = min[ dij(k–1) , dik

(k –1) ( k –1)
 dkj ]

 (ijk–1) if dij(k –1)  dik

( k –1) ( k 1)
 dkj
 (ijk) = 
( k1) ( k –1) ( k1) ( k1)
 kj if dij  dik  dkj

0 3 8   4  NIL 1 1 NIL 1 
 0  1 7  NIL NIL NIL 2 2 
  (0) 
D =  4
(0) 0   ;    NIL 3 NIL NIL NIL 
   
2  5 0  4 NIL 4 NIL NIL 
    6 0 NIL NIL NIL 5 NIL 

 4–12 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
0 3 8   4 NIL 1 1 NIL 1 
 0  1 7 NIL NIL NIL 2 2 
  (1) 
D(1) =   4 0   ;   NIL 3 NIL NIL NIL 
   
 2 5  5 0  2 4 1 4 NIL 1 
    6 0 NIL NIL NIL 5 NIL 


0 3 8 4  4 NIL 1 1 2 1 
 0  1 7  NIL NIL NIL 2 2 
  (2) 
D(2) =   4 0 5 11 ;   NIL 3 NIL 2 2 
   
 2 5  5 0  2 4 1 4 NIL 1 
    6 0 NIL NIL NIL 5 NIL 

0 3 8 4  4  NIL 1 1 2 1 
 0  1 7  NIL NIL NIL 2 2 
  (3) 
D(3) =   4 0 5 11 ;    NIL 3 NIL 2 2 
   
 2  1  5 0  2 4 3 4 NIL 1 
    6 0  NIL NIL NIL 5 NIL 

0 3  1 4  4  NIL 1 4 2 1 
3 0  4 1  1 4 NIL 4 2 1 
(4)
  (4) 
D = 7 4 0 5 3 ;    4 3 NIL 2 1 
   
 2  1  5 0  2 4 3 4 NIL 1 
 8 5 1 6 
0 4 3 4 5 NIL 

0 1  3 2  4  NIL 3 4 5 1 
3 0  4 1  1 4 NIL 4 2 1 
  (5) 
D(5) =  7 4 0 5 3 ;    4 3 NIL 2 1 
   
 2  1  5 0  2 4 3 4 NIL 1 
 8 5 1 6 0 4 3 4 5 NIL 


PART-3
Backtracking, Branch and Bound with Examples such as
Travelling Salesman Problem.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

 More pdf : www.motivationbank.in
Design and Analysis of Algorithms4–13 B (CS/IT-Sem-5)

Que 4.12. What is backtracking ? Write general iterative algorithm

for backtracking. AKTU 2016-17, Marks 10

Answer
1. Backtracking is a general algorithm for finding all solutions to some
computational problems.
2. Backtracking is an important tool for solving constraint satisfaction
problems, such as crosswords, verbal arithmetic, and many other
puzzles.
3. It is often the most convenient (if not the most efficient) technique for
parsing, for the knapsack problem and other combinational optimization
problem.
4. It can be applied only for problems which admit the concept of a ‘‘partial
candidate solution’’ and a relatively quick test of whether it can possibly
be completed to a valid solution.
Iterative backtracking algorithm :
algorithm ibacktrack (n)
// Iterative backtracking process
// All solutions are generated in x[1 : n] and printed as soon as they are
found
{
k = 1;
while (k != 0)
{
if ( there remains an untried x[k] in T(x[1], x[2], ..., x[k – 1])
and B_k(x[1], ..., x[k] ) is true )
{
if (x[1], ..., x[k] is a path to an answer node )
write (x[1 : k]);
k = k + 1; // Consider the next set
}
else
k = k – 1; // Backtrack to the previous set
}
}

Que 4.13. Des cribe backtracking algorithm for Travelling

Salesman Problem (TSP). Show that a TSP can be solved using
backtracking method in the exponential time.
Dynamic Programming & Backtracking 4–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
OR
Explain TSP (Travelling Salesman) problem with example. Write
an approach to solve TSP problem. AKTU 2015-16, Marks 10

Answer
Travelling Salesman Problem (TSP) :
Travelling salesman problem is the problem to find the shortest possible
route for a given set of cities and distance between the pair of cities that
visits every city exactly once and returns to the starting point.
Backtracking approach is used to solve TSP problem.
Backtracking algorithm for the TSP :
1. Let G be the given complete graph with positive weights on its edges.
2. Use a search tree that generates all permutations of V = {1 ... n},
specifically the one illustrated in Fig. 4.13.1 for the case n = 3.

1 2 3

1 2 3

2 3 1 3 1 2

1, 2 1,3 2,1 2,3 3,1 3,2

3 2 3 1 2 1

1,2,3 1,3,2 2,1,3 2,3,1 3,1,2 3,2,1

Fig. 4.13.1.
3. A node at depth i of this tree (the root is at depth 0) stores an
i-permutation of {1, ..., n}. A leaf stores a permutation of {1, ..., n}, which
is equivalent to saying that it stores a particular Hamiltonian cycle (tour)
of G.
4. For the travelling salesman problem, we will not do any static pruning
on this tree, we will do dynamic pruning, during the search.
Proof :
1. At some point during the search, let v be a non-leaf node of this tree that
is just being visited, and let w be the weight of the shortest tour found to
this point.
2. Let (π1π ... k) be the k-permutation of {1 ... n} stored at v.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 4–15 B (CS/IT-Sem-5)

Let, wv =  w ii1 denote the sum of the weights on edges whose

i 1.... k 1
endpoints are adjacent vertices in this k-permutation.
3. Then, if wv  w, the entire subtree of the search tree rooted at v can be
pruned, i.e., not searched at all, since every leaf of this subtree
represents a tour whose weight must be greater than wv.
4. This follows from the fact that all edges in the graph have positive
weights.
5. There are atmost O(n*2n) subproblem, and each one takes linear time
to solve.
6. The total running time is therefore O(n2*2n).
7. The time complexity is much less than O(n!) but still exponential.
Hence proved.

Que 4.14. Explain the working rule for Travelling Salesman

Problem using branch and bound technique.

Answer
For solving travelling salesman problem we represent the solution space
by a state space tree. We define three cost functions C, l and u where
li  Ci  ui for all the nodes i of the state space tree.
Step 1 : First find the cost matrix as given cost on the edges of the graph.
Step 2 : Now, find reduced matrix by subtracting the smallest element
from row i (column j) introduce a zero in to row i (column j). Repeating
this process as often as needed, the cost matrix can be reduced.
Add the total amount subtracted from the column and rows and make this
as the root of the state space tree.
Step 3 : Let M be the reduced cost matrix for node A and let B be a child of
A such that the tree edge (A, B) corresponds to inclusion of edge (i, j) in the
tour. If B is not a leaf, then the reduced cost matrix for B may be obtained
by apply following given steps :
a. Change all the entries in row i, column j of M to . This includes
use of any more edges leaving vertex i or entering vertex j.
b. Set M (j, 1) = . This excludes use of the edge (j, 1).
c. Reduce all the rows and columns in the resulting matrix except
for rows and columns containing only .
Suppose T is the total amount subtracted in step (c), then
I(B) = l(A) + M(i, j) + T
Dynamic Programming & Backtracking4–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
For leaf nodes l = c is easily computed as each leaf defines a unique tour.
For the upper bound u, we assume ui =  for all nodes i.
Step 4 : Find the root of the node as, combine the total amount subtracted
from cost matrix to find the reduced cost matrix M.

Que 4.15. What is travelling salesman problem ? Find the solution

of following travelling salesman problem using branch and bound
method.
  20 30 10 11
15  16 4 2 
 
Cost matrix =  3 5  2 4
 
19 6 6  3
16 4 7 16  

AKTU 2016-17, Marks 10

Answer
Travelling salesman problem : Refer Q. 4.13, Page 4–13B, Unit-4.
Numerical :
  20 30 10 11
15  16 4 2 

Cost matrix =  3 5  2 4
 
19 6 6  3
16 4 7 16  

1. Reduce each column and row by reducing the minimum value from
each element in row and column.
Row Column
1 3

10  10 20 0 1  10 17 0 1
2 13  14 2 0 12  11 2 0
2 1 3  0 2 0 3  0 2 =M
1
3 16 3 3  0 15 3 0  0
4 12 0 3 12  11 0 0 12 
2. So, total expected cost is : 10 + 2 + 2 + 3 + 4 + 1 + 3 = 25.
3. We have discovered the root node V1 so the next node to be expanded
will be V2, V3, V4, V5. Obtain cost of expanding using cost matrix for node
2.
4. Change all the elements in 1st row and 2nd column.
 More pdf : www.motivationbank.in
4–17 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

     
12  11 2 0 
 
M2 =  0   0 2 
 
15  0  0 
11  0 12  
 
5. Now, reducing M2 in rows and columns, we get :
     
12  11 2 0 
 
M2 =  0   0 2 
 
15  0  0 
11  0 12  
 
 Total cost for M2 = 25 + 10 + 0 = 35
6. Similarly, for node 3, we have :
     
12   2 0 
 
M3 =  0 3  0 2 
 
15 3   0 
11 0  12  
 
7. Now, reducing M3, we get :
     
12   2 0 
 
M3 =  0 3  0 2 
 
15 3   0 
11 0  12  
 
 Total cost for M3 = 25 + 17 + 0 = 42
8. Similarly, for node 4, we have :
     
12  11  0 
 
M4 =  0 3   2 
 
15 3 0  0 
11 0 0   
 
9. Now, reducing M4, we get :
     
12  11  0 
 
M4 =  0 3   2 
 
15 3 0  0 
11 0 0   
 
 4–18 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
 Total cost = 25 + 0 + 0 = 25
10. Similarly, for node 5, we have :
    
12  11 2  

M5 =  0 3  0 
 
15 3 0  
11 0 0 12  

11. Now, reducing M5, we get :
    
10  9 0  

M5 =  0 3  0 
 
15 3 0  
11 0 0 12  

 Total cost = 25 + 1 + 2 = 28

1 V1 = 25

2 3 4 5 V 5 = 28

V2 = 35 V3 = 42 V4 = 25
Fig. 4.15.1.

12. Now, the promising node is V4 = 25. Now, we can expand V2, V3 and V5.
Now, the input matrix will be M4.
13. Change all the elements in 4th row and 2nd column.
    
12  11  0 

M6 =  0    2
 
    
11  0   

14. On reducing M6, we get :
    
12  11  0 

M6 =  0    2
 
    
11  0   

 More pdf : www.motivationbank.in
4–19 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

 Total cost = 25 + 3 + 0 = 28
15. Similarly, for node 3, we have :
     
12    0 
 
M7 =  0 3   2 
 
  3   0
11 0    
 
16. On reducing M7, we get :
    
12    0 

M7 =  0 3   2
 
 3   0
11 0    

 Total cost = 25 + 0 + 0 = 25
17. Similarly, for node 5, we have :
     
12  11   
 
M8 =  0 3    
 
     
11 0 0   
 
18. On reducing M8, we get :
     
 1  0  
 
M8 =  0 3    
 
     
11 0 0   
 
 Total cost = 25 + 0 + 11 = 36

2 3 4 5

2 3 5

V2 = 28 V 3 = 25 V 5 = 36
Fig. 4.15.2.
19. Now, promising node is V3 = 25. Now, we can expand V2 and V5. Now,
the input matrix will be M7.
 4–20 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
20. Change all the elements in 3rd row and 2nd column.
    
12    0 

M9 =      
 
    0
11     

21. On reducing M9, we get :
    
1    0 

M9 =      
 
    0
0     

 Total cost = 25 + 3 + 0 = 28
22. Similarly, for node 5, we have :
    
12     

M10 =      
 
 3   
11 0    

23. On reducing M10, we get :
    
 0     

M10 =      
 
 0   
11 0    

 Total cost = 25 + 2 + 12 + 3 = 42.
1

2 3 4 5

2 3 5

2 5

V 2 = 28 V 5 = 42
Fig. 4.15.3.
 More pdf : www.motivationbank.in
4–21 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

24. Here V2 is the most promising node so next we are going to expand this
node further. Now, we are left with only one node not yet traversed
which is V5.
10 6 5 2 16
V1  V4  V3  V2  V5  V1
So, total cost of traversing the graph is :
10 + 6 + 5 + 2 + 16 = 39

PART-4
Graph Colouring, N-Queen Problem.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 4.16. Write short notes on graph colouring.

Answer
1. Graph colouring is a simple way of labelling graph components such as
vertices, edges, and regions under some constraints.
2. In a graph, no two adjacent vertices, adjacent edges, or adjacent regions
are coloured with minimum number of colours. This number is called
the chromatic number and the graph is called a properly coloured
graph.
3. While graph colouring, the constraints that are set on the graph are
colours, order of colouring, the way of assigning colour, etc.
4. A colouring is given to a vertex or a particular region. Thus, the vertices
or regions having same colours form independent sets.
For example :
Blue

Yellow Red

Blue Yellow
Fig. 4.16.1. Properly coloured graph.

Que 4.17. Write short notes on N-Queens problem.

Dynamic Programming & Backtracking4–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
OR
Write pseudocode for 8-Queens problem.
AKTU 2016-17, Marks 10

Answer
1. In N-Queens problem, the idea is to place queens one by one in different
columns, starting from the leftmost column.
2. When we place a queen in a column, we check for clashes with already
placed queens.
3. In the current column, if we find a row for which there is no clash, we
mark this row and column as part of the solution.
4. If we do not find such a row due to clashes then we backtrack and
return false.
Procedure for solving N-Queens problem :
1. Start from the leftmost column.
2. If all queens are placed return true.
3. Try all rows in the current column. Do following for every tried row :
a. If the queen can be placed safely in this row then mark this [row,
column] as part of the solution and recursively check if placing
queen here leads to a solution.
b. If placing queen in [row, column] leads to a solution then return
true.
c. If placing queen does not lead to a solution then unmark this [row,
column] (backtrack) and go to step (a) to try other rows.
4. If all rows have been tried and nothing worked, return false to trigger
backtracking.
Algorithm/pseudocode for N-Queens problem :
N-Queens are to be placed on an n × n chessboard so that no two attack i.e.,
no two Queens are on the same row, column or diagonal.
PLACE (k, i)
1. for j  1 to k – 1
2. do if (x(j) = i) or Abs (x[j] – i) = (Abs (j – k))
3. then return false
4. return true
Place (k, i) returns true if a queen can be placed in the kth row and ith column
otherwise return false.
x[ ] is a global array whose first k – 1 values have been set. Abs(r) returns the
absolute value of r.
N-Queens (k, n)
 More pdf : www.motivationbank.in
4–23 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

1. for i  1 to n
2. do if PLACE (k, i)
3. then x[k]  i
4. if k = n, then print x[1 …. N]
5. else N-Queens (k + 1, n)
[Note : For 8-Queen problem put n = 8 in the algorithm.]

Que 4.18. Write an algorithm for solving N-Queens problem.

Show the solution of 4-Queens problem using backtracking
approach. AKTU 2015-16, Marks 10

Answer
Algorithm for N-Queens problem : Refer Q. 4.17, Page 4–21B, Unit-4.
4-Queens problem :
1. Suppose we have 4 × 4 chessboard with 4-queens each to be placed in
non-attacking position.
1 2 3 4
1
2
3
4

Fig. 4.18.1.
2. Now, we will place each queen on a different row such that no two
queens attack each other.
3. We place the queen q1 in the very first accept position (1, 1).
4. Now if we place queen q2 in column 1 and 2 then the dead end is
encountered.
5. Thus, the first acceptable position for queen q2 is column 3 i.e., (2, 3) but
then no position is left for placing queen q3 safely. So, we backtrack one
step and place the queen q2 in (2, 4).
6. Now, we obtain the position for placing queen q3 which is (3, 2). But later
this position lead to dead end and no place is found where queen q2 can
be placed safely.
 4–24 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
1 2 3 4
1 q1
2 q2

3 q3

Fig. 4.18.2.

7. Then we have to backtrack till queen q1 and place it to (1, 2) and then all
the other queens are placed safely by moving queen q2 to (2, 4), queen q3
to (3, 1) and queen q4 to (4, 3) i.e., we get the solution < 2, 4, 1, 3>. This
is one possible solution for 4-queens problem.
1 2 3 4
1 q1

2 q2

3 q3

4 q4

Fig. 4.18.3.

8. For other possible solution the whole method is repeated for all partial
solutions. The other solution for 4-queens problem is <3, 1, 4, 2> i.e.,
1 2 3 4
1 q1
2 q
2

3 q3

4 q4

Fig. 4.18.4.

9. Now, the implicit tree for 4-queen for solution <2, 4, 1, 3> is as follows :
10. Fig. 4.18.5 shows the complete state space for 4-queens problem. But we
can use backtracking method to generate the necessary node and stop
if next node violates the rule i.e., if two queens are attacking.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 4–25 B (CS/IT-Sem-5)

1 5
q1 q1

× × × × ×
2 3
q1 q1 q1
q2 q2 q2

q1 q1
q2 q2
q3 q3

× × q1
q2
q3
q4

Fig. 4.18.5.

Que 4.19. What is branch and bound technique ? Find a solution

to the 4-Queens problem using branch and bound strategy. Draw
the solution space using necessary bounding function.

Answer
Branch and bound technique :
1. It is a systematic method for solving optimization problems.
2. It is used where backtracking and greedy method fails.
3. It is sometimes also known as best first search.
4. In this approach we calculate bound at each stage and check whether it
is able to give answer or not.
5. Branch and bound procedure requires two tools :
a. The first one is a way of covering the feasible region by several
smaller feasible sub-regions. This is known as branching.
b. Another tool is bounding, which is a fast way of finding upper and
lower bounds for the optimal solution within a feasible sub-region.
Solution to 4-Queens problem :
Basically, we have to ensure 4 things :
1. No two queens share a column.
 4–26 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
2. No two queens share a row.
3. No two queens share a top-right to left-bottom diagonal.
4. No two queens share a top-left to bottom-right diagonal.
Number 1 is automatic because of the way we store the solution. For number
2, 3 and 4, we can perform updates in O(1) time and the whole updation
process is shown in Fig. 4.19.1.

5
Q Q
1

2 3 6
Q Q Q
Q Q Q

4 7
Q Q
Q Q
Q Q

8
Q
Q
Q
Q
Fig. 4.19.1.

PART-5
Hamiltonian Cycles and Sum of Subsets.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 4.20. What is backtracking ? Discuss sum of subset problem

with the help of an example. AKTU 2017-18, Marks 10

 More pdf : www.motivationbank.in
4–27 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Answer
Backtracking : Refer Q. 4.12, Page 4–13B, Unit-4.
Sum of subset problem with example :
In the subset-sum problem we have to find a subset s of the given set
S = (S1, S2, S3, ......, Sn) where the elements of the set S are n positive integers
in such a manner that sS and sum of the elements of subset ‘s’ is equal to
some positive integer ‘X’.
Algorithm for sum-subset problem :
Subset-Sum (S, t)
1. C  
2. Z  S
3. K  
4. t1  t
5. while (Z  ) do
6. K  max(Z)
7. if (K < t) then
8. ZZ–K
9. t1  t1 – K
10. CCK
11. else Z  Z – K
12. print C // Subset Sum elements whose
// Sum is equal to t1
This procedure selects those elements of S whose sum is equal to t. Every
time maximum element is found from S, if it is less than t then this element
is removed from Z and also it is subtracted from t.
For example :
Given S = <1, 2, 5, 7, 8, 10, 15, 20, 25> & m = 35
Z  S, m = 35
k  max[Z] = 25
K<m
 Z=Z–K
i.e., Z = <1, 2, 5, 7, 8, 10, 15, 20> & m1  m
Subtracting K from m1, we get
New m1 = m1(old) – K = 35 – 25 = 10
In new step,
K  max[Z] = 20
K > m1
i.e., Z = <1, 2, 5, 7, 8, 10, 15>
In new step,
 4–28 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
K  max[Z] = 15
K > m1
i.e., Z = <1, 2, 5, 7, 8, 10>
In new step,
K  max[Z] = 10
K > m1
i.e., Z = <1, 2, 5, 7, 8>
In new step,
K  max[Z] = 8
K > m1
i.e., Z = <1, 2, 5, 7> & m2  m1
New m2 = m2(old) – K = 10 – 8 = 2
In new step
K  max[Z] = 7
K > m2
i.e., Z = <1, 2, 5>
In new step, K  max[Z] = 5
K > m2
i.e., Z = <1, 2>
In new step, K  max[Z] = 2
K > m2
i.e., Z = <1>
In new step,
K=1
K < m2
 m3 = 01
Now only those numbers are needed to be selected whose sum is 01, therefore
only 1 is selected from Z and rest other number found as max[Z] are subtracted
from Z one by one till Z become .

Que 4.21. Solve the subset sum problem using backtracking, where

n = 4, m = 18, w[4] = {5, 10, 8, 13}. AKTU 2018-19, Marks 07

Answer
n=4
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 4–29 B (CS/IT-Sem-5)

m = 18
w{4} = {5, 10, 8, 13}
Sorted order : w{4} = {5, 8, 10, 13}
Now we construct state-space tree.
I = Include
I1 0 E1 E = Exclude

5 0
I2 I2 E2
E2 8 E3 0
15
I3 ×
I3
5 8
23 18
E3 ×
× E3 I3 
15 5
× I4 E4
15
I4 E4 18 5
28 15  ×
× ×
First Subset S1 = {5, 13}
Similarly, S2 = {8, 10}

Que 4.22. Differentiate between backtracking and branch and

bound approach. Write an algorithm for sum-subset problem
using backtracking approach.

Answer
Difference between backtracking and branch and bound technique :
S. No. Backtracking Branch and bound
1. It is a methodological way of It is a systematic method for
trying out various sequences of solving optimization problems.
decisions.
2. It is applied in dynamic It is applied where greedy and
programming technique. dynamic programming technique
fails.
3. It is sometimes called depth first It is also known as best first
search. search.
4. This approach is effective for This approach is effective for
decision problem. optimization problems.
5. This algorithm is simple and easy This algorithm is difficult to
to understand. understand.

Algorithm for sum-subset problem : Refer Q. 4.20, Page 4–26B, Unit-4.

 4–30 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
Que 4.23. Explain Hamiltonian circuit problem. Consider a graph
G = (V, E) shown in Fig. 4.23.1 and find a Hamiltonian circuit using
backtracking method.
b

a e
c
f
d
Fig. 4.23.1.

Answer
Hamiltonian circuit problem :
1. Given a graph G = (V, E), we have to find the Hamiltonian circuit using
backtracking approach.
2. We start our search from any arbitrary vertex, say ‘a’. This vertex ‘a’
becomes the root of our implicit tree.
3. The first element of our partial solution is the first intermediate vertex
of the Hamiltonian cycle that is to be constructed.
4. The next adjacent vertex is selected on the basis of alphabetical (or
numerical) order.
5. If at any stage any arbitrary vertex makes a cycle with any vertex other
than vertex ‘a’ then we say that dead end is reached.
6. In this case we backtrack one step, and again search begins by selecting
another vertex and backtrack the element from the partial solution
must be removed.
7. The search using backtracking is successful if a Hamiltonian cycle is
obtained.
Numerical :
1. Firstly, we start our search with vertex ‘a’, this vertex ‘a’ becomes the
root of our implicit tree.
a Root
2. Next, we choose vertex ‘b’ adjacent to ‘a’ as it comes first in lexicographical
order (b, c, d).
a Root

b c d
3. Next, we select ‘c’ adjacent to ‘b’
 More pdf : www.motivationbank.in
4–31 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

a Root

b c d

c e

4. Next, we select ‘d’ adjacent to ‘c’

a Root

b c d

c e

d e

5. Next, we select ‘e’ adjacent to ‘d’

a Root

b c d

c e

d e

e f

6. Next, we select vertex ‘f ’ adjacent to ‘e’. The vertex adjacent to ‘f ’ are ‘d’
and ‘e’ but they have already visited. Thus, we get the dead end and we
backtrack one step and remove the vertex ‘f ’ from partial solution.
 4–32 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
a Root

b c d

c e

d e

e f

f
Dead end
7. From backtracking, the vertex adjacent to ‘e’ are ‘b’, ‘c’, ‘d’, ‘f ’ from which
vertex ‘f ’ has already been checked and ‘b’, ‘c’, ‘d’ have already visited.
So, again we backtrack one step. Now, the vertex adjacent to ‘d’ are ‘e’,
‘f ’ from which ‘e’ has already been checked and adjacent of ‘f ’ are ‘d’ and
‘e’. If ‘e’ vertex visited then again we get dead state. So again we backtrack
one step.
8. Now, adjacent to ‘c’ is ‘e’ and adjacent to ‘e’ is ‘f ’ and adjacent to ‘f ’ is ‘d’
and adjacent to ‘d’ is ‘a’. Here, we get the Hamiltonian cycle as all the
vertex other than the start vertex ‘a’ is visited only once (a – b – c – e – f
– d – a).

a Root

b c d
Backtrack
c e

d e

e f

f c b d
Dead end
 More pdf : www.motivationbank.in
4–33 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

a Root

b c d
Backtrack

c e

d e

e f
Dead end

d e
Dead end

Again backtrack
0
a Root

1
b c d
Backtrack
2
c e

d e 3
Dead end
f 4 d

d 5 e

a Solution

9. Here we have generated one Hamiltonian circuit but other Hamiltonian

circuit can also be obtained by considering other vertex.
 4–34 B (CS/IT-Sem-5)
Dynamic Programming & Backtracking
More pdf : www.motivationbank.in
VERY IMPORTANT QUESTIONS
Following questions are very important. These questions
may be asked in your SESSIONALS as well as
UNIVERSITY EXAMINATION.

Q. 1. What do you mean by dynamic programming ?

Ans. Refer Q. 4.1.

Q. 2. Write down an algorithm to compute longest common

subsequence (LCS) of two given strings.
Ans. Refer Q. 4.4.

Q. 3. Give the algorithm of dynamic 0/1-knapsack problem.

Ans. Refer Q. 4.5.

Q. 4. Describe the Warshall’s and Floyd’s algorithm for finding

all pair shortest paths.
Ans. Refer Q. 4.9.

Q. 5. What is backtracking ? Write general iterative algorithm

for backtracking.
Ans. Refer Q. 4.12.

Q. 6. Write short notes on N-Queens problem.

Ans. Refer Q. 4.17.

Q. 7. Explain travelling salesman problem using backtracking.

Ans. Refer Q. 4.13.

Q. 8. Explain TSP using branch and bound technique.

Ans. Refer Q. 4.14.

Q. 9. Write an algorithm for sum of subset problem.

Ans. Refer Q. 4.20.

Q. 10. Compare the various programming paradigms such as

divide-and-conquer, dynamic programming and greedy
approach.
Ans. Refer Q. 4.8.


 More pdf : www.motivationbank.in
5–1 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

5 Selected Topics

CONTENTS
Part-1 : Algebraic Computation, ..................... 5–2B to 5–4B
Fast Fourier Transform

Part-2 : String Matching ................................. 5–4B to 5–13B

Part-3 : Theory of NP-Completeness .......... 5–13B to 5–24B

Part-4 : Approximation Algorithm .............. 5–24B to 5–30B

Part-5 : Randomized Algorithm .................. 5–30B to 5–32B

Selected Topics 5–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
PART-1
Algebraic Computation, Fast Fourier Transform.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 5.1. What is FFT (Fast Fourier Transformation) ? How the

recursive FFT procedure works ? Explain.

Answer
1. The Fast Fourier Transform (FFT) is a algorithm that computes a
Discrete Fourier Transform (DFT) of n-length vector in O(n log n)
time.
2. In the FFT algorithm, we apply the divide and conquer approach to
polynomial evaluation by observing that if n is even, we can divide a
degree (n – 1) polynomial.
A(x) = a0 + a1x + a2x2 +... + an–1 xn–1
n 
into two degree  – 1 polynomials.
2 
A[0](x) = a0 + a2x + a4x2 + ... + an–2 xn/2–1
A[1](x) = a1 + a3x + a5x2 + ... + an–1 xn/2–1
Where A[0] contains all the even index coefficients of A and A[1] contains
all the odd index coefficients and we can combine these two polynomials
into A, using the equation,
A(x) = A[0] (x2) + xA [1] (x2) ...(5.1.1)
So that the problem of evaluating A(x) at kn where k = 0, 1, 2, ..., n–1
reduces to,
n 
i. Evaluating the degree  – 1 polynomial A[0](x) and A[1](x) at the
2 
point (kn)2 i.e.,
(0n)2, (1n)2, ..., ( n 1 2
n )
because we know that if kn is a complex nth root of unity then (kn)2 is
n th
a complex root of unity. Thus, we can evaluate each A[0](x) and
2
A[1](x) at (kn)2 values.
ii. Combining the results according to the equation (5.1.1). This
observation is the basis for the following procedure which computes
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 5–3 B (CS/IT-Sem-5)

the DFT of an n-element vector a = (a0, a1,..., an–1) where for sake of
simplicity, we assume that n is a power of 2.
FFT (a, w) :
1. n  length [a] n is a power of 2.
2. if n = 1
3. then return a
4. n e2i/n
5. x  0 x will store powers of  initially x = 1.
6. a[0]  (a0, a2,...an–2)
7. a[1]  (a1, a3, ... an–1)
8. y[0]  FFT(a[0], 2) Recursive calls with 2 as (n/2)th root of unity.
9. y[1]  FFT(a[0], 2)
10. for k  0 to (n/2) – 1
11. do yk  y[0] [1]
k + x yk
12. yk+(n/2)  y[0] – xy[1]
k k
13. x  xn
14. return y
Line 2-3 represents the basis of recursion; the DFT of one element is the
element itself. Since in this case
y0 = a0 10 = a0 1 = a0
Line 6-7 defines the recursive coefficient vectors for the polynomials A[0]
and A[1].  = kn
Line 8-9 perform the recursive DFTn/2 computations setting for
n
k = 0, 1, 2, ...,
–1 i.e.,
2
y[0] [0] 2k
k = A ( n ) , y[1] [1]
k = A ( n )
2k

Lines 11-12 combine the results of the recursive DFTn/2 calculations.

For y0, y2, ... y(n/2)–1, line 11 yields.
yk = y[0] k [1]
k   n yk

= A[0](2nk )   kn A[1]( 2nk ) = A( kn ) using equation (5.1.1)

For yn/2, y(n/2)+1 ... yn–1, line 12 yields.
yk+(n/2) = y[0] k [1] [0] k ( n /2) [1]
k –  n yk = yk   n yk [  kn ( n / 2)  –  nk ]
= A[0]( 2nk )   kn (n / 2) A[1]( 2nk )
= A[0]( 2nk  nn )   nk  (n /2) A[1]( 2nk  nn ) [  nn  1]
= A[0]( 2nk  n )   kn ( n/ 2) A[1] ( 2nk n )

= A( nk (n / 2) ) using equation (5.1.1)

each k = 0, 1, 2, ...,(n/2) – 1.
Thus, the vector y returned by the FFT algorithm will store the values of
A(x) at each of the roots of unity.
Selected Topics 5–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 5.2. What is the application of Fast Fourier Transform
(FFT) ? Also write the recursive algorithm for FFT.
AKTU 2018-19, Marks 10

Answer
Application of Fast Fourier Transform :
1. Signal processing.
2. Image processing.
3. Fast multiplication of large integers.
4. Solving Poisson’s equation nearly optimally.
Recursive algorithm : Refer Q. 5.1, Page 5–2B, Unit-5.

PART-2
String Matching.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 5.3. What is string matching ? Discuss string matching

problem. Also define string, substring and proper substring.

Answer
String matching is a process of finding one or more occurrences of a pattern
in a text.
String matching problem :
Given a text array T[1 .. n] of n character and a pattern array P[1 .. m] of m
characters.
The problem is to find an integer s, called valid shift where 0  s < n – m and
T[s + 1 .... s + m] = P[1 ... m].
We further assume that the elements of P and T are characters drawn from
a finite alphabet such as {0, 1} or {A, B, … Z, a, b, …, z}.
String : A string is traditionally a sequence of character, either as a literal
constant or as some kind of variable.
Substring : Given a string T[1 .. n], the substring is defined as T[i .. j] for
some 0  i  j  n – 1, that is, the string formed by the characters in T from
index j, inclusive. This means that a string is a substring of itself (simply take
i = 0 and j = m).
Proper substring : The proper substring of string T[1 .. n] is T[i .. j] for
some 0  i  j  n – 1, that is, we must have either i > 0 or j < m – 1.
Using these definition, we can say given any string T[1 .. n], the substring
are
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms5–5 B (CS/IT-Sem-5)

T[i .. j] = T[i] T[i + 1] T[i + 2] … T[j]

for some 0  i  j  n – 1.
And proper substrings are
T[i .. j] = T[i] T[i + 1] T[i + 2] … T[j]
for some 0  i  j  n – 1.
Note that if i > j, then T[i .. j] is equal to the empty string or null, which has
length zero. Using these notations, we can define of a given string T[1 .. n] as
T[0 .. i] for some 0  i  n – 1 and suffix of a given string T[1 – n] as
T[i .. n –1] for some 0  i  n –1.
Que 5.4. What are the different types of string matching ? Explain
one of them.

Answer
Basic types of string matching algorithms are :
String matching algorithm

Naive string matching Rabin-Karp KMP (Knuth-Morris-Pratt)

algorithm algorithm algorithm

Fig. 5.4.1.
Naive string matching :
The Naive approach simply test all the possible placement of pattern P[1 .. m]
relative to text T[1 .. n]. Specifically, we try shifts s = [0, 1, …., n – m],
successively and for each shift, s, compare T[s + 1 .. s + m] to P[1 .. m].
Naive string matcher (T, P)
1. n  length [T]
2. m  length [P]
3. for s  0 to n – m
4. do if P[1 .. m] = T[s +1 .. s + m]
5. then print “pattern occurs with shift” s.
The Naive string matching procedure can be interpreted graphically as a
sliding a pattern P[1 .. m] over the text T[1 .. m] and noting for which shift all
of the characters in the pattern match the corresponding characters in the
text.
To analyze the time of Naive matching, the given algorithm is implemented
as follows, note that in this implementation, we use notation P[1 .. j] to
denote the substring of P from index i to index j. That is, P[1 ... j] = P[i]
P[i + 1] … P[j].
Naive string matcher (T, P)
1. n  length [T]
2. m  length [P]
3. for s  0 to n–m do
4. j1
Selected Topics 5–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
5. while j  m and T[s + j] = P[j] do
6. j  j +1
7. if j > m then
8. return valid shift s
9. return no valid shift exist // i.e., there is no substring of T matching P.

Que 5.5. Show the comparisons that Naive string matcher makes
for the pattern P = {10001} in the text T = {0000100010010}

Answer
Given, P = 10001
T = 0000100010010
i. 0 0 0 0 1 0 0 0 1 0 0 1 0
S=1

1 0 0 0 1

ii 0 0 0 0 1 0 0 0 1 0 0 1 0

S=1

1 0 0 0 1

iii. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=2
1 0 0 0 1

iv. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=3
1 0 0 0 1

v. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=4

1 0 0 0 1

vi. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=5

1 0 0 0 1
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms5–7 B (CS/IT-Sem-5)

vii. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=6

1 0 0 0 1

viii. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=7

1 0 0 0 1

ix. 0 0 0 0 1 0 0 0 1 0 0 1 0

S=8

1 0 0 0 1

Que 5.6. Write down Knuth-Morris-Pratt algorithm for string

matching. Find the prefix function of the string ababababca.

Answer
Knuth-Morris-Pratt algorithm for string matching :
COMPUTE-PREFIX-FUNCTION (P)
1. m  length [P]
2. [1]  0
3. k  0
4. for q  2 to m
5. do while k > 0 and P [k + 1]  P[q]
6. do k  [k]
7. if P[k + 1] = P [q]
8. then k  k + 1
9. [q]  k
10. return 
KMP-MATCHER calls the auxiliary procedure COMPUTE-PREFIX-
FUNCTION to compute .
KMP-MATCHER (T, p)
1. n  length [T]
2. m  length [P]
3.  COMPUTE-PREFIX-FUNCTION (P)
4. q  0
5. for i  1 to n
6. do while q > 0 and P [q + 1]  T[i]
7. do q  [q]
8. if P [q +1] = T[i]
Selected Topics 5–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
9. then q  q + 1
10. if q = m
11. then print “pattern occurs with shift” i – m
12. q  [q]
Prefix function of the string ababababca :
m  length [P]
 m = 10
Initially, P[1] = 0, k = 0
for q  2 to 10
for q = 2, k  0
& P[0 + 1] = P[2]
 [2] = 0
for q = 3, k  0
& P[0 + 1] = P[3]
 k  k + 1= 1
& [3] 1
for q = 4, k > 0
& P[1 + 1] = P[4]
 k1+1= 2
& [4]  2
for q = 5, k > 0
& P[2 + 1] = P[5]
 k2+1= 3
& [5]  3
for q = 6, k > 0
& P[3 + 1] = P[6]
 k3+1= 4
& [6]  4
for q = 7, k > 0
& P[4 + 1] = P[7]
 k4+1= 5
& [7]  5
for q = 8, k > 0
& P[5 + 1] = P[8]
 k5+1= 6
& [8]  6
for q = 9, k > 0
& P[6 + 1] = P[9]
 k  [k] = 6
& [9]  6
for q = 10, k > 0
& P[6 + 1] = P[10]
 k6+1= 7
& [10]  7
 More pdf : www.motivationbank.in
5–9 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

String a b a b a b a b c a

P[i] 1 2 3 4 5 6 7 8 9 10

[i] 0 0 1 2 3 4 5 6 6 7

Que 5.7. Compute the prefix function  for the pattern P = abacab
using KNUTH-MORRIS-PRATT algorithm. Also explain Naive string
matching algorithm. AKTU 2017-18, Marks 10

Answer
Prefix function of the string abacab :
m  length [P]
 m=6
Initially, [1] = 0, k = 0
for q  2 to 6
for q = 2, k  0
& P[0 + 1]  P[2]
 [2] = 0
for q = 3, k  0
& P[0 + 1] = P[3]
 k= k+1=1
& [3] = 1
for q = 4, k > 0
& P[1 + 1]  P[4]
 k  [1] = 0
P[1]  P[4]
& [4] = 0
for q = 5, k > 0
& P[0 + 1] = P[5]
 k0+1= 1
& [5] = 1
for q = 6, k > 0
& P[1 + 1] = P[6]
 k1+1= 2
& [6] = 2

String a b a c a b
P[i] 1 2 3 4 5 6
[i] 0 0 1 0 1 2
Naive string matching algorithm : Refer Q. 5.4, Page 5–5B, Unit-5.
Selected Topics 5–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Que 5.8. Describe in detail Knuth-Morris-Pratt string matching
algorithm. Compute the prefix function  for the pattern
ababbabbabbababbabb when the alphabet is  = {a, b}.
AKTU 2019-20, Marks 07

Answer
Knuth-Morris-Pratt string matching algorithm : Refer Q. 5.6,
Page 5–7B, Unit-5.
Numerical :
pattern = ababbabbabbababbabb
length 19
Initially, (1) = 0 and k = 0
For q  2 to 9
For q = 2 and k / 0
P[0 + 1]  P[2].
[2] = 0
For q= 3
P[0 + 1] = P[3]
k= k+1=1
[3] = 1
For q= 4k>0
P[1 + 1] = P[4]
P[2] = P[4]
k= k+1=2
[4] = 2
For q= 5k>0
P[2 + 1]  P[5]
[5] = 0
For q= 6
P[0 + 1] = P[6]
k= k+1=1
[6] = 1
For q= 7k>0
P[1 + 1] = P[7]
P[2] = P[7]
k= k+1=2
[7] = 0
For q= 8k>0
P[2 + 1] = P[8]
P[3]  P[8]
[8] = 0
For q= 9
P[0 + 1] = P[9]
k= k+1
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms5–11 B (CS/IT-Sem-5)

[9] = 2
For q= 10 k > 0
P[2 + 1]  P[10]
[10] = 0
For q= 11 k > 0
P[0 + 1]  P[11]
[11] = 0
For q= 12 k > 0
P[0 + 1]  P[12]
k= k+1
[12] = 2
For q= 13 k > 0
P[2 + 1]  P[13]
[13] = 0
For q= 14
P[0 + 1] = P[14]
k= k+1
[14] = 2
For q= 15 k > 0
P[2 + 1]  P[15]
[15] = 0
For q= 16 k > 0
P[0 + 1]  P[16]
[16] = 0
For q= 17
P[0 + 1]  P[17]
k= k+1
[17] = 2
For q= 18 k > 0
P[2 + 1]  P[18]
[18] = 0
For q= 19
P[0 + 1]  P[19]
[19] = 0
String a b a b b a b b a b b a b a b b a b b
P[i] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
[i] 0 0 1 2 0 1 2 0 2 0 0 2 0 2 0 0 2 0 0

Que 5.9. Write algorithm for Rabin-Karp method. Give a suitable

example to explain it.
OR
What is string matching algorithm ? Explain Rabin-Karp method
with examples. AKTU 2015-16, Marks 10
Selected Topics 5–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer
String matching algorithm : Refer Q. 5.3, Page 5–4B, Unit-5.
The Rabin-Karp algorithm :
The Rabin-Karp algorithm states that if two strings are equal, their hash
values are also equal. This also uses elementary number-theoretic notions
such as the equivalence of two numbers module a third.
Rabin-Karp-Matcher (T, P, d, q)
1. n length [T]
2. m length [P]
3. h dm–1 mod q
4. p  0
5. to  0
6. for i  1 to m
7. do p  (dp + p [i]) mod q
8. t0  (dt0 + T[i]) mod q
9. for s  0 to n–m
10. do if p = ts
11. then if p [1....m] = T[s + 1......s + m]
12. then “pattern occurs with shift” s
13. if s < n – m
14. then ts + 1  (d(ts – T [s + 1] h) + T[s + m +1]) mod q
Example of Rabin-Karp method : Working modulo q = 11, how many
spurious hits does the Rabin-Karp matcher encounter in the text
T = 3141592653589793 when looking for the pattern p = 26
Given, p = 26 and q = 11
Now we divide 26 by 11 i.e.,
Remainder is 4 and m = 2.
We know m denotes the length of p.
T 3 1 4 1 5 9 2 6 5 3 5 8 9 7 9 3

Now we divide 31 by 11, and get remainder is 9.

Similarly, 14 by 11 and get remainder is 3.
So, continue this step till last i.e., 93 is divided by 11 and get remainder is 5.
After that we will store all remainder in a table.
9 3 8 4 4 4 4 10 9 2 3 1 9 2 5
Now we find valid matching.
3 1 4 1 5 9 2 6 5 3 5 8 9 7 9 3

Valid matching
9 3 8 4 4 4 4 10 9 2 3 1 9 2 5

Spurious
hit
The number of spurious hits is 3.
 More pdf : www.motivationbank.in
5–13 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Que 5.10. Give a linear time algorithm to determine if a text T is a

cycle rotation of another string T. For example : RAJA and JARA are
cyclic rotations of each other. AKTU 2018-19, Marks 10

Answer
Knuth-Morris-Pratt algorithm is used to determine if a text T is a cycle
rotation of another string T.
Knuth-Morris-Pratt algorithm : Refer Q. 5.6, Page 5–7B, Unit-5.

PART-3
Theory of NP – Completeness.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 5.11. Discuss the problem classes P, NP and NP-complete.

Answer
P : Class P are the problems which can be solved in polynomial time, which
take time like O(n), O(n2), O(n3).
Example : Finding maximum element in an array or to check whether a
string is palindrome or not. So, there are many problems which can be
solved in polynomial time.
NP : Class NP are the problems which cannot be solved in polynomial time
like TSP (travelling salesman problem).
Example : Subset sum problem is best example of NP in which given a set
of numbers, does there exist a subset whose sum is zero, but NP problems
are checkable in polynomial time means that given a solution of a problem,
we can check that whether the solution is correct or not in polynomial time.
NP-complete : The group of problems which are both in NP and NP-hard
are known as NP-complete problem.
Now suppose we have a NP-complete problem R and it is reducible to Q
then Q is at least as hard as R and since R is an NP-hard problem, therefore
Q will also be at least NP-hard, it may be NP-complete also.

Que 5.12. Discuss the problem classes P, NP and NP-complete

with class relationship. AKTU 2017-18, Marks 10

Selected Topics 5–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Answer
1. The notion of NP-hardness plays an important role in the relationship
between the complexity classes P and NP.
NP
NP-complete

NP-hard
P
Fig. 5.12.1. Relationship among P, NP, NP-complete
and NP-hard problems.
2. It is also often used to define the complexity class NP-complete which is
the intersection of NP and NP-hard.
3. Consequently class NP-hard can be understood as the class of problems
that are NP-complete or harder.
4. There are no polynomial time algorithms for NP-hard problems.
5. A problem being in NP means that the problem is “easy” (in a certain
specific sense), whereas a problem being NP-hard means that the
problem is “difficult” (in another specific sense).
6. A problem being in NP and a problem being NP-hard are not mutually
exclusive. When a problem is both in NP and NP-hard, we say that the
problem is NP-complete.
7. All problems in NP can be solved deterministically in time O(2n).
8. An example of an NP-hard problem is the decision problem subset-sum.
Given a set of integers, does any non-empty subset of them add up to
zero ? i.e., a yes or no question, and happens to be NP-complete.
9. There are, also decision problems that are NP-hard but not NP-complete.
10. For example, in the halting problem “given a program and its input, will
it run forever” i.e., yes or no question, so this is a decision problem. It is
case to prove that the halting problem is NP-hard but not NP-complete.

Que 5.13. What is NP-completeness ?

Answer
A language L  {0,1}* is NP-complete if it satisfies the following two properties :
i. L ∈ NP ; and
ii. For every Lp L
NP-hard : If a language L satisfies property (ii), but not necessarily property
(i), we say that L is NP-hard.
NP-complete : We use the notation LNPC to denote that L is NP-complete.
Theorem : If any NP-complete problem is polynomial time solvable, then
P = NP. If any problem in NP is not polynomial time solvable, then all NP
complete problems are not polynomial time solvable.
 More pdf : www.motivationbank.in
5–15 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Proof : Suppose that L  P and also that L  NPC. For any L NP, we have
Lp L by property (ii) of the definition of NP-completeness. We know if
Lp L then L  P implies L P, which proves the first statement.
To prove the second statement, suppose that there exists and L  NP such
that L   P. Let L NPC be any NP-complete language, and for the purpose
of contradiction, assume that L P. But then we have L p L and thus L  P.

Que 5.14. Explain NP-hard and NP-complete problems and also

define the polynomial time problems and write a procedure to
solve NP-problems.
OR
Write short note on NP-hard and NP-complete problems.
OR
Define NP-hard and NP-complete problems. What are the steps
involved in proving a problem NP-complete ? Specify the problems
already proved to be NP-complete. AKTU 2019-20, Marks 07

Answer
NP-hard problem :
1. We say that a decision problem Pi is NP-hard if every problem in NP is
polynomial time reducible to Pi.
2. In symbols,
Poly
Pi is NP-hard if, for every Pj  NP, Pj   Pi .
3. This does not require Pi to be in NP.
4. Highly informally, it means that Pi is ‘as hard as’ all the problem in NP.
5. If Pi can be solved in polynomial time, then all problems in NP.
6. Existence of a polynomial time algorithm for an NP-hard problem implies
the existence of polynomial solution for every problem in NP.
NP-complete problem :
1. There are many problems for which no polynomial time algorithms is
known.
2. Some of these problems are travelling salesman problem, optimal graph
colouring, the Knapsack problem, Hamiltonian cycles, integer
programming, finding the longest simple path in a graph, and satisfying
a Boolean formula.
3. These problems belongs to an interesting class of problems called the
‘‘NP-complete’’ problems, whose status is unknown.
4. The NP-complete problems are traceable i.e., require a super polynomial
time.
Polynomial time problem :
An algorithm is said to be solvable in polynomial time if the number of steps
required to complete the algorithm for a given input is O(nk) for some non-
negative integer k where n is the complexity of input.
Selected Topics 5–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Polynomial time verifiable algorithm : A polynomial time algorithm A
is said to be polynomial time verifiable if it has following properties :
1. The input to A consists of an instance I of X (X is a decision problem) and
a string S such that the length of S is bounded by some polynomial in the
size of I.
2. The output of A is either yes or no.
3. If I is a negative instance of X, then the output of A is “no” regardless of
the value of S.
4. If I is a positive instance of X, then there is at least one choice of S for
which A output “yes”.
Procedure to solve NP-problems :
1. The class NP is the set of all decision problems that have instances that
are solvable in polynomial time using a non-deterministic turing machine.
2. In a non-deterministic turing machine, in contrast to a deterministic
turing machine, for each state, several rules with different actions can
be applied.
3. Non-deterministic turing machine branches into many copies that are
represented by a computational tree in which there are different
computational paths.
4. The class NP corresponds to a non-deterministic turing machine that
guesses the computational path that represents the solution.
5. By doing so, it guesses the instances of the decision problem.
6. In the second step, a deterministic turing machine verifies whether the
guessed instance leads to a “yes” answer.
7. It is easy to verify whether a solution is valid or not. This statement does
not mean that finding a solution is easy.

Que 5.15. Differentiate NP-complete with NP-hard.

AKTU 2016-17, Marks 10

Answer

S. No. NP-complete NP-hard

1. An NP-complete problems NP-hard problems is one to which
is one to which every other an N P-co mple te pro blem is
polynomial-time non- Turing-reducible.
deterministic algorithm can
be reduced in polynomial
time.
2. NP-complete problems do NP-hard problems correspond to
no t co rrespo nds to an an NP-complete problem.
NP-hard problem.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 5–17 B (CS/IT-Sem-5)

3. NP-complete problems are NP-hard problems need not to be

exclusively decision decision problem.
problem.
4. N P-co mple te pro blems NP-hard problems do not have to
have to be in NP-hard and be in NP.
also in NP.
5. For example : 3-SAT For example : Halting problem is
vertex cover problem is NP-hard.
NP-complete.

Que 5.16. Discuss NP-complete problem and also explain minimum

vertex cover problem in context to NP-completeness.

Answer
NP-complete problem : Refer Q. 5.14, Page 5–15B, Unit-5.
Minimum vertex cover problem :
1. A vertex cover of a graph is a set of vertices such that each edge of the
graph is incident to at least one vertex of the set.
2. The problem of finding a minimum vertex cover is a classical optimization
problem in computer science and is a typical example of an NP-hard
optimization problem that has an approximation algorithm.
3. Its decision version, the vertex cover problem, was one of Karp's 21 NP-
complete problems and is therefore a classical NP-complete problem in
computational complexity theory.
4. Furthermore, the vertex cover problem is fixed-parameter tractable
and a central problem in parameterized complexity theory.
5. The minimum vertex cover problem can be formulated as a half-integral
linear program whose dual linear program is the maximum matching
problem.
6. Formally, a vertex cover V of undirected graph G = (V, E) is a subset of
V such that uv  V  v  V i.e., it is a set of vertices V where every edge
has at least one endpoint in the vertex cover V. Such a set is said to
cover the edges of G.
Example : Fig. 5.16.1 shows examples of vertex covers, with some
vertex cover V marked in dark.

Fig. 5.16.1.
7. A minimum vertex cover is a vertex cover of smallest possible size.
8. The vertex cover number  is the size of a minimum cover, i.e.,  = |V|.
The Fig. 5.16.2 shows examples of minimum vertex covers in the previous
graphs.
Selected Topics 5–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in

Fig. 5.16.2.

Que 5.17. Discuss different types of NP-complete problem.

Answer
Types of NP-complete problems :
NP-complete problems

Hamiltonian cycle problem Clique problem

Fig. 5.17.1.
Hamiltonian cycle problem :
1. A Hamiltonian cycle of an undirected graph G = (V, E) is a simple cycle
that contains each vertex in V.
2. A graph that contains a Hamiltonian cycle is said to be Hamiltonian,
otherwise it is said to be non-Hamiltonian.
The clique problem :
1. A clique in an undirected graph G = (V, E) is a subset V  V of vertices,
each pair of which is connected by an edge in E.
2. The size of a clique is the number of vertices it contains.
CLIQUE = {<G, K> : G is a graph with a clique of size K }
3. The clique problem is the optimization problem of finding a clique of
maximum size in a graph.
4. As a decision problem, we ask simply whether a clique of a given size k
exists in the graph.

Que 5.18. Show that Hamiltonian circuit is NP-complete.

Answer
Theorem : Hamiltonian circuit (HC) is NP-complete.
Proof :
1. Let us define a non-deterministic algorithm A that takes, as input, a
graph G encoded as an adjacency list in binary notation, with the vertices
numbered 1 to N.
2. We define A to first iteratively call the choose method to determine a
sequence S of N + 1 numbers from 1 to N.
3. Then, we have A to check that each number from 1 to N appears exactly
once in S (for example, by sorting S), except for the first and last numbers
in S, which should be the same.
 More pdf : www.motivationbank.in
5–19 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

4. Then, we verify that a sequence S defines a cycle of vertices and edges

in G.
5. A binary encoding of the sequence S is clearly of size at most n, where n
is the size of the input. Moreover, both of the checks made on the
sequence S can be done in polynomial time in n.
6. Observe that if there is a cycle in G that visits each vertex of G exactly
once; returning to its starting vertex, then there is a sequence S for
which A will output “yes.”
7. Likewise, if A outputs “yes,” then it has found a cycle in G that visits
each vertex of G exactly once, returning to its starting point. Hence,
Hamiltonian circuit is NP-complete.

Que 5.19. Show that CLIQUE problem is NP-complete.

Answer
Problem : The CLIQUE problem is defined as {< G, k >|, G is a graph with
a k-clique}. Show that CLIQUE is NP-complete.
Proof :
1. First, to show that CLIQUE is in NP. Given an instance of < G, k > and
a k-clique we can easily verify in O(n2) time that we do, in fact, have a
k-clique.
2. Now, we want to show that 3-SAT is CLIQUE. Let F be a boolean
formula in CNF.
3. For each literal in F we will make a vertex in the graph i.e.,
(x1 + x2 + x3) ( x1  x2  x3 ) has 6 vertices.
Let k be the number of clauses in F.
4. We will connect each vertex to all of the other vertices that are logically
compatible except for the ones that are in the same clause.
5. Now, if we have a satisfiable assignment we will have a k-clique because
the satisfying vertices will all be connected to one another.
6. Thus, we can use CLIQUE to solve 3-SAT so CLIQUE is NP-complete.

Que 5.20. Define different complexity classes in detail with

suitable example. Show that TSP problem is NP-complete.

Answer
Different complexity classes :
There are some complexity classes involving randomized algorithms :
1. Randomized polynomial time (RP) : The class RP consists of all
languages L that have a randomized algorithm A running in worst case
polynomial time such that for any input x in *
1
x  L  P[A(x) accepts] 
2
x  L  P[A(x) accepts] = 0
Selected Topics 5–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Independent repetitions of the algorithms can be used to reduce the
probability of error to exponentially small.
2. Zero-error probabilistic polynomial time (ZPP) : The class ZPP is
the class of languages which have Las Vegas algorithms running in
expected polynomial time.
ZPP = RP  co-RP
where a language L is in co-X where X is a complexity class if and only
if its complement * – L is in X.
3. Probabilistic polynomial time (PP) : The class PP consists of all
languages L that have a randomized algorithm A running in worst case
polynomial time such that for any input x in *.
1
x  L  P[A(x) accepts] 
2
1
x  L  P[A(x) accepts] <
2
To reduce the error probability, we cannot repeat the algorithm several
times on the same input and produce the output which occurs in the
majority of those trials.
4. Bounded-error probabilistic polynomial time (BPP) : The class
BPP consists of all languages that have a randomized algorithm A
running in worst case polynomial time such that for any input x in *.
3
x  L  P[A(x) accepts] 
4
1
x  L  P[A(x) accepts] 
4
For this class of algorithms, the error probability can be reduced to 1/
2n with only a polynomial number of iterations.
For a given a graph G = (V, E) and a number k, does there exist a tour
C on G such that the sum of the edge weights for edges in C is less than
or equal to k.
Proof :
Part 1 : TSP is in NP.
Proof :
1. Let a hint S be a sequence of vertices V = v1,..., vn.
2. We then check two things :
a. First we check that every edge traversed by adjacent vertices is an
edge in G, such that the sum of these edge weights is less than or
equal to k.
b. Secondly we check that every vertex in G is in V, which assures
that every node has been traversed.
3. We accept S if and only if S satisfies these two questions, otherwise
reject.
4. Both of these checks are clearly polynomial, thus our algorithm forms a
verifier with hint S, and TSP is consequently in NP.
 More pdf : www.motivationbank.in
5–21 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Part 2 : TSP is NP-Hard.

Proof :
1. To show that TSP is NP-Hard, we must show that every problem y in NP
reduces to TSP in polynomial time.
2. To do this, consider the decision version of Hamiltonian Cycle (HC).
3. Take G = (V, E), set all edge weights equal to 1, and let k = |V|= n, that
is, k equals the number of nodes in G.
4. Any edge not originally in G then receives a weight of 2 (traditionally
TSP is on a complete graph, so we need to add in these extra edges).
5. Then pass this modified graph into TSP, asking if there exists a tour on
G with cost at most k. If the answer to TSP is YES, then HC is YES.
Likewise if TSP is NO, then HC is NO.
First direction : HC has a YES answer => TSP has a YES answer.
Proof :
1. If HC has a YES answer, then there exists a simple cycle C that visits
every node exactly once, thus C has n edges.
2. Since every edge has weight 1 in the corresponding TSP instance for the
edges that are in the HC graph, there is a Tour of weight n. Since k = n,
and given that there is a tour of weight n, it follows that TSP has a YES
answer.
Second direction : HC has a NO answer => TSP has a NO answer.
Proof :
1. If HC has a NO answer, then there does not exist a simple cycle C in G
that visits every vertex exactly once. Now suppose TSP has a YES
answer.
2. Then there is a tour that visits every vertex once with weight at most k.
3. Since the tour requires every node be traversed, there are n edges, and
since k = n, every edge traversed must have weight 1, implying that
these edges are in the HC graph. Then take this tour and traverse the
same edges in the HC instance. This forms a Hamiltonian Cycle, a
contradiction.
This concludes Part 2. Since we have shown that TSP is both in NP and
NP-Hard, we have that TSP is NP-Complete.
Que 5.21. Prove that three colouring problem is NP-complete.

AKTU 2016-17, Marks 10

Answer
1. To show the problem is in NP, let us take a graph G(V, E) and a colouring
c, and checks in O(n2 ) time whether c is a proper colouring by checking
if the end points of every edge e  E have different colours.
2. To show that 3-COLOURING is NP-hard, we give a polytime reduction
from 3-SAT to 3-COLOURING.
Selected Topics 5–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
3. That is, given an instance  of 3-SAT, we will construct an instance of
3-COLOURING (i.e., a graph G(V, E)) where G is 3-colourable iff  is
satisfiable.
4. Let  be a 3-SAT instance and C1, C2, ..., Cm be the clauses of  defined
over the variables {x1, x2, ..., xn}.
5. The graph G(V, E) that we will construct needs to capture two things :
a. Somehow establish the truth assignment for x1, x2, ..., xn via the
colours of the vertices of G; and
b. Somehow capture the satisfiability of every clause Ci in .
6. To achieve these two goals, we will first create a triangle in G with three
vertices {T, F, B} where T stands for True, F for False and B for Base.
7. Consider {T, F, B} as the set of colours that we will use to colour (label)
the vertices of G.
8. Since this triangle is part of G, we need 3 colours to colour G.
9. Now we add two vertices vi, vi for every literal xi and create a triangle B,

vi , vi for every (vi , vi ) pair, as shown in Fig. 5.21.1.

T F

vn
B
v1
vn
v1
v2
v2

Fig. 5.21.1.
10. This construction captures the truth assignment of the literals.
11. Since if G is 3-colourable, then either vi or vi gets the colour T, and we
interpret this as the truth assignment to vi.
12. Now we need to add constraints to G to capture the satisfiability of the
clauses of .
13. To do so, we introduce the Clause Satisfiability Gadget, (the OR-gadget).
For a clause Ci = (a  b  c), we need to express the OR of its literals using
our colours {T, F, B}.
14. We achieve this by creating a small gadget graph that we connect to the
literals of the clause. The OR-gadget is constructed as follows :
a ab
abc
b

The output node

c

Fig. 5.21.2.
 More pdf : www.motivationbank.in
5–23 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

15. Consider this gadget graph as a circuit whose output is the node labeled
a  b  c. We basically want this node to be coloured T if Ci is satisfied and
F otherwise.
16. This is a two step construction : The node labelled a  b captures the
output of (a  b) and we repeat the same operation for ((a  b)  c). If we
play around with some assignments to a, b, c, we will notice that the
gadget satisfies the following properties :
a. If a, b, c are all coloured F in a 3-colouring, then the output node of
the OR-gadget has to be coloured F. Thus capturing the
unsatisfiability of the clause Ci = (a  b  c).
b. If one of a, b, c is coloured T, then there exists a valid
3-colouring of the OR-gadget where the output node is coloured T.
Thus again capturing the satisfiability of the clause.
17. Once we add the OR-gadget of every Ci in , we connect the output node
of every gadget to the Base vertex and to the False vertex of the initial
triangle, as follows :
a ab
abc T
b
B

c F
Fig. 5.21.3.
18. Now we prove that our initial 3-SAT instance  is satisfiable if and only
the graph G as constructed above is 3-colourable. Suppose  is satisfiable
and let (x1*, x2*, ..., xn* ) be the satisfying assignment.
19. If xi* is assigned True, we colour vi with T and vi with F (recall they are
connected to the Base vertex, coloured B, so this is a valid colouring).
20. Since  is satisfiable, every clause Ci = (a b c) must be satisfiable, i.e.,
at least of a, b, c is set to True. By the property of the OR-gadget, we
know that the gadget corresponding to Ci can be 3-coloured so that the
output node is coloured T.
21. And because the output node is adjacent to the False and Base vertices
of the initial triangle only, this is a proper 3-colouring.
22. Conversely, suppose G is 3-colourable. We construct an assignment of
the literals of  by setting xi to True if vi is coloured T and vice versa.
23. Now consider this assignment is not a satisfying assignment to , then
this means there exists at least one clause Ci = (a  b  c) that was not
satisfiable.
24. That is, all of a, b, c were set to False. But if this is the case, then the
output node of corresponding OR-gadget of Ci must be coloured F.
25. But this output node is adjacent to the False vertex coloured F; thus
contradicting the 3-colourability of G.
26. To conclude, we have shown that 3-COLOURING is in NP and that it is
NP-hard by giving a reduction from 3-SAT.
Selected Topics 5–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
27. Therefore 3-COLOURING is NP-complete.
Que 5.22. Prove that P is the subset of NP.

Answer
To prove : P is the subset of NP.
Proof :
1. If L P, then LNP, as there is a polynomial time algorithm to decide L,
this algorithm can easily be converted into a row argument verification
algorithm that simply ignores any exception and accepts exactly those
input strings it determines to be in L.
2. Thus, P  NP.
NP
P P = NP P N

(a) (b) (c)

Fig. 5.22.1. P Vs NP.

PART-4
Approximation Algorithm.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 5.23. Describe approximation algorithm in detail. How it

differ with deterministic algorithm. Show that TSP is 2 approximate.
OR
Explain approximation algorithms with suitable examples.
AKTU 2015-16, 2017-18; Marks 10

Answer
Approximation algorithm :
1. An approximation algorithm is a way of dealing with NP-completeness
for optimization problem. This technique does not guarantee the best
solution.
2. The best of an approximation algorithm is to come as close as possible to
the optimum value in a reasonable amount of time which is at most
polynomial time.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms 5–25 B (CS/IT-Sem-5)

3. Let c(i) be the cost of solution produced by approximate algorithm and

c*(i) be the cost of optimal solution for some optimization problem
instance i.
4. For minimization and maximization problem, we are interested in finding
a solution of a given instance i in the set of feasible solutions, such that
c(i) / c*(i) and c*(i) / c(i) be as small as possible respectively.
5. We say that an approximation algorithm for the given problem instance
i, has a ratio bound of p(n) if for any input of size n, the cost c of the
solution produced by the approximation algorithm is within a factor of
p(n) of the cost c* of an optimal solution. That is
max(c(i) / c*(i), c*(i) / c(i))  p(n)
The definition applies for both minimization and maximization problems.
6. p(n) is always greater than or equal to 1. If solution produced by
approximation algorithm is true optimal solution then clearly we have
p(n) = 1.
7. For a minimization problem, 0 < c*(i) < c(i), and the ratio c(i) / c*(i) gives
the factor by which the cost of the approximate solution is larger than
the cost of an optimal solution.
8. Similarly, for a maximization problem, 0 < c(i)  c*(i), and the ratio
c*(i) / c(i) gives the factor by which the cost of an optimal solution is
larger than the cost of the approximate solution.
Difference between deterministic algorithm and approximation
algorithm :
S. No. Deterministic algorithm Approximation algorithm
1. It does not deal with optimization It deals with optimization
problem. problem.
2. It has initial and final step. It does not have initial or final
state.
3. It require finite state machine. It does not require finite state
machine.
4. It fails to deliver a result. It gives an optimal result.
5. It does not apply to maximization It applies to maximization and
or minimization problem. minimization problem.

Proof :
TSP is 2-approximate :
Let H* denote the optimal tour. Observe that a TSP with one edge removed
is a spanning tree (not necessarily MST).
It implies that the weight of the MST ‘T ’ is in lower bound on the cost of an
optimal tour.
c(T)  c(H*)
A “Full” walk, W, traverse every edge of MST, T, exactly twice. That is,
c(W) = 2c(T)
which means
c(W)  2c(H*)
Selected Topics 5–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
and we have
c(W) / c(H*)  p(n) = 2
That is, the cost of walk, c(W), of the solution produced by the algorithm is
within a factor of p(n) = 2 of the cost c(H*) of an optimal solution.

Que 5.24. Explain vertex cover problem with algorithm and

analysis.

Answer
A vertex cover of an undirected graph G = (V, E) is a subset of V V such
that if edge (u, v)  G then u  V or v V (or both).
Problem : Find a vertex cover of maximum size in a given undirected graph.
This optimal vertex cover is the optimization version of an NP-Complete
problem but it is not too hard to find a vertex cover that is near optimal.
Approx-vertex-cover (G : Graph)
1. c 
2. E  E[G]
3. while E is not empty
4. do Let (u, v) be an arbitrary edge of E
5. c  c {u, v}
6. Remove from E every edge incident on either u or v
7. return c
Analysis : It is easy to see that the running time of this algorithm is
O(V + E), using adjacency list to represent E.
Que 5.25. Describe approximation algorithm in detail. What is
the approximation ratio ? Show that vertex cover problem is
2-approximate.

Answer
Approximation algorithm : Refer Q. 5.23, Page 5–24B, Unit-5.
Proof :
Vertex cover problem is 2-approximate :
Goal : Since this is a minimization problem, we are interested in smallest
possible c / c*. Specifically we want to show c / c* = 2 = p(n).
In other words, we want to show that Approx-Vertex-Cover algorithm returns
a vertex-cover that is almost twice the size of an optimal cover.
Proof : Let the set c and c* be the sets output by Approx-Vertex-Cover and
Optimal-Vertex-Cover respectively. Also, let A be the set of edges.
Because, we have added both vertices, we get c = 2|A| but Optimal-Vertex-
Cover would have added one of two.
c / c*  p(n) = 2.
Formally, since no two edge in A are covered by the same vertex from c* and
the lower bound : |c*| A ...(5.25.1)
on the size of an Optimal-Vertex-Cover.
Now, we pick both end points yielding an upper bound on the size of Vertex-
Cover :
 More pdf : www.motivationbank.in
5–27 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

|c| 2|A|
Since, upper bound is an exact in this case, we have
|c|= 2|A| ...(5.25.2)
Take |c|/ 2 = |A| and put it in equation (5.25.1)
|c*| |c|/ 2
|c*|/|c| 1/2
|c*|/|c| 2 = p(n) Hence the theorem proved.

Que 5.26. Explain Travelling Salesman Problem (TSP) with the

triangle inequality.

Answer
Problem : Given a complete graph with weights on the edges, find a cycle of
least total weight that visits each vertex exactly once. When the cost function
satisfies the triangle inequality, we can design an approximate algorithm for
TSP that returns a tour whose cost is not more than twice the cost of an
optimal tour.
APPROX-TSP-TOUR (G, c) :
1. Select a vertex r  V[G] to be a “root” vertex.
2. Compute a minimum spanning tree T for G from root r using MST-
PRIM (G, c, r).
3. Let L be the list of vertices visited in a pre-order tree walk of T.
4. Return the Hamiltonian cycle H that visits the vertices in the order L.
Outline of an approx-TSP tour : First, compute a MST (minimum spanning
tree) whose weight is a lower bound on the length of an optimal TSP tour.
Then, use MST to build a tour whose cost is no more than twice that of MST’s
weight as long as the cost function satisfies triangle inequality.

Que 5.27. Write short notes on the following using approximation

algorithm with example.
i. Nearest neighbour
ii. Multifragment heuristic

Answer
i. Nearest neighbour :
The following well-known greedy algorithm is based on the nearest-
neighbour heuristic i.e., always go next to the nearest unvisited city.
Step 1 : Choose an arbitrary city as the start.
Step 2 : Repeat the following operation until all the cities have been
visited : go to the unvisited city nearest the one visited last (ties can be
broken arbitrarily).
Step 3 : Return to the starting city.
Example :
1. For the instance represented by the graph in Fig. 5.27.1, with a as the
starting vertex, the nearest-neighbour algorithm yields the tour
(Hamiltonian circuit) sa: a – b – c – d – a of length 10.
Selected Topics 5–28 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. The optimal solution, as can be easily checked by exhaustive search, is
the tour s*: a – b – d – c – a of length 8. Thus, the accuracy ratio of this
approximation is
f (sa ) 10
r(sa) = = =1.25
f (s*) 8
i.e., tour sa is 25 % longer than optimal tour s*.
1
a b

6 2
3 3

c d
1
Fig. 5.27.1. Instance of the traveling salesman problem.
3. Unfortunately, except for its simplicity, not many good things can be
said about the nearest-neighbour algorithm.
4. In particular, nothing can be said in general about the accuracy of
solutions obtained by this algorithm because it can force us to traverse
a very long edge on the last leg of the tour.
5. Indeed, if we change the weight of edge (a, d) from 6 to an arbitrary
large number w  6 in given example, the algorithm will still yield the
tour a – b – c – d – a of length 4 + w, and the optimal solution will still be
a – b – d – c – a of length 8. Hence,
f (s a ) 4w
r(sa) = =
f (s*) 8
which can be made as large as we wish by choosing an appropriately
large value of w. Hence, RA =  for this algorithm.
ii. Multifragment heuristic :
Another natural greedy algorithm for the traveling salesman problem
considers it as the problem of finding a minimum-weight collection of
edges in a given complete weighted graph so that all the vertices have
degree 2.
Step 1 : Sort the edges in increasing order of their weights. (Ties can be
broken arbitrarily.) Initialize the set of tour edges to be constructed to
the empty set.
Step 2 : Repeat this step n times, where n is the number of cities in the
instance being solved : add the next edge on the sorted edge list to the
set of tour edges, provided this addition does not create a vertex of
degree 3 or a cycle of length less than n; otherwise, skip the edge.
Step 3 : Return the set of tour edges.
Example :
1. Applying the algorithm to the graph in Fig. 5.27.1 yie lds
{(a, b), (c, d), (b, c), (a, d)}.
 More pdf : www.motivationbank.in
5–29 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

2. There is, however, a very important subset of instances, called Euclidean,

for which we can make a non-trivial assertion about the accuracy of
both the nearest-neighbour and multifragment-heuristic algorithms.
3. These are the instances in which intercity distances satisfy the following
natural conditions :
a. Triangle inequality : d[i, j ]  d[i, k] + d[k, j ] for any triple of
cities i, j, and k (the distance between cities i and j cannot exceed
the length of a two-leg path from i to some intermediate city k to j).
b. Symmetry : d[i, j ] = d[j, i] for any pair of cities i and j (the distance
from i to j is the same as the distance from j to i).
4. A substantial majority of practical applications of the traveling salesman
problem are its Euclidean instances.
5. They include, in particular, geometric ones, where cities correspond to
points in the plane and distances are computed by the standard Euclidean
formula.
6. Although the performance ratios of the nearest-neighbour and
multifragment-heuristic algorithms remain unbounded for Euclidean
instances, their accuracy ratios satisfy the following inequality for any
such instance with n  2 cities :
f (s a ) 4w
r(sa) = =
f (s*) 8
where f (sa) and f (s*) are the lengths of the heuristic tour and shortest
tour.

Que 5.28. What is an approximation algorithm ? What is meant

by p(n) approximation algorithms ? Discuss approximation
algorithm for Travelling Salesman Problem.
AKTU 2019-20, Marks 07

Answer
Approximation algorithm : Refer Q. 5.23, Page 5–24B, Unit-5.
p(n) approximation algorithm : A is a p(n) approximate algorithm if and
only if for every instance of size n, the algorithm achieves an approximation
ratio of p(n). It is applied to both maximization (0 < C(i)  C*(i)) and
minimization (0 < C* (i)  C(i)) problem because of the maximization factor
and costs are positive. p(n) is always greater than 1.
Approximation algorithm for Travelling Salesman Problem (TSP) :
1. The key to designing approximation algorithm is to obtain a bound on
the optimal value (OPT).
2. In the case of TSP, the minimum spanning tree gives a lower bound on
OPT.
3. The cost of a minimum spanning tree is not greater than the cost of an
optimal tour.
The algorithm is as follows :
1. Find a minimum spanning tree of G.
Selected Topics 5–30 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. Duplicate each edge in the minimum spanning tree to obtain a Eulerian
graph.
3. Find a Eulerian tour (J) of the Eulerian graph.
4. Convert J to a tour T by going through the vertices in the same order of
T, skipping vertices that were already visited.

PART-5
Randomized Algorithm.

Questions-Answers

Long Answer Type and Medium Answer Type Questions

Que 5.29. Write short notes on randomized algorithms.

OR
Explain approximation and randomized algorithms.
AKTU 2017-18, Marks 10

Answer
Approximation algorithm : Refer Q. 5.23, Page 5–24B, Unit-5.
Randomized algorithm :
1. A randomized algorithm is defined as an algorithm that is allowed to
access a source of independent, unbiased bits and it is then allowed to
use these random bits to influence its computation.
2. An algorithm is randomized if its output is determined by the input as
well as the values produced by a random number generator.
3. A randomized algorithm makes use of a randomizer such as a random
number generator.
4. The execution time of a randomized algorithm could also vary from
run to run for the same input.
5. The algorithm typically uses the random bits as an auxiliary input to
guide its behaviour in the hope of achieving good performance in the
“average case”.
6. Randomized algorithms are particularly useful when it faces a malicious
attacker who deliberately tries to feed a bad input to the algorithm.
Randomized algorithm are categorized into two classes :
i. Las Vegas algorithm : This algorithm always produces the same
output for the same input. The execution time of Las Vegas algorithm
depends on the output of the randomizer.
ii. Monte Carlo algorithm :
a. In this algorithm output might differ from run to run for the same
input.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms5–31 B (CS/IT-Sem-5)

b. Consider any problem for which there are only two possible
answers, say yes and no.
c. If a Monte Carlo algorithm is used to solve such a problem then
the algorithm might give incorrect answers depending on the
output of the randomizer.
d. Then the requirement is that the probability of an incorrect answer
from a Monte Carlos algorithm be low.
Que 5.30. Write a short note on randomized algorithm. Write its
merits, and applications.

Answer
Randomized algorithm : Refer Q. 5.29, Page 5–30B, Unit-5.
Merits :
1. Simple.
2. High efficiency.
3. Better complexity bounds.
4. Random selection of good and bad choices.
5. Cost efficient.
Applications :
1. Randomized quick sort algorithm
2. Randomized minimum-cut algorithm
3. Randomized algorithm for N-Queens problem
4. Randomized algorithm for majority element
Que 5.31. Write the EUCLID’S GCD algorithm. Compute gcd (99,
78) with EXTENDED-EUCLID.

Answer
Euclid’s GCD algorithm :
The inputs a and b are arbitrary non-negative integers.
EUCLID (a, b) {
if (b = = 0)
then return a;
else return EUCLID (b, a mod b); }
EXTEUCLID (a, b) {
//returns a triple (d, x, y) such that d = gcd (a, b)
//d = = (a × x + b × y)
if (b = = 0) return (a, 1, 0);
(d1, x1, y1) = EXTEUCLID (b, a % b);
d = d1;
x = y 1;
y = x1 – (a div b) × y1; //div = integer division
return (d, x, y);
}
Selected Topics 5–32 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Numerical :
Let a = 99 and b = 78
a b [a/b] d x y
99 78 1 3 – 11 14
78 21 3 3 3 – 11
21 15 1 3 –2 3
15 6 2 3 1 –2
6 3 2 3 0 1
3 0 – 3 1 0
i. In the 5th receive (a = 6, b = 3), values from the 6th call (b = 0) has
d1 = 3, x1 = 1 and y1 = 0. Still within the 5th call we calculate that
d = d1 = 3, x = y1 = 0 and y = x1 – (a div b) × y1 = 1 – 2 × 0 = 1.
ii. In the 4th receive (a = 15, b = 6), the values d1 = 3, x1 = 0 and y1 = 1 from
the 5th call, then compute x = y1 = 1 and
y = x1 – (a div b) × y1 = 0 – 2 × 1 = – 2.
iii. In the 3rd receive (a = 21, b = 15), x = – 2 and
y = x1 – (a div b) × y1 = 1 – (1 × – 2) = 3.
iv. In the 2nd receive (a = 78, b = 21), x = 3 and
y = x1 – (a div b) × y1 = (– 2) – 3 × 3 = – 11.
v. In the 1st receive (a = 99, b = 78), x = – 11 and
y = x1 – (a div b) × y1 = 3 – 1 × (– 11) = 14.
vi. The call EXTEUCLID (99, 78) return (3, – 11, 14), so gcd (99, 78) = 3 and
gcd (99, 78) = 3 = 99 × (–11) + 78 × 14.

VERY IMPORTANT QUESTIONS

Following questions are very important. These questions
may be asked in your SESSIONALS as well as
UNIVERSITY EXAMINATION.

Q. 1. What is Fast Fourier Transformation and how it works ?

Ans. Refer Q. 5.1.
Q. 2. Explain the following string matching algorithms :
a. Naive string matching
b. Rabin-Karp algorithm
c. Knuth-Morris-Pratt algorithm
Ans.
a. Refer Q. 5.4.
b. Refer Q. 5.8.
c. Refer Q. 5.6.
 More pdf : www.motivationbank.in
Design and Analysis of Algorithms5–33 B (CS/IT-Sem-5)

Q. 3. Discuss the problem classes P, NP and NP-complete.

Ans. Refer Q. 5.11.

Q. 4. Differentiate NP-complete with NP-hard.

Ans. Refer Q. 5.15.

Q. 5. Show that CUQUE NP-complete.

Ans. Refer Q. 5.19.

Q. 6. Explain the following :

a. Approximation algorithm
b. Randomized algorithm
Ans.
a. Refer Q. 5.23.
b. Refer Q. 5.29.

Q. 7. Describe in detail Knuth-Morris-Pratt string matching

algorithm. Compute the prefix function  for the pattern
ababbabbabbababbabb when the alphabet is  = {a, b}.
Ans. Refer Q. 5.8.


 More pdf : www.motivationbank.in
Design and Analysis of AlgorithmsSQ–1 B (CS/IT-Sem-5)

1 Introduction
(2 Marks Questions)

1.1. Define the term algorithm.

Ans. Algorithm is a set of rules for carrying out calculation either by
hand or on a machine. It is a finite step-by-step procedure to achieve
a required result.

1.2. What are the steps needed to be followed while designing

an algorithm ?
Ans. Steps of designing an algorithm :
1. Understanding the problem.
2. Decision making on capabilities of computational devices.
3. Specification of algorithm.
4. Algorithm verification.
5. Analysis of algorithm.
6. Implementation or coding of algorithm.

1.3. Define the following terms :

i. Time efficiency
ii. Space efficiency
Ans.
i. Time efficiency : Time efficiency of an algorithm or a program is
the amount of time it needs to run to completion.
ii. Space efficiency : Space efficiency of an algorithm or program is
the amount of memory it needs to run to completion.

1.4. Give an example of worst case time complexity.

Ans. While searching a particular element using linear searching method,
if desired element is placed at the end of the list then we get worst
time complexity.

1.5. Compare time complexity with space complexity.

AKTU 2017-18, Marks 02
 More pdf : www.motivationbank.in
2 Marks Questions SQ–2 B (CS/IT-Sem-5)

Ans.
S. No. Time complexity Space complexity
1. Time complexity is the Space complexity is the amount
amount of time required for of memory ne eded for an
an algorithm to complete its algorithm to solve the problem.
process.
2. It is expressed using Big It is expressed only using Big
Oh(O), theta () and Oh(O) notation.
omega () notation.

1.6. What are the characteristics of the algorithm ?

AKTU 2017-18, Marks 02
Ans. Characteristics of algorithm :
1. Input and output : The algorithm must accept zero or more
inputs and must produce at least one output.
2. Definiteness : Each step of algorithm must be clear and
unambiguous.
3. Effectiveness : Every step must be basic and essential.
4. Finiteness : Total number of steps used in algorithm should be
finite.

1.7. What do you mean by asymptotic notations ?

Ans. Asymptotic notation is a shorthand way to represent the fastest
possible and slowest possible running times for an algorithm.

1.8. Differentiate between Big Oh(O) and theta() notations with

suitable examples.
Ans.
S. No. O-Notation (Upper bound) -Notation (Same order)
1. Big-oh is formal method of -notation bounds a function to
expressing the upper bound of within constant factors.
an algorithm’s running time.
2. It is the measure of the longest It is the measure of the average
amount of time it could possibly amount of time it could possibly
take for the algorithm to take for the algorithm to
complete. complete.
3. For example : Time complexity For e xample : Worst case
of searching an element in complexity of insertion sort is
array is O(n). (n2).

1.9. Why are asymptotic notations important ?

 More pdf : www.motivationbank.in
SQ–3 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Ans. Asymptotic notations are important because :

i. They give a simple characterization of algorithm efficiency.
ii. They allow the comparisons of the performances of various
algorithms.

1.10. Solve the given recurrence T(n) = 4T (n/4) + n.

AKTU 2015-16, Marks 02
Ans. T(n) = 4T (n/4) + n
We will map this equation with
T(n) = aT (n/b) + f(n)
a = 4 and b = 4, f(n) = n
Now, nlogba = nlog44 = n
i.e., case 2 of Master’s theorem is applied. Thus, resulting solution is
T(n) = (nlogba log n)
= (n log n)

1.11. What is a heap ? What are the different types of heaps ?

Ans. Heap : A heap is a specialized tree-based data structure that satisfies
the heap property.
Different types of heap are :
1. Min-heap
2. Binary heap
3. Binomial heap
4. Fibonacci heap
5. Max-heap
6. Radix heap

1.12. Justify why quick sort is better than merge sort ?

AKTU 2015-16, Marks 02
Ans. Theoretically both quick sort and merge sort take O (n log n) time
and hence time taken to sort the elements remains same. However,
quick sort is superior to merge sort in terms of space.
Quick sort is in-place sorting algorithm where as merge sort is not
in-place. In-place sorting means, it does not use additional storage
space to perform sorting. In merge sort, to merge the sorted arrays
it requires a temporary array and hence it is not in-place.

1.13. Why counting sort is stable ?

Ans. Counting sort is stable because numbers with the same value appear
in the output array in the same order as they do in the input array.

1.14. What are the fundamental steps involved in algorithmic

problem solving ? AKTU 2016-17, Marks 02
 More pdf : www.motivationbank.in
2 Marks Questions SQ–4 B (CS/IT-Sem-5)

Ans. Steps involved in algorithmic problem solving are :

i. Characterize the structure of optimal solution.
ii. Recursively define the value of an optimal solution.
iii. By using bottom-up technique, compute value of optimal solution.
iv. Compute an optimal solution from computed information.
1.15. Write recursive function to find nth Fibonacci number.
AKTU 2016-17, Marks 02
Ans. int fibo(int num)
{
if (num == 0)
{
return 0;
}
else if (num == 1)
{
return 1;
}
else
{
return(fibo(num – 1) + fibo(num – 2));
}
}
1.16. Write the names of various design techniques of algorithm.
AKTU 2016-17, Marks 02
Ans. Various design techniques of algorithm are :
1. Divide and conquer
2. Greedy approach
3. Dynamic programming
4. Branch and bound
5. Backtracking algorithm

1.17. Solve the following recurrence using master method :

T(n) = 4T (n/3) + n2
AKTU 2017-18, Marks 02

Ans. T(n) =4T(n/3) + n2

a= 4, b = 3, f(n) = n2
nlogba = nlog34 = n1.261
f(n) = (nlogba + E)
Now, af(n/b)  c f(n)
4
f(n)  c f(n)
3
4 2
n  cn2
3
 More pdf : www.motivationbank.in
Design and Analysis of AlgorithmsSQ–5 B (CS/IT-Sem-5)

4
c=
3
Hence, T(n) =  (n2)

1.18. Name the sorting algorithm that is most practically used

and also write its time complexity.
AKTU 2017-18, Marks 02
Ans. Quick sort algorithm is most practically used in sorting.
Time complexity of quick sort is O(n log n).

1.19. Find the time complexity of the recurrence relation

T(n) = n + T(n/10) + T(7n/5)
AKTU 2017-18, Marks 02
Ans.

n n

n 7n 3n
10 5 2
log n
n 7n 7n 49n 9n
100 50 50 25 4

n 7n 7n 49n 7n 49n 49n 343n 27n

1000 500 500 250 500 250 250 125 8

30 n 31 n 32 33
T(n) = 0
 1  1 n  3 n + ................+ log n times
2 2 2 2
=  (n log n)

1.20. What is priority queue ?

AKTU 2015-16, Marks 02
Ans. A priority queue is a collection of elements such that each element
has been assigned a priority. The order in which the elements are
deleted and processed according to the following rules :
i. An element of higher priority is processed before any element of
lower priority.
A 1 B 2 C 2 D 3
Fig. 1.
ii. Two elements with the same priority are processed according to
the order in which they were added to queue. This can be
implemented as linked list, or as 2D array.
 More pdf : www.motivationbank.in
2 Marks Questions SQ–6 B (CS/IT-Sem-5)

1.21. How do you compare the performance of various

algorithms ? AKTU 2019-20, Marks 02
Ans. To compare the performance of various algorithms first we measure
its performance which depends on the time taken and the size of
the problem. For this we measure the time and space complexity of
various algorithms which is divided into different cases such as
worst case, average case and best case.

1.22. Take the following list of functions and arrange them in

ascending order of growth rate. That is, if function g(n)
immediately follows function f(n) in your list, then it should
be the case that f(n) is O(g(n)).
f1(n) = n2.5, f2(n) = 2n, f3(n) = n + 10, f4(n) = 10n, f5(n) = 100n,
and f6(n) = n2 log n AKTU 2019-20, Marks 02
Ans. f3(n) = f4(n) = f5(n) < f2(n) < f6(n) < f1(n)

1.23. Rank the following by growth rate :

n, 2 log n , log n, log (log n), log2 n, (log n)log n, 4, (3/2)n, n!

AKTU 2018-19, Marks 02

Ans. Rank in increasing order of growth rate is given as :
n
 3
4, log n, log (log n), log2 n, (log n)log n, log n, 2 log n , n, n!,  
 2

1.24. What do you mean by stability of a sorting algorithm ?

Explain its application. AKTU 2018-19, Marks 02
Ans. Stability of a sorting algorithm : Let A be an array, and let < be
a strict weak ordering on the elements of A.
Sorting algorithm is stable if :
i < j and A[i]  A[j] i.e., A[i] comes before A[j].
Stability means that equivalent elements retain their relative
positions, after sorting.
Application : One application for stable sorting algorithms is sorting
a list using a primary and secondary key. For example, suppose we
wish to sort a hand of cards such that the suits are in the order
clubs, diamonds, hearts, spades and within each suit, the cards are
sorted by rank. This can be done by first sorting the cards by rank
(using any sort), and then doing a stable sort by suit.


 More pdf : www.motivationbank.in
SQ–7 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

2 Advanced
Data Structures
(2 Marks Questions)

2.1. What are the advantages of red-black tree over binary

search tree ?
Ans. The red-black tree is a self balanced tree which ensures the worst
case time complexity as O(log N) with N numbers of nodes and
worst case time complexity of binary search tree is O(N). Hence,
searching of any node in RB-tree becomes efficient than the binary
search tree.

2.2. What is the largest possible number of internal nodes in a

red-black tree with black height k ?
Ans. Consider a red-black tree with black height k. If every node is
black, then total number of internal nodes is 2k – 1.

2.3. Discuss the different kinds of rotations in RB-tree.

Ans. Rotations in RB-tree :
1. Left rotation : When left rotation on node ‘x’ is made then its right
child becomes its parent.
2. Right rotation : When right rotation on node ‘y’ is made then its
left child becomes its parent.

2.4. What is the running time of RB-Delete ?

Ans. Running time of RB-delete is O(log n) time.

2.5. Draw BSTs of height 2, 3 and 4 on the set of keys {10, 4, 5, 16,
1, 17, 21}
Ans. Keys = {10, 4, 5, 16, 1, 17, 21}
BST of height 2 :
10

4 17

1 5 16 21
 More pdf : www.motivationbank.in
2 Marks Questions SQ–8 B (CS/IT-Sem-5)

BST of height 3 :
5

4 10

1 17

16 21
BST of height 4 :
5

4 10

1 16

17

16

2.6. Draw all legal B-trees of minimum degree 2 that represent

{10, 12, 13, 14, 15}.
Ans. Keys : 10, 12 ,13, 14, 15
Minimum degree : 2
Insert 10 : 10

Insert 12 : 10 12

Insert 13 : 10 12 13

12

Split,
10 13

12

Insert 14 :
10 13 14
 More pdf : www.motivationbank.in
SQ–9 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

12

Insert 15 :
10 13 14 15

12 14

Split,
10 13 15

2.7. What are the operations performed for mergeable heaps ?

Ans. Various operations performed for mergeable heaps are :
i. Make-heap ( ) ii. Insert (H, x)
iii. Minimum (H) iv. Extract-min (H)
v. Union (H1, H2)
2.8. Explain binomial heap with properties.
AKTU 2015-16, Marks 02
Ans. Binomial heap is a type of data structure which keeps data sorted
and allows insertion and deletion in amortized time.
Properties of binomial heap :
i. There are 2k nodes.
ii. The height of the tree is k.
 k
iii. There are exactly nodes at depth i for i = 0, 1, …. , k (this is why
 i 
the tree is called a “binomial” tree).
iv Ro ot has degre e k (childre n) and its childre n are B k-1 ,
Bk-2, …, B0 from left to right.
2.9. Discuss the application of Fibonacci heaps.
Ans. Application of Fibonacci heaps :
i. In Dijkstra’s algorithm for computing shortest path.
ii. In Prim’s algorithm for finding minimum spanning tree.
2.10. What is a disjoint set ?
Ans. A disjoint set is a data structure S = {S1 ,….., Sk}, or a collection of
disjoint dynamic sets. Each set has a representative element, which
never changes unless union with another set.

2.11. Define binary heap. AKTU 2016-17, Marks 02

Ans. The binary heap data structure is an array that can be viewed as a
complete binary tree. Each node of the binary tree corresponds to
an element of the array. The tree is completely filled on all levels
except possibly lowest.
 More pdf : www.motivationbank.in
2 Marks Questions SQ–10 B (CS/IT-Sem-5)

2.12. Differentiate between complete binary tree and binary tree.

AKTU 2017-18, Marks 02
Ans.
S. No. Complete binary tree Binary tree

1. In a complete binary tree every A binary tree is a tree data

level, except possibly the last is structure in which each node has
completely filled, and all nodes at most two children, which are
in the last level are as far left referred to as the left child and
as possible. the right child.
2 It can have between 1 and 2h–1 It can have between 2h+1 and
nodes at the last level h. 2h+1 – 1 nodes at the last level h.

2.13. What is advantage of binary search over linear search?

Also, state limitations of binary search.
AKTU 2019-20, Marks 02
Ans. Advantages of binary search over linear search :
1. Input data needs to be sorted in binary search but not in linear
search.
2. Linear search does the sequential access whereas binary search
access data randomly.
3. Time complexity of linear search is O(n) where binary search has
time complexity O(log n).
Limitation of binary search :
1. List must be sorted.
2. It is more complicated to implement and test.
2.14. Prove that if n > = 1, then for any n-key B-tree of height h
and minimum degree t > = 2, h < = logt ((n + 1)/2).
AKTU 2018-19, Marks 02
Ans. Proof :
1. The root contains at least one key.
2. All other nodes contain at least t – 1 keys.
3. There are at least 2 nodes at depth 1, at least 2t nodes at depth 2, at
least 2ti – 1 nodes at depth i and 2th – 1 nodes at depth h.
h
 th  1
n  1 + (t – 1)
i1
 2t i 1
= 1 + 2(t – 1) 
 t  1 
= 2th – 1

4. So th  (n + 1)/2 as required.
Taking log both sides we get,
h  logt (n + 1)/2


 More pdf : www.motivationbank.in
SQ–11 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

3 Graph Algorithm
(2 Marks Questions)

3.1. Describe the general method for divide and conquer.

Ans. In divide and conquer method, a given problem is :
i. Divided into smaller sub-problems.
ii. These sub-problems are solved independently.
iii. Combining all the solutions of sub-problems into a solution of the
large problem.

3.2. Give various applications of divide and conquer.

Ans. Various applications divide and conquer strategy are :
i. Binary search
ii. Merge sort
iii. Quick sort
iv. Strassen’s matrix multiplication
v. Finding maximum and minimum element

3.3. Describe convex hull problem.

Ans. There exists a set of points on a plane which is said to be convex if
for any two points A and B in the set, the entire line segment with
the end points at A and B belongs to the set.

3.4. Define activity selection problem.

Ans. Activity selection problem is a problem of scheduling a resource
among several competing activity.

3.5. List out the disadvantages of divide and conquer algorithm.

AKTU 2016-17, Marks 02
Ans. Disadvantages of divide and conquer algorithm :
i. Recursion is slow.
ii. Algorithm becomes complicated for large value of n.

3.6. When we can say that optimization problem has overlapping

sub-problem ?
Ans. When a recursive algorithm revisits the same problem over and
over again, we say that optimization problem has overlapping
sub-problems.
 More pdf : www.motivationbank.in
2 Marks Questions SQ–12 B (CS/IT-Sem-5)

3.7. Define greedy technique.

Ans. Greedy technique solves problem by making the choice that seems
best at the particular moment.
Greedy technique works if a problem exhibit two properties :
i. Greedy choice property ii. Optimal sub-structure
3.8. Compare between greedy method and divide and conquer
algorithm.
Ans.
S. No. Greedy algorithm Divide and Conquer
1. It is used to obtain optimum It is used to obtain solution
solution. to given problem.
2. In greedy method, a set In this technique, the problem
of feasible solution is is divided into smaller
generated and optimum sub-problems.
solution is picked up.

3.9. How greedy method is used to solve the optimization

problem ?
Ans. In greedy technique, the solution is computed through a sequence
of steps. In each step, the partially constructed solution is expanded.
This process is repeated until a complete solution to the problem is
reached.

3.10. Give comparison between Prim’s and Kruskal’s algorithm.

Ans.
S. No. Prim’s algorithm Kruskal’s algorithm

1. It is used to obtain minimum It is also use d to o btain

spanning tree (MST). minimum spanning tre e
(MST).
2. It starts to build MST from any of It starts to build MST from
the node. minimum weighted edge in the
graph.
3. Run faster in dense graph. Run faster in sparse graph.

3.11. Discuss principle of optimality.

AKTU 2016-17, Marks 02

Ans. Principle of optimality states that in an optimal sequence of decisions

or choices, each subsequence must also be optimal.

3.12. Define bottleneck spanning tree.

 More pdf : www.motivationbank.in
SQ–13 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Ans. A bottleneck spanning tree T of an undirected graph G is a spanning

tree of G whose largest edge weight is minimum over all spanning
trees of G.
3.13. Can the minimum spanning tree change ?
Ans. Yes, the spanning tree will change only in the case when some of
the weights of edges are changed.
3.14. Briefly explain the Prim’s algorithm.
AKTU 2016-17, Marks 02
Ans. First it chooses a vertex and then chooses an edge with smallest
weight incident on that vertex. The algorithm involves following
steps :
Step 1 : Choose any vertex V1 of G.
Step 2 : Choose an edge e1 =V1V2 of G such that V2  V1 and e1 has
smallest weight among the edge e of G incident with V1.
Step 3 : If edges e1, e2, ......., ei have been chosen involving end
points V1, V2, ............., V i+1, choose an edge e i+1 = V jVk with
Vj = {V1 ....... Vi+1} and Vk  {V1 ...........Vi+1} such that ei+1 has
smallest weight among the edges of G with precisely one end in
{V1 ............... Vi+1}.
Step 4 : Stop after n – 1 edges have been chosen. Otherwise goto
step 3.
3.15. Explain element searching techniques using divide and
conquer approach. AKTU 2015-16, Marks 02
Ans. Binary search is a searching technique which uses divide and
conquer.
In each step, the algorithm compares the input element x with the
value of the middle element in array. If the values match, return
the index of middle. Otherwise, if x is less than the middle element,
then recursively call the algorithm for left side of middle element,
else recursively call the algorithm for right side of middle element.
3.16. What are greedy algorithms ? Explain their characteristics ?
AKTU 2019-20, Marks 02
Ans. Greedy algorithms : Greedy algorithms are shortsighted in their
approach in the sense that they take decisions on the basis of
information at hand without worrying about the effect these
decisions may have in the future.
Characteristics of greedy algorithm :
1. Greedy algorithms are most efficient.
2. For every instance of input greedy algorithms makes a decision
and continues to process further set of input.
3. The other input values at the instance of decision are not used in
further processing.
 More pdf : www.motivationbank.in
2 Marks Questions SQ–14 B (CS/IT-Sem-5)

3.17. Define feasible and optimal solution.

AKTU 2019-20, Marks 02
Ans. Feasible solution : A feasible solution is a set of values for the
decision variables that satisfies all of the constraints in an
optimization problem. The set of all feasible solutions defines the
feasible region of the problem.
Optimal solution : An optimal solution is a feasible solution where
the objective function reaches its maximum (or minimum) value.

3.18. Compare adjacency matrix and linked adjacency lists

representation of graph with suitable example/diagram.
AKTU 2018-19, Marks 02
Ans.
S. No. Adjacency matrix Linked adjacency list

1. An adjacency matrix is a Linke d adjace ncy list is a

square matrix used to collection of unordered lists used
represent a finite graph. to represent a finite graph.
2. The elements of the matrix Each list describes the set of
indicate whether pairs of adjacent vertices in the graph.
vertices are adjacent or not in
the graph.
3. Space complexity in the worst Space complexity in the worst
case is O(|V|2). case is O(|V|+|E|).

For example : Consider the graph :

0 1
2
4 3
Using adjacency matrix : Using adjacency list :
0 1 2 3 4
0 1 4 /
0 0 1 0 0 1
1 0 / 2 /
1 1 0 1 0 0
2 0 1 0 1 0 2 1 3 /

3 0 0 1 0 1 3 4 / 2 /

4 1 0 0 1 0 4 3 / 0 /


 More pdf : www.motivationbank.in
SQ–15 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Dynamic Programming,

4 Backtracking and
Branch and Bound
(2 Marks Questions)

4.1. Describe dynamic programming.

Ans. Dynamic programming is a technique for solving problems with
overlapping subproblems.
4.2. Give two differences between dynamic programming and
divide and conquer techniques.
Ans.
S. No. Dynamic programming Divide and conquer
1. In dynamic programming, many The problem is divided into
decision sequences are generated small subproblems. These
and all the o ve rlapping sub- subproble ms are solve d
instances are considered. independently.
2. In dynamic computing, duplication In this method, duplications in
in solutions is avoided totally. subsolutions are neglected.

4.3. Explain dynamic programming. How it is different from

greedy approach ? AKTU 2015-16, Marks 02
OR
Differentiate between greedy technique and dynamic
programming. AKTU 2017-18, Marks 02
Ans. Dynamic programming is a technique for solving problems with
overlapping subproblems.
In dynamic programming, problems are solved by using divide and
conquer method but in greedy approach problem are solved by
making the choice that seems best at the particular moment.
Difference :
S. No. Dynamic programming Greedy method
1. Dynamic programming can A greedy algorithm is one that at
be thought o f as ‘smart’ a given point in time makes a local
recursion. It often requires optimization.
one to break down a problem
into smaller components that
can be cached.
 More pdf : www.motivationbank.in
2 Marks Questions SQ–16 B (CS/IT-Sem-5)

2. Dynamic programming Greedy algorithms have a local

would solve all dependent choice of the subproblem that will
subproblems and then select lead to an optimal answer.
one that would lead to an
optimal solution.

4.4. State the single source shortest path.

AKTU 2017-18, Marks 02
Ans. Single source shortest path problem states that in a given graph
G = (V, E) we can find a shortest path from given source vertex
s  V to every vertex v  V.

4.5. What is the purpose of Floyd-Warshall’s algorithm ?

Ans. The purpose of Floyd-Warshall’s algorithm is to find the shortest
path between all pairs of vertices in a graph. It uses a number of
matrices of size n × n where n is the number of vertices.

4.6. What are the constraints used for solving the problem in
backtracking algorithm ?
Ans. There are two types of constraints used to solve the problem
in backtracking :
i. Explicit constraints
ii. Implicit constraints

4.7. Write the major advantage of backtracking algorithm.

Ans. The major advantage of backtracking algorithm is that we can
realize the fact that the partial vector generated does not lead to an
optimal solution. In such a situation that vector can be ignored.

4.8. State the difference between backtracking and branch and

bound. AKTU 2015-16, 2016-17, 2017-18; Marks 02
Ans.
S. No. Backtracking Branch and bound

1. Solution for backtracking is In this method, it is not necessary

trace d using de pth first to use depth first search fo r
search. obtaining the solution, even the
breadth first search, best first
search can be applied.
2. Typically decision problems Typically optimization problems can
can be solve d using be solved using branch and bound.
backtracking.
 More pdf : www.motivationbank.in
SQ–17 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

4.9. What is the running time complexity of 8-Queens problem ?

AKTU 2016-17, Marks 02
Ans. The running time complexity of 8-Queens problem is O(P(n)n!)
where P(n) is polynomial in n.

4.10. Define graph colouring. AKTU 2017-18, Marks 02

Ans. Graph colouring is a problem of colouring each vertex in graph in
such a way that no two adjacent vertices have same colour and
m-colours are used.

4.11. What is optimal colouring ?

Ans. If the degree of given graph is d then we colour it with
d + 1 colours. The least number of colours needed to colour the
graph is called its chromatic number. Such type of colouring is
called optimal colouring.

4.12. What is Travelling Salesman Problem (TSP) ?

Ans. Travelling salesman problem is a problem in which a salesman
visits ‘m’ cities, in a way that all cities must be visited at once and in
the end, he returns to the city from where he started, with minimum
cost.

4.13. Mention the three common techniques used in amortized

analysis and write the characteristics of one of its
technique.
Ans. Three common techniques used in amortized analysis are :
i. Aggregate method
ii. Accounting method
iii. Potential method
Characteristics of aggregate method are :
i. The amortized cost is T(n)/n per operation.
ii. It gives the average performance of each operation in the worst case.

4.14. Find out Hamiltonian cycles in complete graph having ‘n’

vertices. AKTU 2015-16, Marks 02
Ans. The answer depends on how we think that two Hamiltonian cycles
are equal. Take K3K3 as example. Let’s call its vertices 11, 22 and 33.
Then do we consider 1  2  3  11  2  3  1 same as 2  3 
1  22  3  1  2 ? If yes, 1  2  3  11  2  3  1 is the same
as 2  3  1  22  3  1  2, which is the same as 3  1  2 
33  1  2  3 in K3K3. Then we will have two Hamiltonian cycles
1  2  3  11  2  3  1 and 1  3  2  11  3  2  1.
Moreover, if we consider 1  2  3  11  2  3  1 and 1  3 
2  11  3  2  1 being the same because the second one is
 More pdf : www.motivationbank.in
2 Marks Questions SQ–18 B (CS/IT-Sem-5)

obtained by reversing direction the first one, then we have only

one Hamiltonian cycle in K3K3.
For general KnKn, it’s the same. 1  2  ...  n  11  2  ... 
n  1 is the same as 2  ...  n  1  22  ...  n  1  2 is the
same as ... n  1  2  ...  nn  1  2  ...  n. And the 1  2
 ...  n  11  2  ...  n  1 and 1  n  ...  2  11  n 
...  2  1 being the same because the second one is obtained by
reversing direction the first one. So, we have altogether 12(n – 1)!
12(n – 1)! Hamiltonian cycles in KnKn.

4.15. Find the subsets of sum of following problem. Given total

elements are (S) = {4, 2, 7, 6, 8} and maximum SUM is (X) = 8.
AKTU 2015-16, Marks 02
Ans.
Initially subset Sum(x) = 8
= {s}
4 4 Add next element
4, 2 66<8 Add next element
4, 2, 7 13 Sum exceeds X = 8
Hence backtrack
4, 2, 6 12 Sum exceeds X = 8
 Backtrack
4, 2, 8 14 Sum exceeds X = 8
 Not feasible
Hence backtrack
4, 2 Add next element
4, 7 11 Backtrack
2, 7 9 Backtrack
2, 6 8 Solutio n obtained as
sum = 8 = X.
Also
2 Backtrack
8 Backtrack and add element
8 Solutio n obtained as
sum = 8 = X

4.16. Explain application of graph colouring problem.

AKTU 2018-19, Marks 02
Ans. Application of graph colouring problem :
1. Sudoku : Sudoku is a variation of graph colouring problem where
every cell represents a vertex. There is an edge between two vertices
if they are in same row or same column or same block.
2. Register allocation : In compiler optimization, register allocation
is the process of assigning a large number of target program variables
 More pdf : www.motivationbank.in
SQ–19 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

onto a small number of CPU registers. This problem is also a graph

colouring problem.
3. Bipartite graphs : We can check if a graph is bipartite or not by
colouring the graph using two colours. If a given graph is
2-colourable, then it is bipartite, otherwise not.
4. Map colouring : Geographical maps of countries or states where
no two adjacent cities cannot be assigned same colour.

4.17. Define principle of optimality. When and how dynamic

programming is applicable ? AKTU 2018-19, Marks 02
Ans. Principle of optimality : Refer Q. 3.11, Page SQ–12B, Unit-3,
Two Marks Questions.
Dynamic programming is mainly applicable where the solution of
one sub-problem is needed repeatedly. In this procedure, the
solutions of sub-problems are stored in a table, so that there is no
need to re-compute the sub-problems and can be directly accessed
from the table if required.


 More pdf : www.motivationbank.in
2 Marks Questions SQ–20 B (CS/IT-Sem-5)

5 Selected Topics
(2 Marks Questions)

5.1. Define Fast Fourier Transformation (FFT).

Ans. The Fast Fourier Transform (FFT) is a algorithm that computes a
Discrete Fourier Transform (DFT) of n-length vector in O(n log n)
time. In the FFT algorithm, we apply the divide and conquer
approach to polynomial evaluation by observing that if n is even,
we can divide a degree (n – 1) polynomial.

5.2. Discuss the term string matching.

Ans. String matching algorithms, are an important class of string
algorithms that try to find a place where one or several strings (also
called patterns) are found within a larger string or text.

5.3. Write about Knuth-Morris-Pratt algorithm.

Ans. It is the first linear time string matching algorithm. In this, we have
to examine all the characters in the text and pattern at least once.
The running time is O(n + m).

5.4. Differentiate between decision problem and optimization

problem.
Ans.
S. No. Decision problem Optimization problem

i. Any problem for which Any problem that involves the

answer is either yes or no identification of optimal cost is called
is called decision problem. optimization problem.
ii. The algorithm for decision Typically optimization problems can
problem is called decision be solved using branch and bound.
algorithm.

5.5. Define P, NP and NP-complete in decision problems.

AKTU 2015-16, 2016-17; Marks 02
Ans. P-polynomial time : These problems can be solved in polynomial
time, which take time like O(n), O(n2), O(n3) like finding maximum
 More pdf : www.motivationbank.in
SQ–21 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

element in an array or to check whether a string is palindrome or

not are P problems.
Non deterministic polynomial time : These problem cannot be
solved in polynomial time like TSP (Travelling Salesman Problem)
or subset sum are NP problem.
But NP problems are checkable in polynomial time means that
given a solution of a problem, we can check that whether the
solution is correct or not in polynomial time.
NP-complete : The group of problems which are both in NP and
NP-hard are known as NP-complete problem.

5.6. When a language is said to be NP-complete and NP-hard ?

Ans. A language L  {0, 1}* is NP-complete, if it satisfies the following
two properties :
i. L  NP and
ii. For every L  NP, L  PL
If a language L satisfies property (ii), but not necessarily property
(i), we say that L is NP-hard.

5.7. Describe satisfiability.

Ans. A given boolean formula is satisfiable if there is a way to assign
truth values (0 or 1) to the variable such that the final result is 1.

5.8. Define circuit satisfiability.

Ans. A combinational circuit is called circuit satisfiable if for set of inputs
applied, output of this circuit should always be one.

5.9. Name some NP-complete problems.

Ans. NP-complete problems are :
i. The 0/1 knapsack problem
ii. Hamiltonian cycle
iii. Travelling salesman problem

5.10. Define vertex cover.

Ans. A vertex cover to an undirected graph G = (V, E) is a subset of
V  V such that if edge (u, v)  G then u  V or v  V (or both). This
o ptimal verte x co ve r is the o ptimizatio n ve rsio n of an
NP-complete problem but it is not too hard to find a vertex cover
that is near optimal.

5.11. Name any three problems that cannot be solved by

polynomial time algorithm.
Ans. Three problems that cannot be solved by polynomial time
algorithm are :
1. Halting problem
2. Decision problem
3. Optimization problem
 More pdf : www.motivationbank.in
2 Marks Questions SQ–22 B (CS/IT-Sem-5)

5.12. What are polynomial-time solvable and polynomial-time

verifiable algorithms ?
Ans. Polynomial-time solvable algorithm :
An algorithm is said to be solvable in polynomial time if the number
of steps required to complete the algorithm for a given input is
O(nk) for some non-negative integer k where n is the complexity of
input.
Polynomial-time verifiable algorithm : A polynomial-time
algorithms A is said to be polynomial time verifiable if it has following
properties :
1. The input to A consists of an instance I of X(X is a decision problem)
and a string s such that the length of s is bounded by some
polynomials in the size of I.
2. The output of A is either yes or no.
3. If I is a negative instance of X. Then the output of A is ‘‘no’ regardless
of the value of s.
4. If I is a positive instance of X1 then there is at least one choice of s
for which A output ‘‘yes’’.

5.13. List three problems that have polynomial time algorithm.

Ans. Problems that have polynomial time algorithm are :
i. Binary search
ii. Evaluation of polynomial
iii. Sorting a list

5.14. List any two approximation algorithms.

Ans. Two approximation algorithms are :
i. Nearest neighbour algorithm
ii. Greedy algorithm
5.15. Mention the types of randomized algorithms.
Ans. Two types of randomized algorithms are :
i. Las Vegas algorithms
ii. Monte Carlo algorithm
5.16. Describe the advantages of randomized algorithms.
Ans. Advantages of randomized algorithms are :
i. These algorithms are simple to implement.
ii. These algorithms are many times efficient than traditional
algorithms.
5.17. What do you mean by clique ?
Ans. A clique in an undirected graph G = (V, E) is a subset
V  V of vertices, each pair of which is connected by an edge in E.
CLIQUE = {(G, R) : G is a graph with a clique of size R}.

5.18. Explain applications of FFT. AKTU 2019-20, Marks 02

 More pdf : www.motivationbank.in
SQ–23 B (CS/IT-Sem-5)
Design and Analysis of Algorithms

Ans. Application of Fast Fourier Transform :

1. Signal processing.
2. Image processing.
3. Fast multiplication of large integers.
4. Solving Poisson’s equation nearly optimally.

5.19. What do you mean by polynomial time reduction ?

AKTU 2019-20, Marks 02
Ans. A polynomial time reduction is a method for solving one problem
using another. For example, if a hypothetical subroutine solving
the second problem exists, then the first problem can be solved by
transforming or reducing it to inputs for the second problem and
calling the subroutine one or more times.

5.20. What are approximation algorithms ? What is meant by

p(n) approximation algorithms ?
AKTU 2018-19, Marks 02
Ans. Approximation algorithm : An approximation algorithm is a
way of dealing with NP-completeness for optimization problem.
This technique does not guarantee the best solution.
p(n) approximation algorithm : A is a p(n) approximate
algorithm if and only if for every instance of size n, the algorithm
achieves an approximation ratio of p(n). It is applied to both
maximization (0 < C(i)  C*(i)) and minimization (0 < C* (i)  C(i))
problem because of the maximization factor and costs are positive.
p(n) is always greater than 1.


Design and Analysis of Algorithms SP–1 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B. Tech.
(SEM. V) ODD SEMESTER THEORY
EXAMINATION, 2015-16
DESIGN AND ANALYSIS OF ALGORITHMS

Time : 3 Hours Max. Marks : 100

SECTION – A

Note : Attempt all parts. All parts carry equal marks. Write answer
of each part in short. (2 × 10 = 20)

1. a. Justify why quick sort is better than merge sort ?

b. What is priority queue ?

c. Find out Hamiltonian cycles in complete graph having ‘n’

vertices.

d. Explain binomial heap with properties.

e. Explain element searching techniques using divide and

conquer approach.

f. Find the subsets of sum of following problem. Given total

elements are (S) = {4, 2, 7, 6, 8} and maximum SUM is (X) = 8.

g. Explain dynamic programming. How it is different from

greedy approach ?

h. Solve the given recurrence T(n) = 4T (n/4) + n.

i. Differentiate between backtracking and branch and bound

programming approach.

j. Explain the P, NP and NP-complete in decision problems.

SECTION – B

Note : Attempt any five questions from this section : (10 × 5 = 50)
2. Explain insertion in red-black tree. Show steps for inserting
1, 2, 3, 4, 5, 6, 7, 8 and 9 into empty RB-tree.
Solved Paper (2015-16) SP–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
3. Discuss knapsack problem with respect to dynamic
programming approach. Find the optimal solution for
given problem, w (weight set) = {5, 10, 15, 20} and W
(Knapsack size) = 25 and v = {50, 60, 120, 100}.

4. What is heap sort ? Apply heap sort algorithm for sorting

1, 2, 3, 4, 5, 6, 7, 8, 9, 10. Also deduce time complexity of heap
sort.

5. Explain B-tree and insert elements B, Q, L, F into B-tree

Fig. 1 then apply deletion of elements F, M, G, D, B on
resulting B-tree.

G M P X

A C D E I K N O R S T U V Y Z

Fig. 1.

6. Write an algorithm for solving N-Queens problem. Show

the solution of 4-Queens problem using backtracking
approach.

7. Explain greedy single source shortest path algorithm with

example.

8. What is string matching algorithm ? Explain Rabin-Karp

method with examples.

9. Explain approximation algorithms with suitable examples.

SECTION – C

Note : Attempt any two questions from this section : (15 × 2 = 30)

10. What is Fibonacci heap ? Explain CONSOLIDATE operation

with suitable example for Fibonacci heap.

11. What is minimum spanning tree ? Explain Prim’s algorithm

and find MST of graph Fig. 2.
Design and Analysis of Algorithms SP–3 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6 7
b c d
4 9
2
a 11 i 14 e
4
7 6
5 10
h g f
1 2
Fig. 2.

12. Explain TSP (Travelling Salesman) problem with example.

Write an approach to solve TSP problem.


Solved Paper (2015-16) SP–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
SOLUTION OF PAPER (2015-16)

SECTION – A

Note : Attempt all parts. All parts carry equal marks. Write answer
of each part in short. (2 × 10 = 20)

1. a. Justify why quick sort is better than merge sort ?

Ans. Theoretically both quick sort and merge sort take O (n log n) time
and hence time taken to sort the elements remains same. However,
quick sort is superior to merge sort in terms of space.
Quick sort is in-place sorting algorithm where as merge sort is not
in-place. In-place sorting means, it does not use additional storage
space to perform sorting. In merge sort, to merge the sorted arrays
it requires a temporary array and hence it is not in-place.

b. What is priority queue ?

Ans. A priority queue is a collection of elements such that each element
has been assigned a priority. The order in which the elements are
deleted and processed according to the following rules :
i. An element of higher priority is processed before any element of
lower priority.
ii. Two elements with the same priority are processed according to
the order in which they were added to queue. This can be
implemented as linked list, or as 2D array.

c. Find out Hamiltonian cycles in complete graph having ‘n’

vertices.
Ans. The answer depends on how we think that two Hamiltonian cycles
are equal. Take K3K3 as example. Let’s call its vertices 11, 22 and 33.
Then do we consider 1  2  3  11  2  3  1 same as 2  3 
1  22  3  1  2 ? If yes, 1  2  3  11  2  3  1 is the same
as 2  3  1  22  3  1  2, which is the same as 3  1  2 
33  1  2  3 in K3K3. Then we will have two Hamiltonian cycles
1  2  3  11  2  3  1 and 1  3  2  11  3  2  1.
Moreover, if we consider 1  2  3  11  2  3  1 and 1  3 
2  11  3  2  1 being the same because the second one is
obtained by reversing direction the first one, then we have only
one Hamiltonian cycle in K3K3.
For general KnKn, it’s the same. 1  2  ...  n  11  2  ... 
n  1 is the same as 2  ...  n  1  22  ...  n  1  2 is the
same as ... n  1  2  ...  nn  1  2  ...  n. And the 1  2
 ...  n  11  2  ...  n  1 and 1  n  ...  2  11  n 
...  2  1 being the same because the second one is obtained by
reversing direction the first one. So, we have altogether 12(n – 1)!
12(n – 1)! Hamiltonian cycles in KnKn.
Design and Analysis of Algorithms SP–5 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
d. Explain binomial heap with properties.
Ans. Binomial heap is a type of data structure which keeps data sorted
and allows insertion and deletion in amortized time.
Properties of binomial heap :
i. There are 2k nodes.
ii. The height of the tree is k.
 k
iii. There are exactly nodes at depth i for i = 0, 1, …. , k (this is why
 i 

the tree is called a “binomial” tree).

iv Ro ot has degre e k (childre n) and its childre n are B k-1 ,
Bk-2, …, B0 from left to right.

e. Explain element searching techniques using divide and

conquer approach.
Ans. Binary search is a searching technique which uses divide and
conquer.
In each step, the algorithm compares the input element x with the
value of the middle element in array. If the values match, return
the index of middle. Otherwise, if x is less than the middle element,
then recursively call the algorithm for left side of middle element,
else recursively call the algorithm for right side of middle element.
f. Find the subsets of sum of following problem. Given total
elements are (S) = {4, 2, 7, 6, 8} and maximum SUM is (X) = 8.
Ans.
Initially subset Sum(x) = 8
= {s}
4 4 Add next element
4, 2 66<8 Add next element
4, 2, 7 13 Sum exceeds X = 8
Hence backtrack
4, 2, 6 12 Sum exceeds X = 8
 Backtrack
4, 2, 8 14 Sum exceeds X = 8
 Not feasible
Hence backtrack
4, 2 Add next element
4, 7 11 Backtrack
2, 7 9 Backtrack
2, 6 8 Solutio n obtained as
sum = 8 = X.
Also
2 Backtrack
8 Backtrack and add element
8 Solutio n obtained as
sum = 8 = X
Solved Paper (2015-16) SP–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
g. Explain dynamic programming. How it is different from
greedy approach ?
Ans. Dynamic programming is a technique for solving problems with
overlapping subproblems.
In dynamic programming, problems are solved by using divide and
conquer method but in greedy approach problem are solved by
making the choice that seems best at the particular moment.
Difference :
S. No. Dynamic programming Greedy method
1. Dynamic programming can A greedy algorithm is one that at
be thought o f as ‘smart’ a given point in time makes a local
recursion. It often requires optimization.
one to break down a problem
into smaller components that
can be cached.
2. Dynamic programming Greedy algorithms have a local
would solve all dependent choice of the subproblem that will
subproblems and then select lead to an optimal answer.
one that would lead to an
optimal solution.

h. Solve the given recurrence T(n) = 4T (n/4) + n.

Ans. T(n) = 4T (n/4) + n
We will map this equation with
T(n) = aT (n/b) + f(n)
a = 4 and b = 4, f(n) = n
Now, nlogba = nlog44 = n
i.e., case 2 of Master’s theorem is applied. Thus, resulting solution is
T(n) = (nlogba log n)
= (n log n)

i. Differentiate between backtracking and branch and bound

programming approach.
Ans.
S. No. Backtracking Branch and bound

1. Solution for backtracking is In this method, it is not necessary

trace d using de pth first to use depth first search fo r
search. obtaining the solution, even the
breadth first search, best first
search can be applied.
2. Typically decision problems Typically optimization problems can
can be solve d using be solved using branch and bound.
backtracking.
Design and Analysis of Algorithms SP–7 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
j. Explain the P, NP and NP-complete in decision problems.
Ans. P-polynomial time : These problems can be solved in polynomial
time, which take time like O(n), O(n2), O(n3) like finding maximum
element in an array or to check whether a string is palindrome or
not are P problems.
Non deterministic polynomial time : These problem cannot be
solved in polynomial time like TSP (Travelling Salesman Problem)
or subset sum are NP problem.
But NP problems are checkable in polynomial time means that
given a solution of a problem, we can check that whether the
solution is correct or not in polynomial time.
NP-complete : The group of problems which are both in NP and
NP-hard are known as NP-complete problem.

SECTION – B

Note : Attempt any five questions from this section : (10 × 5 = 50)
2. Explain insertion in red-black tree. Show steps for inserting
1, 2, 3, 4, 5, 6, 7, 8 and 9 into empty RB-tree.
Ans. Insertion :
i. We begin by adding the node as we do in a simple binary search
tree and colouring it red.
RB-INSERT(T, z)
1. y  nil [T]
2. x  root [T]
3. while x  nil [T]
4. do y  x
5. if key[z] < key [x]
6. then x  left [x]
7. else x  right [x]
8. p[z]  y
9. if y = nil [T]
10. then root [T]  z
11. else if key [z] < key[y]
12. then left [y] z
13. else right [y]  z
14. left [z]  nil[T]
15. right [z]  nil[T]
16. colour [z]  RED
17. RB-INSERT-FIXUP(T, z)
ii. Now, for any colour violation, RB-INSERT-FIXUP procedure is
used.
RB-INSERT-FIXUP(T, z)
1. while colour [p[z]] = RED
2. do if p[z] = left[p [p [z]]]
3. then y  right[p [p [z]]]
4. if colour[y] = RED
Solved Paper (2015-16) SP–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
5. then colour[p [z]]  BLACK  case 1
6. colour[y]  BLACK  case 1
7. colour[p [p [z]]] RED  case 1
8. z  p [p [ z]]  case 1
9. else if z = right[p[z]]
10. then z  p[z]  case 2
11. LEFT-ROTATE(T, z)  case 2
12. colour[p[z]]  BLACK  case 3
13. colour[p[p[z]]]  RED  case 3
14. RIGHT-ROTATE(T, p[p[z]])  case 3
15. else (same as then clause with “right” and “left”exchanged)
16. colour[root[T]]  BLACK
Numerical :
B
Insert 1 : 1

B
1
Insert 2 : R
2
Insert 3 :
B
B
1 Case 2 2
R
2 1 3
R R R
3

Insert 4 :
B B
2 2

1 3 1 3
R R B B 4
4
R R

Insert 5 :

B B
2 Case 2 2 B
B B
1 3 1 4
R B R
4 3 5
R
5 R
Design and Analysis of Algorithms SP–9 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 6 :

B B
2 2
B B B R
1 4 1 4
R
3 5 3 5
R
R 6 B B 6
R
Insert 7 :
B B
2 Case 2 2
B B B R
1 4 1 4
R B
3 5 B 3 6
R R
R 6 R 5 7
R
7
Insert 8 :
B B
2 2
B R B R
1 4 1 4
B R
B 3 6 B 3 6
R B
R 5 7 B 5 7
R R
8 8
Insert 9 :
B
2 Case 2 2
B R
1 4 1 4
R
B 3 6 3 6
B
R 5 7 5 8
R
8 7 9
R
9

3. Discuss knapsack problem with respect to dynamic

programming approach. Find the optimal solution for
given problem, w (weight set) = {5, 10, 15, 20} and W
(Knapsack size) = 25 and v = {50, 60, 120, 100}.
Ans. Knapsack problem with respect to dynamic programming
approach :
Dynamic 0/1-knapsack(v, w, n, W) :
1. for (w = 0 to W) V[0, w] = 0
2. for (i = 1 to n)
Solved Paper (2015-16) SP–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
3. for (w = 0 to W)
4. if (w[i]  w) then
5. V[i, w] = max{V[i – 1, w], v[i] + V[i – 1, w – w[i]]};
6. else V[i, w] = V[i – 1, w];
7. return V[n, W];
Now, as we know that V [n, W] is the total value of selected items,
the can be placed in the knapsack. Following steps are used
repeatedly to select actual knapsack item.
Let, i = n and k = W then
while (1 > 0 and k > 0)
{
if (V[i, k]  V[i – 1, k]) then
mark ith item as in knapsack
i = i – 1 and k = k – wi // selection of ith item
else
i=i–1 //do not select ith item
}
Numerical :
w = {5, 10, 15, 20}
W = 25
v = {50, 60, 120, 100}
Initially,
Item wi vi
I1 5 50
I2 10 60
I3 15 120
I4 20 100

Taking value per weight ratio, i.e., pi = vi/wi

Item wi vi pi = vi/wi
I1 5 50 10
I2 10 60 6
I3 15 120 8
I4 20 100 5

Now, arrange the value of pi in decreasing order.

Design and Analysis of Algorithms SP–11 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Item wi vi pi = vi/wi
I1 5 50 10
I3 15 120 8
I2 10 60 6
I4 20 100 5
Now, fill the knapsack according to decreasing value of pi.
First we choose item I1 whose weight is 5, then choose item I3 whose
weight is 15. Now the total weight in knapsack is 5 + 15 = 20. Now, next
item is I2 and its weight is 10, but we want only 5. So, we choose fractional
part of it, i.e.,

5
15 25

5
The value of fractional part of I2 is,
60
= × 5 = 30
10
Thus, the maximum value is,
= 50 + 120 + 3 = 200

4. What is heap sort ? Apply heap sort algorithm for sorting

1, 2, 3, 4, 5, 6, 7, 8, 9, 10. Also deduce time complexity of heap
sort.
Ans. Heap sort :
1. Heap sort is a comparison based sorting technique based on binary
heap data structure.
2. Heap sort finds the largest element and puts it at the end of array,
then the second largest item is found and this process is repeated
for all other elements.
MAX-HEAPIFY (A, i) :
1. i  left [i]
2. r  right [i]
3. if l  heap-size [A] and A[l] > A[i]
4. then largest  l
5. else largest  i
6. if r  heap-size [A] and A[r] > A [largest]
7. then largest  r
8. if largest  i
9. then exchange A[i]  A[largest]
10. MAX-HEAPIFY [A, largest]
Solved Paper (2015-16) SP–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
HEAP-SORT(A) :
1. BUILD-MAX-HEAP (A)
2. for i  length [A] down to 2
3. do exchange A[1]  A[i]
4. heap-size [A]  heap-size [A] – 1
5. MAX-HEAPIFY (A, 1)
BUILD-MAX-HEAP (A)
1. heap-size (A)  length [A]
2. for i  (length [A]/2) down to 1 do
3. MAX-HEAPIFY (A, i)
We can build a heap from an unordered array in linear time.
Average case and worst case complexity :
1. We have seen that the running time of BUILD-HEAP is O(n).
2. The heap sort algorithm makes a call to BUILD-HEAP for creating
a (max) heap, which will take O(n) time and each of the (n – 1)
calls to MAX-HEAPIFY to fix up the new heap (which is created
after exchanging the root and by decreasing the heap size).
3. We know ‘MAX-HEAPIFY’ takes time O(log n).
4. Thus the total running time for the heap sort is O(n log n).
Numerical : Since the given problem is already in sorted form.
So, there is no need to apply any procedure on given problem.

5. Explain B-tree and insert elements B, Q, L, F into B-tree

Fig. 1 then apply deletion of elements F, M, G, D, B on
resulting B-tree.

G M P X

A C D E I K N O R S T U V Y Z

Fig. 1.
Ans. B-tree :
A B-tree of order m is an m-ary search tree with the following
properties :
1. The root is either leaf or has atleast two children.
2. Each node, except for the root and the leaves, has between m/2
and m children.
3. Each path from the root to a leaf has the same length.
4. The root, each internal node and each leaf is typically a disk block.
5. Each internal node has upto (m – 1) key values and upto m children.
Design and Analysis of Algorithms SP–13 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Numerical :
Insertion :

G M P X

A C D E I K N O R S T U V Y Z

Assuming, order of B-tree = 5

G M P T X

A C D E I K N O R S U V Y Z

G M T X

A C D E I K N O R S U V Y Z

Insert B :

G M T X

A B C D E I K N O R S U V Y Z

C G M T X

A B D E I K N O R S U V Y Z
Solved Paper (2015-16) SP–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert Q :

C G M T X

A B D E I K N O Q R S U V Y Z

Insert L :
P

C G M T X

A B D E I K L N O Q R S U V Y Z
Insert F :

C G M T X

A B D E F I K L N O Q R S U V Y Z

Deletion :
Delete F :

C G M T X

A B D E I K L N O Q R S U V Y Z

Delete M :

C G L T X

A B D E I K N O Q R S U V Y Z
Design and Analysis of Algorithms SP–15 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Delete G :
P

C L T X

A B D E I K N O Q R S U V Y Z
Delete D :

C L P T X

A B E I K N O Q R S U V Y Z
Delete B :
E L P T X

A C I K NO Q R S UV Y Z

6. Write an algorithm for solving N-Queens problem. Show

the solution of 4-Queens problem using backtracking
approach.
Ans.
1. In N-Queens problem, the idea is to place queens one by one in
different columns, starting from the leftmost column.
2. When we place a queen in a column, we check for clashes with
already placed queens.
3. In the current column, if we find a row for which there is no clash,
we mark this row and column as part of the solution.
4. If we do not find such a row due to clashes then we backtrack and
return false.
Algorithm for N-Queens problem :
N-Queens are to be placed on an n × n chessboard so that no two
attack i.e., no two Queens are on the same row, column or diagonal.
PLACE (k, i)
1. for j  1 to k – 1
2. do if (x(j) = i) or Abs (x[j] – i) = (Abs (j – k))
3. then return false
4. return true
Place (k, i) returns true if a queen can be placed in the kth row and
ith column otherwise return false.
x[ ] is a global array whose first k – 1 values have been set. Abs(r)
returns the absolute value of r.
Solved Paper (2015-16) SP–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
N-Queens (k, n)
1. for i  1 to n
2. do if PLACE (k, i)
3. then x[k]  i
4. if k = n, then print x[1 …. N]
5. else N-Queens (k + 1, n)
[Note : For 8-Queen problem put n = 8 in the algorithm.]
4-Queens problem :
1. Suppose we have 4 × 4 chessboard with 4-queens each to be placed
in non-attacking position.

1 2 3 4
1
2
3
4

Fig. 2.
2. Now, we will place each queen on a different row such that no two
queens attack each other.
3. We place the queen q1 in the very first accept position (1, 1).
4. Now if we place queen q2 in column 1 and 2 then the dead end is
encountered.
5. Thus, the first acceptable position for queen q2 is column 3 i.e., (2, 3)
but then no position is left for placing queen q3 safely. So, we
backtrack one step and place the queen q2 in (2, 4).
6. Now, we obtain the position for placing queen q3 which is (3, 2). But
later this position lead to dead end and no place is found where
queen q2 can be placed safely.

1 2 3 4
1 q1
2 q2

3 q3

Fig. 3.
7. Then we have to backtrack till queen q1 and place it to (1, 2) and
then all the other queens are placed safely by moving queen q2 to
(2, 4), queen q3 to (3, 1) and queen q4 to (4, 3) i.e., we get the solution
< 2, 4, 1, 3>. This is one possible solution for 4-queens problem.
Design and Analysis of Algorithms SP–17 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1 2 3 4
1 q1

2 q2

3 q3

4 q4

Fig. 4.
8. For other possible solution the whole method is repeated for all
partial solutions. The other solution for 4-queens problem is <3, 1,
4, 2> i.e.,
1 2 3 4
1 q1
2 q
2

3 q3

4 q4

Fig. 5.
9. Now, the implicit tree for 4-queen for solution <2, 4, 1, 3> is as
follows :
10. Fig. 6 shows the complete state space for 4-queens problem. But we
can use backtracking method to generate the necessary node and
stop if next node violates the rule i.e., if two queens are attacking.
0

1 5
q1 q1

× × × × ×
2 3
q1 q1 q1
q2 q2 q2

q1 q1
q2 q2
q3 q3

× × q1
q2
q3
q4

Fig. 6.
Solved Paper (2015-16) SP–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
7. Explain greedy single source shortest path algorithm with
example.
Ans.
1. Dijkstra’s algorithm, is a greedy algorithm that solves the single
source shortest path problem for a directed graph G = (V, E) with
non-negative edge weights, i.e., we assume that w(u, v)  0 each
edge (u, v)  E.
2. Dijkstra’s algorithm maintains a set S of vertices whose final
shortest path weights from the source s have already been
determined.
3. That is, for all vertices v  S, we have d[v] = (s, v).
4. The algorithm repeatedly selects the vertex u  V – S with the
minimum shortest path estimate, inserts u into S, and relaxes all
edges leaving u.
5. We maintain a priority queue Q that contains all the vertices in
v – s, keyed by their d values.
6. Graph G is represented by adjacency list.
7. Dijkstra’s always chooses the “lightest or “closest” vertex in V – S
to insert into set S that it uses as a greedy strategy.
For example :
Initialize :
 50  S:{}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20

100
5 4
 60 
Extract min (1) :
 50  S : {1}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20

100
5 4
 60 
All edges leaving (1) :
10 
50 S : {1}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20
10  30 100
100
5 4
60
100 30
Design and Analysis of Algorithms SP–19 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Extract min(2) :
50 60 S : {1, 2}
2 3
10 Q:1 2 3 4 5
10 0    
1 30 20
10 60 30 100
100
5 4
60
All edges leaving (2) :
50 60 S : {1, 2}
2 3
10 Q:1 2 3 4 5
10 0    
1 30 20
10 60 30 100
100
5 4 60 30 100
60 30
100
Extract min(4) :
10 60
50 S : {1, 2, 4}
2 3
10 Q:1 2 3 4 5
10 0    
0 1 30 20
10 60 30 100
100
5 4 60 30 100
60
100 30
All edges leaving (4) :
50 50 S : {1, 2, 4}
2 3 Q:1 2 3 4 5
10
0    
10
1 30 20 10 60 30 100
100 60 30 100
5 4
60 50
100 30
Extract min(3) :
10 S : {1, 2, 4, 3}
50
2 3 50 Q:1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50
100
Solved Paper (2015-16) SP–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
All edges leaving (3) :
10 S : {1, 2, 4, 3, }
50
2 3 50 Q:1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50 60
100
Extract min(5) :
10 S : {1, 2, 4, 3, 5}
50
2 3 50 Q: 1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50
60
60

Shortest path
2 3
10

1 30 10 20

5 4

8. What is string matching algorithm ? Explain Rabin-Karp

method with examples.
Ans. String matching algorithm :
String matching is a process of finding one or more occurrences of
a pattern in a text.
String matching problem :
Given a text array T[1 .. n] of n character and a pattern array P[1 ..
m] of m characters.
The problem is to find an integer s, called valid shift where 0  s < n
– m and T[s + 1 .... s + m] = P[1 ... m].
We further assume that the elements of P and T are characters
drawn from a finite alphabet such as {0, 1} or {A, B, … Z, a, b, …,
z}.
The Rabin-Karp algorithm :
The Rabin-Karp algorithm states that if two strings are equal, their
hash values are also equal. This also uses elementary number-
Design and Analysis of Algorithms SP–21 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
theoretic notions such as the equivalence of two numbers module
a third.
Rabin-Karp-Matcher (T, P, d, q)
1. n length [T]
2. m length [P]
3. h dm–1 mod q
4. p0
5. to  0
6. for i  1 to m
7. do p  (dp + p [i]) mod q
8. t0  (dt0 + T[i]) mod q
9. for s  0 to n–m
10. do if p = ts
11. then if p [1....m] = T[s + 1......s + m]
12. then “pattern occurs with shift” s
13. if s < n – m
14. then ts + 1  (d(ts – T [s + 1] h) + T[s + m +1]) mod q
Example of Rabin-Karp method : Working modulo q = 11, how
many spurious hits does the Rabin-Karp matcher encounter in the
text T = 3141592653589793 when looking for the pattern p = 26
Given, p = 26 and q = 11
Now we divide 26 by 11 i.e.,
Remainder is 4 and m = 2.
We know m denotes the length of p.
T 3 1 4 1 5 9 2 6 5 3 5 8 9 7 9 3

Now we divide 31 by 11, and get remainder is 9.

Similarly, 14 by 11 and get remainder is 3.
So, continue this step till last i.e., 93 is divided by 11 and get remainder
is 5. After that we will store all remainder in a table.
9 3 8 4 4 4 4 10 9 2 3 1 9 2 5
Now we find valid matching.
3 1 4 1 5 9 2 6 5 3 5 8 9 7 9 3

Valid matching
9 3 8 4 4 4 4 10 9 2 3 1 9 2 5

Spurious
hit
The number of spurious hits is 3.

9. Explain approximation algorithms with suitable examples.

Ans. Approximation algorithm :
1. An approximation algorithm is a way of dealing with NP-
completeness for optimization problem. This technique does not
guarantee the best solution.
Solved Paper (2015-16) SP–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. The best of an approximation algorithm is to come as close as possible
to the optimum value in a reasonable amount of time which is at
most polynomial time.
3. Let c(i) be the cost of solution produced by approximate algorithm
and c*(i) be the cost of optimal solution for some optimization
problem instance i.
4. For minimization and maximization problem, we are interested in
finding a solution of a given instance i in the set of feasible solutions,
such that c(i) / c*(i) and c*(i) / c(i) be as small as possible respectively.
5. We say that an approximation algorithm for the given problem
instance i, has a ratio bound of p(n) if for any input of size n, the
cost c of the solution produced by the approximation algorithm is
within a factor of p(n) of the cost c* of an optimal solution. That is
max(c(i) / c*(i), c*(i) / c(i)) ≤ p(n)
The definition applies for both minimization and maximization
problems.
6. p(n) is always greater than or equal to 1. If solution produced by
approximation algorithm is true optimal solution then clearly we
have p(n) = 1.
7. For a minimization problem, 0 < c*(i) < c(i), and the ratio c(i) / c*(i)
gives the factor by which the cost of the approximate solution is
larger than the cost of an optimal solution.
8. Similarly, for a maximization problem, 0 < c(i)  c*(i), and the ratio
c*(i) / c(i) gives the factor by which the cost of an optimal solution
is larger than the cost of the approximate solution.
Travelling salesman problem (TSP) is an example of approximation
algorithm.

SECTION – C

Note : Attempt any two questions from this section : (15 × 2 = 30)

10. What is Fibonacci heap ? Explain CONSOLIDATE operation

with suitable example for Fibonacci heap.
Ans. Fibonacci heap :
1. A Fibonacci heap is a set of min-heap-ordered trees.
2. Trees are not ordered binomial trees, because
a. Children of a node are unordered.
b. Deleting nodes may destroy binomial construction.
3. Fibonacci heap H is accessed by a pointer min[H] to the root of a
tree containing a minimum key. This node is called the minimum
node.
4. If Fibonacci heap H is empty, then min[H] = NIL.
Design and Analysis of Algorithms SP–23 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
min(H)

12 7 4 8

14 15 9 10 11 6

20
Fig. 7.
CONSOLIDATE operation :
CONSOLIDATE(H)
1. for i  0 to D(n[H])
2. do A[i]  NIL
3. for each node w in the root list of H
4. do x  w
5. d  degree[x]
6. while A[d]  NIL
7. do y  A[d]  Another node with the same degree as x.
8. if key[x] > key[y]
9. then exchange x  y
10. FIB-HEAP-LINK(H, y, x)
11. A[d]  NIL
12. dd+1
13. A[d]  x
14. min[H]  NIL
15. for i  0 to D(n[H])
16. do if A[i]  NIL
17. then add A[i] to the root list of H
18. if min[H] = NIL or key [A[i]] < key[min[H]]
19. then min[H]  A[i]
FIB-HEAP-LINK(H, y, x)
1. remove y from the root list of H
2. make y a child of x, incrementing degree[x]
3. mark[y]  FALSE
11. What is minimum spanning tree ? Explain Prim’s algorithm
and find MST of graph Fig. 8.
6 7
b c d
4 9
2
a 11 i 14 e
4
7 6
5 10
h g f
1 2
Fig. 8.
Solved Paper (2015-16) SP–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans. Minimum spanning tree :
1. Given a connected weighted graph G, it is often desired to create a
spanning tree T for G such that the sum of the weights of the tree
edges in T is as small as possible.
2. Such a tree is called a minimum spanning tree and represents the
‘‘cheapest’’ way of connecting all the nodes in G.
3. There are number of techniques for creating a minimum spanning
tree for a weighted graph but the most famous methods are Prim’s
and Kruskal’s algorithm.
Prim’s algorithm :
First it chooses a vertex and then chooses an edge with smallest
weight incident on that vertex. The algorithm involves following
steps :
Step 1 : Choose any vertex V1 of G.
Step 2 : Choose an edge e1 =V1V2 of G such that V2  V1 and e1 has
smallest weight among the edge e of G incident with V1.
Step 3 : If edges e1, e2, ......., ei have been chosen involving end
points V1, V2, ............., Vi+1, choose an edge ei+1 = VjVk with Vj = {V1
....... Vi+1} and Vk  {V1 ...........Vi+1} such that ei+1 has smallest
weight among the edges of G with precisely one end in {V1 ...............
Vi+1}.
Step 4 : Stop after n – 1 edges have been chosen. Otherwise goto
step 3.
Numerical :
6 7
b c d
4 2 9
4
a 11 i 14 e
6
5 7 10
h g f
1 2
Fig. 9.
Let a be the source node. Select edge (a, b) as distance between
edge (a, b) is minimum.
b
4

a
Now, select edge (b, c)
6
b c
4

a
Design and Analysis of Algorithms SP–25 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Now, select edge (c, i)
6
b c
4
2
a i
Now, select edge (i, g)
6
b c
4
2
a i
6

Now, select edge (g, h)

6
b c
4
2
a i
6

h g
1
Now, select edge (g, f)
6
b c
4
2
a i
6
1 2
h g f

Now, select edge (f, e)

6
b c
4
2
a i e
6
10
1 g 2
h f
Solved Paper (2015-16) SP–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Now, select edge (e, d)
6
b c
4 2
a i d
6

h 1 g 2 f e

Thus, we obtained MST for Fig. 9.

12. Explain TSP (Travelling Salesman) problem with example.

Write an approach to solve TSP problem.
Ans. Travelling Salesman Problem (TSP) :
Travelling salesman problem is the problem to find the shortest
possible route for a given set of cities and distance between the
pair of cities that visits every city exactly once and returns to the
starting point.
For example :
The following graph shows a set of cities and distance between
every pair of cities :
City A
20
10 15
City D
25 30

City B City C
35
If salesman starting city is A. Then a TSP tour in the graph is
ABDCA
Cost of the tour = 10 + 25 + 30 + 15 = 80 Units.
Backtracking approach is used to solve TSP problem.
Backtracking algorithm for the TSP :
1. Let G be the given complete graph with positive weights on its
edges.
2. Use a search tree that generates all permutations of V = {1 ... n},
specifically the one illustrated in Fig. 10 for the case n = 3.
Design and Analysis of Algorithms SP–27 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
0

1 2 3

1 2 3

2 3 1 3 1 2

1, 2 1,3 2,1 2,3 3,1 3,2

3 2 3 1 2 1

1,2,3 1,3,2 2,1,3 2,3,1 3,1,2 3,2,1

Fig. 10.
3. A node at depth i of this tree (the root is at depth 0) stores an
i-permutation of {1, ..., n}. A leaf stores a permutation of {1, ..., n},
which is equivalent to saying that it stores a particular Hamiltonian
cycle (tour) of G.
4. For the travelling salesman problem, we will not do any static
pruning on this tree, we will do dynamic pruning, during the search.


Design and Analysis of Algorithms SP–1 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B. Tech.
(SEM. V) ODD SEMESTER THEORY
EXAMINATION, 2016-17
DESIGN AND ANALYSIS OF ALGORITHMS

Time : 3 Hours Max. Marks : 100

Section-A

1. Attempt all parts. All parts carry equal marks. Write answer of
each part in short. (2 × 10 = 20)

a. List out the disadvantages of divide and conquer algorithm.

b. What are the fundamental steps involved in algorithmic

problem solving ?

c. Write recursive function to find nth Fibonacci number.

d. Define binary heap.

e. Briefly explain the Prim’s algorithm.

f. Define principle of optimality.

g. Write the names of various design techniques of algorithm.

h. Differentiate between branch and bound and backtracking

technique.

i. What is the running time complexity of 8-Queens problem ?

j. Define P, NP and NP-complete in decision problem.

Section-B

Note : Attempt any five questions from this section. (10 × 5 = 50)
2. Explain the concepts of quick sort method and analyze its
complexity with suitable example.

3. Explain the concept of merge sort with example.

4. Insert the nodes 15, 13, 12, 16, 19, 23, 5, 8 in empty red-black
tree and delete in the reverse order of insertion.
Solved Paper (2016-17) SP–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
5. Write short note on Dijkstra’s algorithm shortest paths
problems.

6. Write pseudocode for 8-Queens problem.

7. Write non-deterministic algorithm for sorting.

8. What is backtracking ? Write general iterative algorithm

for backtracking.

9. Differentiate NP-complete with NP-hard.

Section-C

Note : Attempt any two questions from this section. (15 × 2 = 30)
10. i. State Bellman-Ford algorithm.
ii. Consider following instance for simple knapsack problem.
Find the solution using greedy method.
N=8
P = {11, 21, 31, 33, 43, 53, 55, 65}
W = {1, 11, 21, 23, 33, 43, 45, 55}
M = 110

11. What is travelling salesman problem ? Find the solution of

following travelling salesman problem using branch and
bound method.
  20 30 10 11
 
15  16 4 2 
Cost matrix =  3 5  2 4
 
19 6 6  3
16 4 7 16  


12. Prove that three colouring problem is NP-complete.


Design and Analysis of Algorithms SP–3 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
SOLUTION OF PAPER (2016-17)

Section-A

1. Attempt all parts. All parts carry equal marks. Write answer of
each part in short. (2 × 10 = 20)

a. List out the disadvantages of divide and conquer algorithm.

Ans. Disadvantages of divide and conquer algorithm :
i. Recursion is slow.
ii. Algorithm becomes complicated for large value of n.

b. What are the fundamental steps involved in algorithmic

problem solving ?
Ans. Steps involved in algorithmic problem solving are :
i. Characterize the structure of optimal solution.
ii. Recursively define the value of an optimal solution.
iii. By using bottom-up technique, compute value of optimal solution.
iv. Compute an optimal solution from computed information.

c. Write recursive function to find nth Fibonacci number.

Ans. int fibo(int num)
{
if (num == 0)
{
return 0;
}
else if (num == 1)
{
return 1;
}
else
{
return(fibo(num – 1) + fibo(num – 2));
}
}

d. Define binary heap.

Ans. The binary heap data structure is an array that can be viewed as a
complete binary tree. Each node of the binary tree corresponds to
an element of the array. The tree is completely filled on all levels
except possibly lowest.

e. Briefly explain the Prim’s algorithm.

Solved Paper (2016-17) SP–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans. First it chooses a vertex and then chooses an edge with smallest
weight incident on that vertex. The algorithm involves following
steps :
Step 1 : Choose any vertex V1 of G.
Step 2 : Choose an edge e1 =V1V2 of G such that V2  V1 and e1 has
smallest weight among the edge e of G incident with V1.
Step 3 : If edges e1, e2, ......., ei have been chosen involving end
points V1, V2, ............., V i+1, choose an edge e i+1 = V jVk with
Vj = {V1 ....... Vi+1} and Vk  {V1 ...........Vi+1} such that ei+1 has
smallest weight among the edges of G with precisely one end in
{V1 ............... Vi+1}.
Step 4 : Stop after n – 1 edges have been chosen. Otherwise goto
step 3.

f. Define principle of optimality.

Ans. Principle of optimality states that in an optimal sequence of decisions
or choices, each subsequence must also be optimal.

g. Write the names of various design techniques of algorithm.

Ans. Various design techniques of algorithm are :
1. Divide and conquer
2. Greedy approach
3. Dynamic programming
4. Branch and bound
5. Backtracking algorithm

h. Differentiate between branch and bound and backtracking

technique.
Ans.
S. No. Branch and bound Backtracking

1. In this metho d, it is not Solution for backtracking is traced

necessary to use depth first using depth first search.
se arch fo r o btaining the
solution, even the breadth
first search, best first search
can be applied.
2. Typically o ptimizatio n Typically decision problems can be
problems can be solved using solved using backtracking.
branch and bound.

i. What is the running time complexity of 8-Queens problem ?

Ans. The running time complexity of 8-Queens problem is O(P(n)n!)
where P(n) is polynomial in n.
Design and Analysis of Algorithms SP–5 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
j. Define P, NP and NP-complete in decision problem.
Ans. P-polynomial time : These problems can be solved in polynomial
time, which take time like O(n), O(n2), O(n3) like finding maximum
element in an array or to check whether a string is palindrome or
not are P problems.
Non deterministic polynomial time : These problem cannot be
solved in polynomial time like TSP (Travelling Salesman Problem)
or subset sum are NP problem.
But NP problems are checkable in polynomial time means that
given a solution of a problem, we can check that whether the
solution is correct or not in polynomial time.
NP-complete : The group of problems which are both in NP and
NP-hard are known as NP-complete problem.

Section-B

Note : Attempt any five questions from this section. (10 × 5 = 50)
2. Explain the concepts of quick sort method and analyze its
complexity with suitable example.
Ans. Quick sort :
Quick sort works by partitioning a given array A[p ... r] into two
non-empty subarray A[p ... q – 1] and A [q + 1 ... r] such that every
key in A[p ... q – 1] is less than or equal to every key in A[q + 1 ... r].
Then the two subarrays are sorted by recursive calls to quick sort.
Quick_Sort (A, p, r)
1. If p < r then
2. q  Partition (A, p, r)
3. Recursive call to Quick_Sort (A, p, q – 1)
4. Recursive call to Quick_Sort (A, q + 1, r)
As a first step, Quick sort chooses as pivot one of the items in the
array to be sorted. Then array is partitioned on either side of the
pivot. Elements that are less than or equal to pivot will move toward
the left and elements that are greater than or equal to pivot will
move toward the right.
Partition (A, p, r)
1. x  A[r]
2. i  p – 1
3. for j  p to r – 1
4. do if A[j]  x
5. then i  i + 1
6. then exchange A[i]  A[j]
7. exchange A[i + 1]  A[r]
8. return i + 1
Example : Given array to be sorted
3 1 4 1 5 9 2 6 5 3 5 8 9
Sort the array A using quick sort algorithm.
Solved Paper (2016-17) SP–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Step 1 : The array is Pivoted about it first element i.e., Pivot (P) = 3

3 1 4 1 5 9 2 6 5 3 5 8 9
P
Step 2 : Find first element larger then pivot (make underline) and
find element not larger than pivot from end make over line.

3 1 4 1 5 9 2 6 5 3 5 8 9
P
Underline Overline
Step 3 : Swap these element and scan again.
3 1 3 1 5 9 2 6 5 4 5 8 9
P
Array after swapping

3 1 3 1 5 9 2 6 5 4 5 8 9
P
Underline Overline
Apply swapping,

3 1 3 1 2 9 5 6 5 4 5 8 9
Again apply scanning,

3 1 3 1 2 9 5 6 5 4 5 8 9

Overline Underline
The pointers have crossed
i.e., overline on left of underlined
Then, in this situation swap pivot with overline.

2 1 3 1 3 9 5 6 5 4 5 8 9
P
Now, pivoting process is complete.
Step 4 : Recursively sort subarrays on each side of pivot.
Subarray 1 : 2 1 3 1

Subarray 2 : 9 5 6 5 1 5 8 9
First apply Quick sort for subarray 1.
Design and Analysis of Algorithms SP–7 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2 1 3 1
P
Underline Overline

2 1 1 3
P
Underline Overline
The pointers have crossed.
i.e., overline on left of underlined.
Swap pivot with overline
1 1 2 3
Sorted array
Now, for subarray 2 we apply Quick sort procedure.

9 5 6 5 4 5 8 9
P
Overline Underline
The pointer has crossed. Then swap pivot with overline.

8 5 6 5 4 5 9 9

Subarray 4
Subarray 3

8 5 6 5 4 5
P Overline
Swap overline with pivot.

5 5 6 5 4 8
P
5 5 6 5 4
Underline Overline
5 5 4 5 6
Overline on left of underlined.
Swap pivot with overline.

4 5 5 5 6
P
Now combine all the subarrays
Sorted array 1 1 2 3 3 4 5 5 5 6 8 9 9
Pivot
Solved Paper (2016-17) SP–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Analysis of complexity :
i. Worst case :
1. Let T(n) be the worst case time for quick sort on input size n. We
have a recurrence
T(n) = max (T(q) + T(n – q – 1)) + (n) ...(1)
o q n 1
where q ranges from 0 to n – 1, since the partition produces two
regions, each having size n – 1.
2. Now we assume that T(n)  cn2 for some constant c.
Substituting our assumption in eq. (1.19.1) we get
T(n)  max (cq2 + c(n – q – 1)2) + (n)
o q n 1

= max (q2 + (n – q – 1)2) + (n)

o q n 1
3. Since the second derivative of expression q2 + (n – q – 1)2 with
respect to q is positive. Therefore, expression achieves a maximum
over the range 0  q  n – 1 at one of the endpoints.
4. This gives the bound
max0  q  n – 1 (q2 + (n – q – 1)2) (n – 1)2 = n2 – 2n + 1
5. Continuing with the bounding of T(n) we get
T(n)  cn2 – c(2n – 1) + (n) cn2
6. Since we can pick the constant c large enough so that the c(2n – 1)
term dominates the (n) term. We have
T(n) = O(n2 )
7. Thus, the worst case running time of quick sort is (n2).
ii. Average case :
1. If the split induced of RANDOMIZED_PARTITION puts constant
fraction of elements on one side of the partition, then the
recurrence tree has depth (log n) and (n) work is performed at
each level.
2. This is an intuitive argument why the average case running time
of RANDOMIZED_QUICKSORT is (n log n).
3. Let T(n) denotes the average time required to sort an array of n
elements. A call to RANDOMIZED_ QUICKSORT with a 1 element
array takes a constant time, so we have T(1) = (1).
4. After the split RANDOMIZED_QUICKSORT calls itself to sort
two subarrays.
5. The average time to sort an array A[1 .. q] is T[q] and the average
time to sort an array A[q+1 .. n] is T[n – q]. We have
T(n) = 1/n (T(1) + T(n – 1) + n–1Sq=1 T(q)
T(n – q))) + (n) ...(1)
We know from worst-case analysis
T(1) =  (1) and T(n – 1) = O(n2)
T(n) = 1/n ( (1) + O(n2)) + 1/n n–1q=1 (r(q)
+ T(n – q)) + Q(n)
= 1/n n–1Sq=1 (T(q) + T(n – q)) + Q(n) ...(2)
= 1/n [2 n–1Sk=1(T(k))] +  (n)
= 2/n n–1Sk=1 (T(k)) +  (n) ...(3)
Design and Analysis of Algorithms SP–9 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6. Solve the above recurrence using substitution method. Assume
that T(n)  an n log n + b for some constants a > 0 and b > 0.
If we can pick ‘a’ and ‘b’ large enough so that n log n + b > T(1).
Then for n > 1, we have
T(n)  n–1k=1 2/n (ak log k + b) + (n)
= 2a/n n–1Sk=1 k log k – 1/8(n2) + 2b/n
(n – 1) + n ...(4)
At this point we are claiming that
n–1 2 2
k=1 k log k  1/2 n log n – 1/8(n )
Substituting this claim in the eq. (4), we get
T(n) 2a/n [1/2n2 log n – 1/8(n2)] + 2/n b(n – 1) +  (n)
 an log n – an/4 + 2b + (n) ...(5)
In the eq. (5), (n) + b and an/4 are polynomials and we can choose
‘a’ large enough so that an/4 dominates (n) + b.
We conclude that QUICKSORT’s average running time is (n log
n).

3. Explain the concept of merge sort with example.

Ans.
1. Merge sort is a sorting algorithm that uses the idea of divide and
conquer.
2. This algorithm divides the array into two halves, sorts them
separately and then merges them.
3. This procedure is recursive, with the base criteria that the number
of elements in the array is not more than 1.
Algorithm :
MERGE_SORT (a, p, r)
1. if p < r
2. then q (p + r)/2
3. MERGE-SORT (A, p, q)
4. MERGE-SORT (A, q + 1, r)
5. MERGE (A, p, q, r)
MERGE (A, p, q, r)
1. n1 = q – p + 1
2. n2 = r – q
3. Create arrays L [1 .........n1 + 1] and
R [1......n2 + 1]
4. for i = 1 to n1
do
L[i] = A [p + i – 1]
endfor
5. for j = 1 to n2
do
R[j] = A[q + j]
endfor
6. L[n1 + 1] = , R[n2 + 1] = 
Solved Paper (2016-17) SP–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
7. i = 1, j = 1
8. for k = p to r
do
if L[i]  R[j]
then A[k]  L[i]
i=i+1
else A[k] = R[j]
j= j+1
endif
endfor
9. exit
Example :
10, 25, 16, 5, 35, 48, 8
1. Divide into two halves : 10, 25, 16, 5 35, 48, 8
2. Consider the first part : 10, 25, 16, 5 again divide into two
sub- arrays
10 , 25 16 , 5

10, 25 5, 16

5, 10, 16, 25
3. Consider the second half : 35, 48, 5 again divide into two sub-
arrays
35 , 48 8

35, 48 8

8, 35, 48
4. Merge these two sorted sub-arrays,
5, 10, 16, 25 8, 35, 48

5, 8, 10, 16, 25, 35, 45

This is the sorted array.

4. Insert the nodes 15, 13, 12, 16, 19, 23, 5, 8 in empty red-black
tree and delete in the reverse order of insertion.
Ans. Insertion :
B
Insert 15 : 15
Design and Analysis of Algorithms SP–11 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B
R 15
Insert 13 : 13
Insert 12 :
B
B
15 13
R
R 13 12 15
12 R R

Insert 16 :
B B
13 Case 3 13

12 15 12 15
R R B B 16
16
R R

Insert 19 :

B B
13 Case 2 13
B B B
12 15 12 16
R B R
16 15 19
R
19 R

Insert 23 :

B B
13 13
B B B R
12 16 12 16
R
15 19 15 19
R
R 23 B B 23
R
Insert 5 :

B
B 13
R
12 16
B
5 15 19
R B
23
R
Solved Paper (2016-17) SP–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 8 :
B B
13 13
B R
B
B 12 16 8 16
R R
2nd 15 19 B
R 5 5 12 15 19
B B
23 B R
1st 8 23
R
R
Deletions :
Delete 8 :
B B
13 13
B R B R
8 16 12 16
R
5 12 15 19 B 5 R B 15 19 B
R B
R 23 23 R
Delete 5 :
B
B
13 13
R R B
R
12 16 12 16
R B B
5 15 19 15 19 B
B
23 R 23 R
Delete 23 :

B B
13 13
B 12 R B 12 R
16 16

B 15 19 B B 15 19 B

23 R

Delete 19 :
B B B
13 13 13
R Case 2 B
B 12 16 B 12 16 12 16 B
R
B 15 19 B B 15 15 R
Design and Analysis of Algorithms SP–13 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Delete 16 :
B B
13 13
B B
B 12 16 B 12 15

15 R

Delete 12 :
B B
13 13
B 12 15 B 15 B

Delete 13 :
B B
13 15
15 B
Delete 15 :
No tree

5. Write short note on Dijkstra’s algorithm shortest paths

problems.
Ans.
1. Dijkstra’s algorithm, is a greedy algorithm that solves the single
source shortest path problem for a directed graph G = (V, E) with
non-negative edge weights, i.e., we assume that w(u, v)  0 each
edge (u, v)  E.
2. Dijkstra’s algorithm maintains a set S of vertices whose final
shortest path weights from the source s have already been
determined.
3. That is, for all vertices v  S, we have d[v] = (s, v).
4. The algorithm repeatedly selects the vertex u  V – S with the
minimum shortest path estimate, inserts u into S, and relaxes all
edges leaving u.
5. We maintain a priority queue Q that contains all the vertices in
v – s, keyed by their d values.
6. Graph G is represented by adjacency list.
7. Dijkstra’s always chooses the “lightest or “closest” vertex in V – S
to insert into set S that it uses as a greedy strategy.
Dijkstra’s algorithm :
DIJKSTRA (G, w, s)
1. INITIALIZE-SINGLE-SOURCE (G, s)
2. s  
3. Q  V[G]
Solved Paper (2016-17) SP–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
4. while Q  
5. do u  EXTRACT-MIN (Q)
6. S  S  {u}
7. for each vertex v  Adj [u]
8. do RELAX (u, v, w)
RELAX (u, v, w) :
1. If d[u] + w(u, v) < d[v]
2. then d[v]  d[u] + w(u, v)
3. [v]  u

6. Write pseudocode for 8-Queens problem.

Ans. Pseudocode for N-Queens problem :
N-Queens are to be placed on an n × n chessboard so that no two
attack i.e., no two Queens are on the same row, column or diagonal.
PLACE (k, i)
1. for j  1 to k – 1
2. do if (x(j) = i) or Abs (x[j] – i) = (Abs (j – k))
3. then return false
4. return true
Place (k, i) returns true if a queen can be placed in the kth row and
ith column otherwise return false.
x[ ] is a global array whose first k – 1 values have been set. Abs(r)
returns the absolute value of r.
N-Queens (k, n)
1. for i  1 to n
2. do if PLACE (k, i)
3. then x[k]  i
4. if k = n, then print x[1 …. N]
5. else N-Queens (k + 1, n)
[Note : For 8-Queen problem put n = 8 in the algorithm.]

7. Write non-deterministic algorithm for sorting.

Ans. Non-deterministic algorithms are algorithm that, even for the same
input, can exhibit different behaviours on different runs, iterations
and executions.
N SORT(A, B) :
1. for i = 1 to n do
2. j = choice(1 . . . n)
3. if B[j] != 0 then failure
4. B[j] = A[i]
5. endfor
Design and Analysis of Algorithms SP–15 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6. for i = 1 to n – 1 do
7. if B[i] < B[i + 1] then failure
8. endfor
9. print(B)
10. success

8. What is backtracking ? Write general iterative algorithm

for backtracking.
Ans.
1. Backtracking is a general algorithm for finding all solutions to
some computational problems.
2. Backtracking is an important tool for solving constraint satisfaction
problems, such as crosswords, verbal arithmetic, and many other
puzzles.
3. It is often the most convenient (if not the most efficient) technique
for parsing, for the knapsack problem and other combinational
optimization problem.
4. It can be applied only for problems which admit the concept of a
‘‘partial candidate solution’’ and a relatively quick test of whether
it can possibly be completed to a valid solution.
Iterative backtracking algorithm :
algorithm ibacktrack (n)
// Iterative backtracking process
// All solutions are generated in x[1 : n] and
printed as soon as they are found
{
k = 1;
while (k != 0)
{
if ( there remains an untried x[k] in T(x[1], x[2], ..., x[k – 1])
and B_k(x[1], ..., x[k] ) is true )
{
if (x[1], ..., x[k] is a path to an answer node )
write (x[1 : k]);
k = k + 1; // Consider the next set
}
else
k = k – 1; // Backtrack to the previous set
}
}

9. Differentiate NP-complete with NP-hard.

Solved Paper (2016-17) SP–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans.
S. No. NP-complete NP-hard
1. An NP-complete problems NP-hard problems is one to which
is one to which every other an N P-co mple te pro blem is
polynomial-time non- Turing-reducible.
deterministic algorithm can
be reduced in polynomial
time.
2. NP-complete problems do NP-hard problems correspond to
no t co rrespo nds to an an NP-complete problem.
NP-hard problem.
3. NP-complete problems are NP-hard problems need not to be
exclusively decision decision problem.
problem.
4. N P-co mple te pro blems NP-hard problems do not have to
have to be in NP-hard and be in NP.
also in NP.
5. For example : 3-SAT For example : Halting problem is
vertex cover problem is NP-hard.
NP-complete.

Section-C
Note : Attempt any two questions from this section. (15 × 2 = 30)
10. i. State Bellman-Ford algorithm.
ii. Consider following instance for simple knapsack problem.
Find the solution using greedy method.
N= 8
P = {11, 21, 31, 33, 43, 53, 55, 65}
W = {1, 11, 21, 23, 33, 43, 45, 55}
M = 110
Ans.
i.
1. Bellman-Ford algorithm finds all shortest path length from a source
s  V to all v  V or determines that a negative-weight cycle exists.
2. Bellman-Ford algorithm solves the single source shortest path
problem in the general case in which edges of a given digraph G
can have negative weight as long as G contains no negative cycles.
3. This algorithm, uses the notation of edge relaxation but does not
use with greedy method.
4. The algorithm returns boolean TRUE if the given digraph contains
no negative cycles that are reachable from source vertex otherwise
it returns boolean FALSE.
Design and Analysis of Algorithms SP–17 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Bellman-Ford (G, w, s) :
1. INITIALIZE-SINGLE-SOURCE (G, s)
2. for each vertex i  1 to V[G] – 1
3. do for each edge (u, v) in E[G]
4. do RELAX (u, v, w)
5. for each edge (u, v) in E[G] do
6. do if d[u] + w(u, v) < d[v] then
7. then return FALSE
8. return TRUE
RELAX (u, v, w) :
1. If d[u] + w(u, v) < d[v]
2. then d[v]  d[u] + w(u, v)
3. [v]  u
If Bellman-Ford returns true, then G forms a shortest path tree,
else there exists a negative weight cycle.
ii.
N=8
W = {1, 11, 21, 23, 33, 43, 45, 55}
P = {11, 21, 31, 33, 43, 53, 55, 65}
M = 110
Now, arrange the value of Pi in decreasing order
N Wi Pi V i = Wi × P i
1 1 11 11
2 11 21 231
3 21 31 651
4 23 33 759
5 33 43 1419
6 43 53 2279
7 45 55 2475
8 55 65 3575
Now, fill the knapsack according to decreasing value of Pi. First we
choose item N = 1 whose weight is 1.
Then choose item N = 2 whose weight is 11.
Then choose item N = 3 whose weight is 21.
Now, choose item N = 4 whose weight is 23.
Then choose item N = 5 whose weight is 33.
Total weight in knapsack is = 1 + 11 + 21 + 23 + 33 = 89
Now, the next item is N = 6 and its weight is 43, but we want only
21 because M = 110.
So, we choose fractional part of it, i.e.,
The value of fractional part of N = 6 is,
2279
× 21 =1113
43
Solved Paper (2016-17) SP–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
21

33

23

21  110

11

Thus, the maximum value is,

= 11 + 231 + 651 + 759 + 1419 + 1113
= 4184

11. What is travelling salesman problem ? Find the solution of

following travelling salesman problem using branch and
bound method.
  20 30 10 11
 
15  16 4 2 
Cost matrix =  3 5  2 4
 
19 6 6  3
16 4 7 16  

Ans. Travelling salesman problem :
Travelling salesman problem is the problem to find the shortest
possible route for a given set of cities and distance between the
pair of cities that visits every city exactly once and returns to the
starting point.
Branch and bound method and backtracking approach are used to
solve TSP problem.
Numerical :
  20 30 10 11
15  16 4 2 
 
Cost matrix =  3 5  2 4 
 
19 6 6  3
16 4 7 16  
 
1. Reduce each column and row by reducing the minimum value from
each element in row and column.
Design and Analysis of Algorithms SP–19 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Row Column
1 3

10  10 20 0 1  10 17 0 1
2 13  14 2 0 12  11 2 0
2 1 3  0 2 0 3  0 2 =M
1
3 16 3 3  0 15 3 0  0
4 12 0 3 12  11 0 0 12 
2. So, total expected cost is : 10 + 2 + 2 + 3 + 4 + 1 + 3 = 25.
3. We have discovered the root node V1 so the next node to be expanded
will be V2, V3, V4, V5. Obtain cost of expanding using cost matrix for
node 2.
4. Change all the elements in 1st row and 2nd column.
     
12  11 2 0 
 
M2 =  0   0 2 
 
15  0  0 
11  0 12  
 
5. Now, reducing M2 in rows and columns, we get :
     
12  11 2 0 
 
M2 =  0   0 2 
 
15  0  0 
11  0 12  
 
 Total cost for M2 = 25 + 10 + 0 = 35
6. Similarly, for node 3, we have :
     
12   2 0 
 
M3 =  0 3  0 2 
 
15 3   0 
11 0  12  
 
7. Now, reducing M3, we get :
     
12   2 0 
 
M3 =  0 3  0 2 
 
15 3   0 
11 0  12  
 
 Total cost for M3 = 25 + 17 + 0 = 42
Solved Paper (2016-17) SP–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
8. Similarly, for node 4, we have :
     
12  11  0 
 
M4 =  0 3   2 
 
15 3 0  0 
11 0 0   
 
9. Now, reducing M4, we get :
     
12  11  0 
 
M4 =  0 3   2 
 
15 3 0  0 
11 0 0   
 
 Total cost = 25 + 0 + 0 = 25
10. Similarly, for node 5, we have :
    
12  11 2  

M5 =  0 3  0 
 
15 3 0  
11 0 0 12  

11. Now, reducing M5, we get :
    
10  9 0  

M5 =  0 3  0 
 
15 3 0  
11 0 0 12  

 Total cost = 25 + 1 + 2 = 28

1 V1 = 25

2 3 4 5 V 5 = 28

V2 = 35 V3 = 42 V4 = 25
Fig. 1.
12. Now, the promising node is V4 = 25. Now, we can expand V2, V3 and
V5. Now, the input matrix will be M4.
13. Change all the elements in 4th row and 2nd column.
Design and Analysis of Algorithms SP–21 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
    
12  11  0 

M6 =  0    2
 
    
11  0   

14. On reducing M6, we get :
    
12  11  0 

M6 =  0    2
 
    
11  0   

 Total cost = 25 + 3 + 0 = 28
15. Similarly, for node 3, we have :
     
12    0 
 
M7 =  0 3   2 
 
  3   0
11 0    
 
16. On reducing M7, we get :
     
12    0 
 
M7 =  0 3   2 
 
  3   0
11 0    
 
 Total cost = 25 + 0 + 0 = 25
17. Similarly, for node 5, we have :
     
12  11   
 
M8 =  0 3    
 
     
11 0 0   
 
18. On reducing M8, we get :
     
 1  0  
 
M8 =  0 3    
 
     
11 0 0   
 
Solved Paper (2016-17) SP–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
 Total cost = 25 + 0 + 11 = 36

2 3 4 5

2 3 5

V2 = 28 V 3 = 25 V 5 = 36
Fig. 2.
19. Now, promising node is V3 = 25. Now, we can expand V2 and V5.
Now, the input matrix will be M7.
20. Change all the elements in 3rd row and 2nd column.
     
12    0 
 
M9 =       
 
     0
11     
 
21. On reducing M9, we get :
     
 1    0
 
M9 =       
 
    0
 0    
 
 Total cost = 25 + 3 + 0 = 28
22. Similarly, for node 5, we have :
     
12     
 
M10 =       
 
  3   
11 0    
 
23. On reducing M10, we get :
     
 0    
 
M10 =       
 
  0   
11 0    
 
 Total cost = 25 + 2 + 12 + 3 = 42.
Design and Analysis of Algorithms SP–23 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1

2 3 4 5

2 3 5

2 5

V 2 = 28 V 5 = 42
Fig. 3.
24. Here V2 is the most promising node so next we are going to expand
this node further. Now, we are left with only one node not yet
traversed which is V5.
10 6 5 2 16
V1  V4  V3  V2  V5  V1
So, total cost of traversing the graph is :
10 + 6 + 5 + 2 + 16 = 39

12. Prove that three colouring problem is NP-complete.

Ans.
1. To show the problem is in NP, let us take a graph G(V, E) and a
colouring c, and checks in O(n2 ) time whether c is a proper colouring
by checking if the end points of every edge e  E have different
colours.
2. To show that 3-COLOURING is NP-hard, we give a polytime
reduction from 3-SAT to 3-COLOURING.
3. That is, given an instance  of 3-SAT, we will construct an instance
of 3-COLOURING (i.e., a graph G(V, E)) where G is 3-colourable iff
 is satisfiable.
4. Let  be a 3-SAT instance and C1, C2, ..., Cm be the clauses of 
defined over the variables {x1, x2, ..., xn}.
5. The graph G(V, E) that we will construct needs to capture two
things :
a. Somehow establish the truth assignment for x1, x2, ..., xn via the
colours of the vertices of G; and
b. Somehow capture the satisfiability of every clause Ci in .
6. To achieve these two goals, we will first create a triangle in G with
three vertices {T, F, B} where T stands for True, F for False and B
for Base.
7. Consider {T, F, B} as the set of colours that we will use to colour
(label) the vertices of G.
8. Since this triangle is part of G, we need 3 colours to colour G.
Solved Paper (2016-17) SP–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
9. Now we add two vertices vi, vi for every literal xi and create a

triangle B, vi , vi for every (vi , vi ) pair, as shown in Fig. 4.

T F

vn
B
v1
vn
v1
v2
v2

Fig. 4.
10. This construction captures the truth assignment of the literals.
11. Since if G is 3-colourable, then either vi or vi gets the colour T, and
we interpret this as the truth assignment to vi.
12. Now we need to add constraints to G to capture the satisfiability of
the clauses of .
a ab
abc
b

The output node

c
Fig. 5.
13. To do so, we introduce the Clause Satisfiability Gadget, (the OR-
gadget). For a clause Ci = (a  b  c), we need to express the OR of
its literals using our colours {T, F, B}.
14. We achieve this by creating a small gadget graph that we connect to
the literals of the clause. The OR-gadget is constructed as follows :
15. Consider this gadget graph as a circuit whose output is the node
labeled a  b  c. We basically want this node to be coloured T if Ci
is satisfied and F otherwise.
16. This is a two step construction : The node labelled a  b captures the
output of (a  b) and we repeat the same operation for ((a  b)  c).
If we play around with some assignments to a, b, c, we will notice
that the gadget satisfies the following properties :
a. If a, b, c are all coloured F in a 3-colouring, then the output node of
the OR-gadget has to be coloured F. Thus capturing the
unsatisfiability of the clause Ci = (a  b  c).
b. If one of a, b, c is coloured T, then there exists a valid
3-colouring of the OR-gadget where the output node is coloured T.
Thus again capturing the satisfiability of the clause.
Design and Analysis of Algorithms SP–25 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
17. Once we add the OR-gadget of every Ci in , we connect the output
node of every gadget to the Base vertex and to the False vertex of
the initial triangle, as follows :
a ab
abc T
b
B

c F
Fig. 6.
18. Now we prove that our initial 3-SAT instance  is satisfiable if and
only the graph G as constructed above is 3-colourable. Suppose  is
satisfiable and let (x1*, x2*, ..., xn* ) be the satisfying assignment.
19. If xi* is assigned True, we colour vi with T and vi with F (recall
they are connected to the Base vertex, coloured B, so this is a valid
colouring).
20. Since  is satisfiable, every clause Ci = (a b c) must be satisfiable,
i.e., at least of a, b, c is set to True. By the property of the OR-
gadget, we know that the gadget corresponding to Ci can be 3-
coloured so that the output node is coloured T.
21. And because the output node is adjacent to the False and Base
vertices of the initial triangle only, this is a proper 3-colouring.
22. Conversely, suppose G is 3-colourable. We construct an assignment
of the literals of  by setting xi to True if vi is coloured T and vice
versa.
23. Now consider this assignment is not a satisfying assignment to ,
then this means there exists at least one clause Ci = (a  b  c) that
was not satisfiable.
24. That is, all of a, b, c were set to False. But if this is the case, then the
output node of corresponding OR-gadget of Ci must be coloured F.
25. But this output node is adjacent to the False vertex coloured F; thus
contradicting the 3-colourability of G.
26. To conclude, we have shown that 3-COLOURING is in NP and that
it is NP-hard by giving a reduction from 3-SAT.
27. Therefore 3-COLOURING is NP-complete.


Design and Analysis of Algorithms SP–1 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B.Tech.
(SEM. V) ODD SEMESTER THEORY
EXAMINATION, 2017-18
DESIGN AND ANALYSIS OF ALGORITHM
Time : 3 Hours Max. Marks : 100

Note : Attempt all sections. Assume any missing data.

Section-A

1. Define/Explain the following : (2 × 10 = 20)

a. Differentiate between complete binary tree and binary tree.

b. Differentiate between greedy technique and dynamic

programming.

c. Solve the following recurrence using master method :

T(n) = 4T (n/3) + n2

d. Name the sorting algorithm that is most practically used

and also write its time complexity.

e. Find the time complexity of the recurrence relation

T(n) = n + T(n/10) + T(7n/5)

f. Explain single source shortest path.

g. Define graph colouring.

h. Compare time complexity with space complexity.

i. What are the characteristics of the algorithm ?

j. Differentiate between backtracking and branch and bound

techniques.

SECTION-B

2. Attempt any three of the following : (10 × 3 = 30)

a. Solve the following by recursion tree method
T(n) = n + T(n/5) + T(4n/5)
Solved Paper (2017-18) SP–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Insert the following information, F, S, Q, K, C, L, H, T, V, W,
M, R, N, P, A, B, X, Y, D, Z, E, G, I into an empty B-tree with
degree t = 3.

c. What is minimum cost spanning tree ? Explain Kruskal’s

algorithm and Find MST of the graph. Also write it’s time
complexity.

7 1 6
2
2 4
6 6
3 3
4 5
7
5 6
7
Fig. 2.

d. What is red-black tree ? Write an algorithm to insert a node

in an empty red-black tree explain with suitable example.

e. Explain HEAP SORT on the array. Illustrate the operation

HEAP SORT on the array
A = {6, 14, 3, 25, 2, 10, 20, 7, 6}

SECTION-C

3. Attempt any one part of the following : (10 × 1 = 10)

a. Explain convex-hull problem.

b. Find the shortest path in the below graph from the source
vertex 1 to all other vertices by using Dijkstra’s algorithm.
50
2 3
10
10
1 30 20
100
5 4
60
Fig. 2.

4. Attempt any one part of the following : (10 × 1 = 10)

a. What is backtracking ? Discuss sum of subset problem with
the help of an example.

b. Write down an algorithm to compute Longest Common

Subsequence (LCS) of two given strings and analyze its
time complexity.
Design and Analysis of Algorithms SP–3 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
5. Attempt any one part of the following : (10 × 1 = 10)
a. The recurrence T(n) = 7T(n/2) + n2 describe the running
time of an algorithm A. A competing algorithm A has a
running time T(n) = aT (n/4) + n 2. What is the largest
integer value for a A is asymptotically faster than A ?

b. Discuss the problem classes P, NP and NP-complete with

class relationship.

6. Attempt any one part of the following : (10 × 1 = 10)

a. Explain properties of binomial heap. Write an algorithm to
perform uniting two binomial heaps. And also to find
minimum key.

b. Given the six items in the table below and a knapsack with
weight 100, what is the solution to the knapsack problem in
all concepts. i.e., explain greedy all approaches and find the
optimal solution.
Item ID Weight Value Value/Weight
A 100 40 .4
B 50 35 .7
C 40 20 .5
D 20 4 .2
E 10 10 1
F 10 6 .6

7. Attempt any one part of the following : (10 × 1 = 10)

a. Compute the prefix function  for the pattern P = a b a c a b
using Knuth-Morris-Pratt algorithm. Also explain Naive
string matching algorithm.

b. Explain approximation and randomized algorithms.


Solved Paper (2017-18) SP–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
SOLUTION OF PAPER (2017-18)

Note : Attempt all sections. Assume any missing data.

Section-A

1. Define/Explain the following : (2 × 10 = 20)

a. Differentiate between complete binary tree and binary tree.
Ans.
S. No. Complete binary tree Binary tree

1. In a complete binary tree every A binary tree is a tree data

level, except possibly the last is structure in which each node has
completely filled, and all nodes at most two children, which are
in the last level are as far left referred to as the left child and
as possible. the right child.
2 It can have between 1 and 2h–1 It can have between 2h+1 and
nodes at the last level h. 2h+1 – 1 nodes at the last level h.

b. Differentiate between greedy technique and dynamic

programming.
Ans. Difference :
S. No. Greedy method Dynamic programming
1. A greedy algorithm is one Dynamic programming can be
that at a given point in time thought of as ‘smart’ recursion. It
makes a local optimization. often requires one to break down
a pro blem into smalle r
components that can be cached.
2. Greedy algorithms have a Dynamic programming would
lo cal choice of the solve all dependent subproblems
subproblem that will lead to and then select one that would lead
an optimal answer. to an optimal solution.

c. Solve the following recurrence using master method :

T(n) = 4T (n/3) + n2
Ans. T(n) =4T(n/3) + n2
a = 4, b = 3, f(n) = n2
nlogba = nlog34 = n1.261
f(n) = (nlogba + E)
Now, af(n/b)  c f(n)
Design and Analysis of Algorithms SP–5 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
4
f(n)  c f(n)
3
4 2
n  cn2
3
4
c=
3
Hence, T(n) =  (n2)

d. Name the sorting algorithm that is most practically used

and also write its time complexity.
Ans. Quick sort algorithm is most practically used in sorting.
Time complexity of quick sort is O(n log n).

e. Find the time complexity of the recurrence relation

T(n) = n + T(n/10) + T(7n/5)
Ans.

n n

n 7n 3n
10 5 2
log n
n 7n 7n 49n 9n
100 50 50 25 4

n 7n 7n 49n 7n 49n 49n 343n 27n

1000 500 500 250 500 250 250 125 8

30 n 31 n 32 33
T(n) = 0
 1  1 n  3 n + ................+ log n times
2 2 2 2
=  (n log n)

f. Explain single source shortest path.

Ans. Single source shortest path problem states that in a given graph
G = (V, E) we can find a shortest path from given source vertex
s  V to every vertex v  V.

g. Define graph colouring.

Ans. Graph colouring is a problem of colouring each vertex in graph in
such a way that no two adjacent vertices have same colour and
m-colours are used.

h. Compare time complexity with space complexity.

Solved Paper (2017-18) SP–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans.
S. No. Time complexity Space complexity
1. Time complexity is the Space complexity is the amount
amount of time required for of memory ne eded for an
an algorithm to complete its algorithm to solve the problem.
process.
2. It is expressed using Big It is expressed only using Big
Oh(O), theta () and Oh(O) notation.
omega () notation.

i. What are the characteristics of the algorithm ?

Ans. Characteristics of algorithm :
1. Input and output : The algorithm must accept zero or more
inputs and must produce at least one output.
2. Definiteness : Each step of algorithm must be clear and
unambiguous.
3. Effectiveness : Every step must be basic and essential.
4. Finiteness : Total number of steps used in algorithm should be
finite.

j. Differentiate between backtracking and branch and bound

techniques.
Ans.
S. No. Backtracking Branch and bound

1. Solution for backtracking is In this method, it is not necessary

trace d using de pth first to use depth first search fo r
search. obtaining the solution, even the
breadth first search, best first
search can be applied.
2. Typically decision problems Typically optimization problems can
can be solve d using be solved using branch and bound.
backtracking.

SECTION-B

2. Attempt any three of the following : (10 × 3 = 30)

a. Solve the following by recursion tree method
T(n) = n + T(n/5) + T(4n/5)
Ans. T(n) =n + n + n + ................+ log n times =  (n log n)
Design and Analysis of Algorithms SP–7 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
n n

n 4n n
5 5
log n
n 4n 4n 16n
25 25 25 25 n

n 4n 4n 16n 4n 16n 16n 64n n

125 125 125 125 125 125 125 125

b. Insert the following information, F, S, Q, K, C, L, H, T, V, W,

M, R, N, P, A, B, X, Y, D, Z, E, G, I into an empty B-tree with
degree t = 3.
Ans. Assume that t=3
2t – 1 = 2 × 3 – 1 = 6 – 1 = 5
and t–1= 3–1=2
So, maximum of 5 keys and minimum of 2 keys can be inserted in
a node. Now, apply insertion process as :
Insert F : F

Insert S : F S

Insert Q : F Q S

Insert K : F K Q S

Insert C : C F K Q S

Insert L : C F K L Q S
As, there are more than 5 keys in this node.
 Find median, n[x] = 6 (even)
n[ x ] 6
Median = = 3
2 2
Now, median = 3,
So, we split the node by 3rd key.

K
C F K L Q S
Median
(splitting point) C F L Q S
Move up
Solved Paper (2017-18) SP–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert H, T :
K

C F H L Q S T
Insert V :
K

C F H L Q S T V
Insert W :
K

C F H L Q S T V W
More than 5 keys split node from Median.
n[x] = 6 [even]
n[ x] 6
Median =  3 (i.e., 3rd key move up)
2 2
K S

C F H L Q T V W
Insert M, R, N :
K S

CF H L MNQR T V W
Insert P :

K S

C F H LM N P QR T V W
More than 5 key
split the node
n[x] = 6
n[ x] 6
Median =   3 (i.e., 3rd key move up)
2 2
Design and Analysis of Algorithms SP–9 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
K N S

C F H LM P QR T V W
Insert A, B :

KN S

A B CF H LM PQ R T V W
Insert X, Y :
K N S

A B CF H LM PQR TV WXY
Insert D :
K N S

A B CD FH LM PQ R TV WXY
More than 5 key
split the node
n[x] = 6 (even)
n[ x] 6
Median =  3 (i.e., 3rd key move up)
2 2

C K NS

AB D FH LM PQR TV WXY
Insert Z, E :

CK NS

A B DEFH LM PQR TVWXYZ

More than 5 key
split the node
n[x] = 6
n[ x] 6
Median =  3 (i.e., 3rd key move up)
2 2
Solved Paper (2017-18) SP–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
CK N SW

A B D E FH L M PQR TV X YZ
Insert G, I :
N

CK SW

A B DEFG HI LM PQR TV X YZ

CF K SW

A B DE GHI LM PQR TV X YZ
Fig. 1. Inserted all given information with degree t = 3.

c. What is minimum cost spanning tree ? Explain Kruskal’s

algorithm and Find MST of the graph. Also write it’s time
complexity.

7 1 6
2
2 4
6 6
3 3
4 5
7
5 6
7
Fig. 2.
Ans. Minimum spanning tree (Minimum cost spanning tree) :
1. Given a connected weighted graph G, it is often desired to create a
spanning tree T for G such that the sum of the weights of the tree
edges in T is as small as possible.
2. Such a tree is called a minimum spanning tree and represents the
‘‘cheapest’’ way of connecting all the nodes in G.
3. There are number of techniques for creating a minimum spanning
tree for a weighted graph but the most famous methods are Prim’s
and Kruskal’s algorithm.
Kruskal’s algorithm :
i. In this algorithm, we choose an edge of G which has smallest weight
among the edges of G which are not loops.
Design and Analysis of Algorithms SP–11 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
ii. This algorithm gives an acyclic subgraph T of G and the theorem
given below proves that T is minimal spanning tree of G. Following
steps are required :
Step 1 :Choose e1, an edge of G, such that weight of e1, w(e1) is as
small as possible and e1 is not a loop.
Step 2 :If edges e1, e2, ........., ei have been selected then choose an
edge ei+1 not already chosen such that
i. the induced subgraph
G[{e1 ..... ei+1}] is acyclic and
ii. w(ei+1) is as small as possible
Step 3 :If G has n vertices, stop after n – 1 edges have been chosen.
Otherwise repeat step 2.
If G be a weighted connected graph in which the weight of the
edges are all non-negative numbers, let T be a subgraph of G
obtained by Kruskal’s algorithm then, T is minimal spanning tree.
Numerical :
Step 1 : Arrange the edge of graph according to weight in ascending
order.
Edges Weight Edge Weight
13 2 32 6
46 3 17 7
25 4 35 7
36 5 56 7
34 6
41 6
Step 2 : Now draw the vertices as given in graph,

2 4
3

5 6
Now draw the edge according to the ascending order of weight. If
any edge forms cycle, leave that edge.
Step 3 : Select edge 13
1

2 2 4
3

5 6
Solved Paper (2017-18) SP–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Step 4 : Select edge 46
1

2 2 4
3
3

5 6
Step 5 : Select edge 25
1

2 2 4

4 3
3

5 6
Step 6 : Select edge 36
1

2 2 4

4 3
3
5
5 6
Step 7 : Select edge 23
1

2 2 4
6
4 3
3
5
5 6
All the remaining edges, such as 34, 41, 12, 35, 56 are rejected
because they form cycle.
All the vertices are covered in this tree. So, the final tree with
minimum cost of given graph is

2 2 4
6
4 3
3
5
5 6

Minimum cost = 2 + 3 + 4 + 5 + 6 = 20
Time complexity : Time complexity is O(|E| log |E|).
Design and Analysis of Algorithms SP–13 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
d. What is red-black tree ? Write an algorithm to insert a node
in an empty red-black tree explain with suitable example.
Ans. Red-black tree :
a. A red-black tree is a binary tree where each node has colour as an
extra attribute, either red or black.
b. It is a self-balancing Binary Search Tree (BST) where every node
follows following properties :
1. Every node is either red or black.
2. The root is black.
3. Every leaf (NIL) is black.
4. If a node is red, then both its children are black.
5. For each node, all paths from the node to descendent leave
contain the same number of black nodes.
Insertion algorithm :
i. We begin by adding the node as we do in a simple binary search
tree and colouring it red.
RB-INSERT(T, z)
1. y  nil [T]
2. x  root [T]
3. while x  nil [T]
4. do y  x
5. if key[z] < key [x]
6. then x  left [x]
7. else x  right [x]
8. p[z]  y
9. if y = nil [T]
10. then root [T]  z
11. else if key [z] < key[y]
12. then left [y] z
13. else right [y]  z
14. left [z]  nil[T]
15. right [z]  nil[T]
16. colour [z]  RED
17. RB-INSERT-FIXUP(T, z)
ii. Now, for any colour violation, RB-INSERT-FIXUP procedure is
used.
RB-INSERT-FIXUP(T, z)
1. while colour [p[z]] = RED
2. do if p[z] = left[p [p [z]]]
3. then y  right[p [p [z]]]
4. if colour[y] = RED
5. then colour[p [z]]  BLACK  case 1
6. colour[y]  BLACK  case 1
7. colour[p [p [z]]] RED  case 1
8. z  p [p [ z]]  case 1
9. else if z = right[p[z]]
10. then z  p[z]  case 2
Solved Paper (2017-18) SP–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
11. LEFT-ROTATE(T, z)  case 2
12. colour[p[z]]  BLACK  case 3
13. colour[p[p[z]]]  RED  case 3
14. RIGHT-ROTATE(T, p[p[z]])  case 3
15. else (same as then clause with “right” and “left”
exchanged)
16. colour[root[T]]  BLACK
For example :
B
Insert 41 : 41

B
41
Insert 38 : R
38
Insert 31 :
B B
41
R 38
Case 3 R R
38
R 31 41
31

Insert 12 :
B
38 B
B R 38
Case 1 R R
31 41
R 31 41
12 R
12

Insert 19 :
B B
38
B B 38
Case 2, 3 B B
31 41
R 19 41
12 R R
R 12 31
19
Insert 8 :
B B
38 38
B B R B
Case 1
19 41 19 41
R R B B
12 31 12 31
R R
8 8
Design and Analysis of Algorithms SP–15 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Thus final tree is

B
38
R B
19 41
B
B
12 31
R
8

e. Explain HEAP SORT on the array. Illustrate the operation

HEAP SORT on the array
A = {6, 14, 3, 25, 2, 10, 20, 7, 6}
Ans. Heap sort :
1. Heap sort is a comparison based sorting technique based on binary
heap data structure.
2. Heap sort finds the largest element and puts it at the end of array,
then the second largest item is found and this process is repeated
for all other elements.
3. The general approach of heap sort is as follows :
a. From the given array, build the initial max heap.
b. Interchange the root (maximum) element with the last element.
c. Use repetitive downward operation from root node to rebuild the
heap of size one less than the starting.
d. Repeat step a and b until there are no more elements.
Numerical :
Originally the given array is : [6, 14, 3, 25, 2, 10, 20, 7, 6]
First we call BUILD-MAX-HEAP
heap size [A] = 9
6
14 3

25 2 10 20

7 6
so, i = 4 to 1, call MAX-HEAPIFY (A, i)
i.e., first we call MAX-HEAPIFY (A, 4)
Az [l] = 7, A [i] = A [4] = 25, A [r] = 6
l  left [4] = 8
r  right [4] = 9
8  9 and 7 > 25 (False)
Then, largest  4
9  9 and 6 > 25 (False)
Then, largest = 4
A [i]  A [4]
Now call MAX-HEAPIFY (A, 2)
Solved Paper (2017-18) SP–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6
14 3
i=4
25 2 10 20
l=8
7 6 r=9
(i)
Similarly for i = 3, 2, 1 we get the following heap tree.
6 i 6
i
14 3 14 r 20
l
25 2 10 20 25 2 10 3
l r
7 6 7 6
(ii) (iii)
i
6 i 25
r
l 25 20 6 r 20
l
14 2 10 3 14 2 10 3

7 6 7 6
(iv) (v)

25 25
14 20 14 20
i
6 2 10 3 7 2 10 3

l 7 6 r 6 6
(vi) (vii)
So final tree after BUILD-MAX-HEAP is

25
14 20

7 2 10 3

6 6
(viii)

Now i = 9 down to 2, and size = size – 1 and call MAX-HEAPIFY

(A, 1) each time.
Design and Analysis of Algorithms SP–17 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
exchanging A [1]  A [9]

25 6
14 20 14 20

7 2 10 3 7 2 10 3

6 6 6 25

6 14 20 7 2 10 3 6 25
Now call MAX-HEAPIFY (A, 1) we get
Now exchange A [1] and A [8] and size = 8 – 1 = 7

20 6
14 10 14 10
7 2 6 3 7 2 3
6
6 20

6 14 10 7 2 6 3 20

Again call MAX-HEAPIFY (A, 1), we get

exchange A [1] and A [7] and size = 7 – 1 = 6

14 3
7 10 7 10

6 2 6 3 6 2 6 14

3 7 10 6 2 6 14

Again call MAX-HEAPIFY (A, 1), we get

exchange A [1] and A [6] and now size = 6 – 1 = 5

10 3
7 6 7 6

6 2 3 6 2 10

3 7 6 6 2 10

Again call MAX-HEAPIFY (A, 1)

exchange A [1] and A [5] and now size = 5 – 1 = 4
Solved Paper (2017-18) SP–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
7 2
6 6 6 6

3 2 3 7

2 6 6 3 7
Again, call MAX-HEAPIFY (A, 1)
exchange A [1] and A [4] and size = 4 – 1 = 3

6 3
6 2 6 2

3 6

3 6 2 6

call MAX-HEAPIFY (A, 1)

exchange A [1] and A [3], size = 3 – 1 = 2

6 2
3 2 3 6

2 3 6

call MAX-HEAPIFY (A, 1)

exchange A [1] and A [2] and size = 2 – 1 = 1

3 2
2 3
2 3

Thus, sorted array :

2 3 6 6 7 10 14 20 25

SECTION-C

3. Attempt any one part of the following : (10 × 1 = 10)

a. Explain convex-hull problem.
Ans.
1. The convex hull of a set S of points in the plane is defined as the
smallest convex polygon containing all the points of S.
2. The vertices of the convex hull of a set S of points form a (not
necessarily proper) subset of S.
Design and Analysis of Algorithms SP–19 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
3. To check whether a particular point p S is extreme, see each
possible triplet of points and check whether p lies in the triangle
formed by these three points.

Fig. 3.
4. If p lies in any triangle then it is not extreme, otherwise it is.
5. We denote the convex hull of S by CH(S). Convex hull is a convex
set because the intersection of convex sets is convex and convex
hull is also a convex closure.
Graham-Scan algorithm :
The procedure GRAHAM-SCAN takes as input a set Q of points,
where|Q|  3. It calls the functions Top(S), which return the point
on top of stack S without changing S, and to NEXT-TO-TOP(S),
which returns the point one entry below the top of stack S without
changing S.
GRAHAM-SCAN(Q)
1. Let p0 be the point in Q with the minimum y-coordinate, or the
leftmost such point in case of a tie.
2. Let <p1, p2, ...., pm> be the remaining points in Q, sorted by polar
angle in counter clockwise order around po (if more than one point
has the same angle remove all but the one that is farthest from po).
3. Top [S]  0
4. PUSH (p0, S)
5. PUSH (p1, S)
6. PUSH (p2, S)
7. for i  3 to m
8. do while the angle formed by points NEXT-To-TOP(S), Top(S), and
pi makes a non left turn.
9. do POP(S)
10. PUSH (pi, S)
11. return S

b. Find the shortest path in the below graph from the source
vertex 1 to all other vertices by using Dijkstra’s algorithm.
50
2 3
10
10
1 30 20
100
5 4
60
Fig. 4.
Solved Paper (2017-18) SP–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans. Initialize :
 50  S:{}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20

100
5 4
 60 
Extract min (1) :
 50  S : {1}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20

100
5 4
 60 
All edges leaving (1) :
10 
50 S : {1}
2 3
10 Q: 1 2 3 4 5
10 0    
0 1 30 20
10  30 100
100
5 4
60
100 30
Extract min(2) :
50 60 S : {1, 2}
2 3
10 Q:1 2 3 4 5
10 0    
1 30 20
10 60 30 100
100
5 4
60
All edges leaving (2) :
50 60 S : {1, 2}
2 3
10 Q:1 2 3 4 5
10 0    
1 30 20
10 60 30 100
100
5 4 60 30 100
60 30
100
Design and Analysis of Algorithms SP–21 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Extract min(4) :

10 60
50 S : {1, 2, 4}
2 3
10 Q:1 2 3 4 5
10 0    
0 1 30 20
10 60 30 100
100
5 4 60 30 100
60
100 30

All edges leaving (4) :

50 50 S : {1, 2, 4}
2 3 Q:1 2 3 4 5
10
0    
10
1 30 20 10 60 30 100
100 60 30 100
5 4 50
60
100 30

Extract min(3) :

10 S : {1, 2, 4, 3}
50
2 3 50 Q:1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50
100

All edges leaving (3) :

10 S : {1, 2, 4, 3, }
50
2 3 50 Q:1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50 60
100
Solved Paper (2017-18) SP–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Extract min(5) :
10 S : {1, 2, 4, 3, 5}
50
2 3 50 Q: 1 2 3 4 5
10
0    
10
0 1 30 20 10  30 100
100 60 30 100
5 4 30
60 50
60
60
Shortest path
2 3
10

1 30 10 20

5 4

4. Attempt any one part of the following : (10 × 1 = 10)

a. What is backtracking ? Discuss sum of subset problem with
the help of an example.
Ans. Backtracking :
1. Backtracking is a general algorithm for finding all solutions to
some computational problems.
2. Backtracking is an important tool for solving constraint satisfaction
problems, such as crosswords, verbal arithmetic, and many other
puzzles.
Sum of subset problem with example :
In the subset-sum problem we have to find a subset s of the given
set S = (S1, S2, S3, ......, Sn) where the elements of the set S are n
positive integers in such a manner that s  S and sum of the
elements of subset ‘s’ is equal to some positive integer ‘X’.
Algorithm for sum-subset problem :
Subset-Sum (S, t)
1. C  
2. Z  S
3. K  
4. t1  t
5. while (Z  ) do
6. K  max(Z)
7. if (K < t) then
8. ZZ–K
9. t1  t1 – K
10. CCK
11. else Z  Z – K
Design and Analysis of Algorithms SP–23 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
12. print C // Subset Sum elements whose
// Sum is equal to t1
This procedure selects those elements of S whose sum is equal to t.
Every time maximum element is found from S, if it is less than t
then this element is removed from Z and also it is subtracted from
t.
For example :
Given S = <1, 2, 5, 7, 8, 10, 15, 20, 25> & m = 35
Z  S, m = 35
k  max[Z] = 25
K<m
 Z=Z–K
i.e., Z = <1, 2, 5, 7, 8, 10, 15, 20> & m1  m
Subtracting K from m1, we get
New m1 = m1(old) – K = 35 – 25 = 10
In new step,
K  max[Z] = 20
K > m1
i.e., Z = <1, 2, 5, 7, 8, 10, 15>
In new step,
K  max[Z] = 15
K > m1
i.e., Z = <1, 2, 5, 7, 8, 10>
In new step,
K  max[Z] = 10
K > m1
i.e., Z = <1, 2, 5, 7, 8>
In new step,
K  max[Z] = 8
K > m1
i.e., Z = <1, 2, 5, 7> & m2  m1
New m2 = m2(old) – K = 10 – 8 = 2
In new step
K  max[Z] = 7
K > m2
i.e., Z = <1, 2, 5>
In new step, K  max[Z] = 5
K > m2
i.e., Z = <1, 2>
In new step, K  max[Z] = 2
K > m2
i.e., Z = <1>
In new step,
K=1
K < m2
 m3 = 01
Solved Paper (2017-18) SP–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Now only those numbers are needed to be selected whose sum is
01, therefore only 1 is selected from Z and rest other number found
as max[Z] are subtracted from Z one by one till Z become .

b. Write down an algorithm to compute Longest Common

Subsequence (LCS) of two given strings and analyze its
time complexity.
Ans. LCS-Length (X, Y) :
1. m  length[X]
2. n  length[Y]
3. for i  1 to m
4. do c[i, 0]  0
5. for j  0 to n
6. do c[0, j]  0
7. for i  1 to m
8. do for j  1 to n
9. do if xi = yj
10. then c[i, j]  c[i – 1, j – 1] + 1
11. b[i, j]  “”
12. else if c[i – 1, j]  c[i, j – 1]
13. then c[i, j]  c[i – 1, j]
14. b[i, j] “”
15. else c[i, j]  c[i, j – 1]
16. b[i, j] “”
17. return c and b
Note :
1. “” means both the same.
2. “” means c[i – 1, j]  c[i, j – 1].
3. “” means c[i – 1, j] < c[i, j – 1].
4. The “” diagonal arrows lengthen the LCS.
Time complexity :
Since, two for loops are present in LCS algorithm first for loop runs
upto m times and second for loop runs upto n times. So, time
complexity of LCS is O(mn).

5. Attempt any one part of the following : (10 × 1 = 10)

a. The recurrence T(n) = 7T(n/2) + n2 describe the running
time of an algorithm A. A competing algorithm A has a
running time T(n) = aT (n/4) + n 2. What is the largest
integer value for a A is asymptotically faster than A ?
Ans. Given that :
 n
T(n) = 7T    n2 ...(1)
 2
 n
T(n) = aT     n2 ...(2)
 4
Design and Analysis of Algorithms SP–25 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Here, eq. (1) defines the running time for algorithm A and eq. (2)
defines the running time for algorithm A. Then for finding value of
a for which A is asymptotically faster than A we find asymptotic
notation for the recurrence by using Master’s method.
 n
Now, compare eq. (1) by T(n) = aT    f (n)
 b
we get, a= 7
b=2
f(n) = n2
log 7
n log b a = n 2 = n2.81
Now, apply cases of Master’s, theorem as :
Case 1 : f(n) = O (n log2 7  E )
 f(n) = O (n2.81 – E)
 f(n) = O (n2.81 – 0.81)
 f(n) = O (n2)
Hence, case 1 of Master’s theorem is satisfied.
Thus, T(n) =  ( nlog b a )
 T(n) =  (n2.81)
Since recurrence given by eq. (1) is asymptotically bounded by
-notation by which is used to show optimum time we have to show
that recurrence given by eq. (2) is bounded by -notation which
shows minimum time (best case).
For the use satisfy the case 3 of Master theorem, let a = 16
 n
T(n) = 16T     n2
 4
 a = 16
b=4
f(n) = n2
(nlog b a E ) = (n2 + E)
Hence, case 3 of Master’s theorem is satisfied.
 T(n) =  (f(n))
 T(n) =  (n2)
Therefore, this shows that A is asymptotically faster than A when
a = 16.

b. Discuss the problem classes P, NP and NP-complete with

class relationship.
Ans.
1. The notion of NP-hardness plays an important role in the
relationship between the complexity classes P and NP.
Solved Paper (2017-18) SP–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
NP
NP-complete

NP-hard
P
Fig. 5. Relationship among P, NP, NP-complete and NP-hard problems.
2. It is also often used to define the complexity class NP-complete
which is the intersection of NP and NP-hard.
3. Consequently class NP-hard can be understood as the class of
problems that are NP-complete or harder.
4. There are no polynomial time algorithms for NP-hard problems.
5. A problem being in NP means that the problem is “easy” (in a
certain specific sense), whereas a problem being NP-hard means
that the problem is “difficult” (in another specific sense).
6. A problem being in NP and a problem being NP-hard are not
mutually exclusive. When a problem is both in NP and NP-hard, we
say that the problem is NP-complete.
7. All problems in NP can be solved deterministically in time O(2n).
8. An example of an NP-hard problem is the decision problem subset-
sum. Given a set of integers, does any non-empty subset of them
add up to zero ? i.e., a yes or no question, and happens to be NP-
complete.
9. There are, also decision problems that are NP-hard but not NP-
complete.
10. For example, in the halting problem “given a program and its input,
will it run forever” i.e., yes or no question, so this is a decision
problem. It is case to prove that the halting problem is NP-hard but
not NP-complete.

6. Attempt any one part of the following : (10 × 1 = 10)

a. Explain properties of binomial heap. Write an algorithm to
perform uniting two binomial heaps. And also to find
minimum key.
Ans. Properties of binomial heap :
1. The total number of nodes at order k are 2k.
2. The height of the tree is k.
 k
3. There are exactly   i.e., kCi nodes at depth i for i = 0, 1, …. , k
 i
(this is why the tree is called a “binomial” tree).
4. Root has degree k (children) and its children are Bk-1, Bk-2, …, B0
from left to right.
Algorithm for union of binomial heap :
1. The BINOMIAL-HEAP-UNION procedure repeatedly links binomial
trees where roots have the same degree.
Design and Analysis of Algorithms SP–27 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. The following procedure links the Bk-1 tree rooted at node to the
Bk-1 tree rooted at node z, that is, it makes z the parent of y. Node z
thus becomes the root of a Bk tree.
BINOMIAL-LINK (y, z)
i p[y]  z
ii. sibling [y]  child[z]
iii. child[z]  y
iv. degree[z]  degree[z] + 1
3. The BINOMIAL-HEAP-UNION procedure has two phases :
a. The first phase, performed by the call of BINOMIAL-HEAP-
MERGE, merges the root lists of binomial heaps H1 and H2 into a
single linked list H that is sorted by degree into monotonically
increasing order.
b. The second phase links root of equal degree until at most one root
remains of each degree. Because the linked list H is sorted by
degree, we can perform all the like operations quickly.
BINOMIAL-HEAP-UNION(H1, H2)
1. H  MAKE-BINOMIAL-HEAP ( )
2. head[H]  BINOMIAL-HEAP-MERGE(H1, H2)
3. Free the objects H1 and H2 but not the lists they point to
4. if head[H] = NIL
5. then return H
6. prev-x  NIL
7. x  head[H]
8. next-x  sibling[x]
9. while next-x  NIL
10. do if (degree[x]  degree[next-x]) or
(sibling[next-x]  NIL and degree[sibling[next-x]] = degree[x])
11. then prev-x  x  case 1 and 2
12. x  next-x  case 1 and 2
13. else if key[x]  key[next-x]
14. then sibling[x]  sibling[next-x]  case 3
15. BINOMIAL-LINK(next-x, x)  case 3
16. else if prev-x = NIL
 case 4
17. then head[H]  next-x
 case 4
18. else sibling[prev-x]  next-x  case 4
19. BINOMIAL-LINK(x, next-x)  case 4
20. x  next-x  case 4
21. next-x  sibling[x]
22. return H
BINOMIAL-HEAP-MERGE(H1, H2)
1. a  head[H1]
2. b  head[H2]
3. head[H1]  min-degree (a, b)
4. if head[H1] = NIL
Solved Paper (2017-18) SP–28 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
5. return
6. if head[H1] = b
7. then b  a
8. a  head[H1]
9. while b  NIL
10. do if sibling[a] = NIL
11. then sibling[a]  b
12. return
13. else if degree [sibling[a]] < degree[b]
14. then a  sibling[a]
15. else c  sibling[b]
16. sibling[b]  sibling[a]
17. sibling[a]  b
18. a  sibling[a]
19. bc
Minimum key :
BINOMIAL-HEAP-EXTRACT-MIN (H) :
1. Find the root x with the minimum key in the root list of H, and
remove x from the root list of H.
2. H'  MAKE-BINOMIAL-HEAP( ).
3. Reverse the order of the linked list of x’s children, and set head[H]
to point to the head of the resulting list.
4. H  BINOMIAL-HEAP-UNION(H, H’).
5. Return x
Since each of lines 1-4 takes O(log n) time of H has n nodes,
BINOMIAL- HEAP-EXTRACT-MIN runs in O(log n) time.

b. Given the six items in the table below and a knapsack with
weight 100, what is the solution to the knapsack problem in
all concepts. i.e., explain greedy all approaches and find the
optimal solution.
Item ID Weight Value Value/Weight
A 100 40 .4
B 50 35 .7
C 40 20 .5
D 20 4 .2
E 10 10 1
F 10 6 .6

Ans. We can use 0/1-knapsack problem when the items cannot be divided
into parts and fractional knapsack problem when the items can be
divided into fractions.
Design and Analysis of Algorithms SP–29 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
First arrange in non-increasing order of value/weight :
Item ID Weight Value Value/Weight
E 10 10 1
B 50 35 0.7
F 10 6 0.6
C 40 20 0.5
A 100 40 0.4
D 20 4 0.2
According to 0/1-knapsack problem, either we select an item or
reject. So the item will be selected according to value per weight.
E is selected W = 10 < 100
B is selected W = 10 + 50
= 60 < 100
F is selected W = 60 + 10
= 70 < 100
C cannot be selected because
W = 70 + 40 = 110 > 100
Hence we select D
W = 70 + 20 = 90 < 100
Total value = 10 + 35 + 6 + 4 = 55
According to fractional knapsack problem, we can select fraction of
any item.
E is selected W = 10 < 100
B is selected W = 10 + 50
= 60 < 100
F is selected W = 60 + 10
= 70 < 100
If we select C W = 70 + 40
= 110 > 100
Hence we select the fraction of item C as
100  W 100  70
=
Weight of C 40
Weight of C = 30/40 = 0.75
So, W= 0.75 × 40 = 30
W= 70 + 30 = 100
Total value = 10 + 35 + 6 + 0.75 (20)
= 10 + 35 + 6 + 15 = 66

7. Attempt any one part of the following : (10 × 1 = 10)

a. Compute the prefix function  for the pattern P = a b a c a b
using Knuth-Morris-Pratt algorithm. Also explain Naive
string matching algorithm.
Solved Paper (2017-18) SP–30 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans. Prefix function of the string abacab :
m  length [P]
 m=6
Initially, [1] = 0, k = 0
for q  2 to 6
for q = 2, k  0
& P[0 + 1]  P[2]
 [2] = 0
for q = 3, k  0
& P[0 + 1] = P[3]
 k= k+1=1
& [3] = 1
for q = 4, k > 0
& P[1 + 1]  P[4]
 k  [1] = 0
P[1]  P[4]
& [4] = 0
for q = 5, k > 0
& P[0 + 1] = P[5]
 k0+1= 1
& [5] = 1
for q = 6, k > 0
& P[1 + 1] = P[6]
 k1+1= 2
& [6] = 2

String a b a c a b
P[i] 1 2 3 4 5 6
[i] 0 0 1 0 1 2
Naive string matching algorithm :
1. The Naive approach simply test all the possible placement of pattern
P[1 .. m] relative to text T[1 .. n]. Specifically, we try shifts s = [0, 1,
…., n – m], successively and for each shift, s, compare T[s + 1 .. s +
m] to P[1 .. m].
2. The Naive string matching procedure can be interpreted graphically
as a sliding a pattern P[1 .. m] over the text T[1 .. m] and noting for
which shift all of the characters in the pattern match the
corresponding characters in the text.
Naive string matcher (T, P)
1. n  length [T]
2. m  length [P]
3. for s  0 to n–m do
4. j1
5. while j  m and T[s + j] = P[j] do
Design and Analysis of Algorithms SP–31 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6. j  j +1
7. if j > m then
8. return valid shift s
9. return no valid shift exist // i.e., there is no substring of T matching
P.

b. Explain approximation and randomized algorithms.

Ans. Approximation algorithm :
1. An approximation algorithm is a way of dealing with NP-
completeness for optimization problem. This technique does not
guarantee the best solution.
2. The best of an approximation algorithm is to come as close as possible
to the optimum value in a reasonable amount of time which is at
most polynomial time.
3. Let c(i) be the cost of solution produced by approximate algorithm
and c*(i) be the cost of optimal solution for some optimization
problem instance i.
4. For minimization and maximization problem, we are interested in
finding a solution of a given instance i in the set of feasible solutions,
such that c(i) / c*(i) and c*(i) / c(i) be as small as possible respectively.
5. We say that an approximation algorithm for the given problem
instance i, has a ratio bound of p(n) if for any input of size n, the
cost c of the solution produced by the approximation algorithm is
within a factor of p(n) of the cost c* of an optimal solution. That is
max(c(i) / c*(i), c*(i) / c(i))  p(n)
The definition applies for both minimization and maximization
problems.
6. p(n) is always greater than or equal to 1. If solution produced by
approximation algorithm is true optimal solution then clearly we
have p(n) = 1.
Randomized algorithm :
1. A randomized algorithm is defined as an algorithm that is allowed
to access a source of independent, unbiased bits and it is then
allowed to use these random bits to influence its computation.
2. An algorithm is randomized if its output is determined by the
input as well as the values produced by a random number
generator.
3. A randomized algorithm makes use of a randomizer such as a
random number generator.
4. The execution time of a randomized algorithm could also vary
from run to run for the same input.
5. The algorithm typically uses the random bits as an auxiliary input
to guide its behaviour in the hope of achieving good performance
in the “average case”.
6. Randomized algorithms are particularly useful when it faces a
malicious attacker who deliberately tries to feed a bad input to the
algorithm.

Design and Analysis of Algorithms SP–1 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B.Tech.
(SEM. V) ODD SEMESTER THEORY
EXAMINATION, 2018-19
DESIGN AND ANALYSIS OF ALGORITHM
Time : 3 Hours Max. Marks : 70

Note : 1. Attempt all sections. If require any missing data; then choose
suitable.
2. Any special paper specific instruction.

Section-A

1. Attempt all questions in brief : (2 × 10 = 20)

a. Rank the following by growth rate :
n, 2 log n , log n, log (log n), log2 n, (log n)log n, 4, (3/2)n, n!

b. Prove that if n > = 1, then for any n-key B-tree of height h

and minimum degree t > = 2, h < = logt ((n + 1)/2).

c. Define principle of optimality. When and how dynamic

programming is applicable ?

d. Explain application of graph colouring problem.

e. Compare adjacency matrix and linked adjacency lists

representation of graph with suitable example / diagram.

f. What are approximation algorithms ? What is meant by

p(n) approximation algorithms ?

g. What do you mean by stability of a sorting algorithm ?

Explain its application.

SECTION-B

2. Attempt any three of the following : (7 × 3 = 21)

a. Use a recursion tree to give an asymptotically tight solution
to the recurrence T(n) = T(n) + T((1 – )n) + cn, where  is
a constant in the range 0 <  <1 and c > 0 is also a constant.
Solved Paper (2018-19) SP–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Define BNP, NP hard and NP complete problems. Prove that
travelling salesman problem is NP complete.

c. Consider the weight and values of item listed below. Note

that there is only one unit of each item. The task is to pick
a subset of these items such that their total weight is no
more than 11 kgs and their total value is maximized.
Moreover, no item may be split. The total value of items
picked by an optimal algorithm is denoted by Vopt. A greedy
algorithm sorts the items by their value-to-weight rations
in descending order and packs them greedily, starting from
the first item in the ordered list. The total value of items
picked by the greedy algorithm is denoted by Vgreedy. Find
the value of Vopt – Vgreedy.
Item I1 I2 I3 I4
W 10 7 4 2
V 60 28 20 24

d. Insert the following key in a 2-3-4 B-tree :

40, 35, 22, 90, 12, 45, 58, 78, 67, 60 and then delete key 35 and 22
one after other.

e. Prove that the weights on the edge of the connected

undirected graph are distinct then there is a unique
minimum spanning tree. Give an example in this regard.
Also discuss prim’s minimum spanning tree algorithm in
detail.

SECTION-C

3. Attempt any one part of the following : (10 × 1 = 10)

a. The recurrence T(n) = 7T(n/3) + n2 describes the running
time of an algorithm A. Another competing algorithm B
has a running time of S(n) = a S(n/9) + n2. What is the
smallest value of a such that B is asymptotically faster
than A ?

b. How will you sort following array A of element using heap

sort : A = (23, 9, 18, 45, 5, 9, 1, 17, 6).

4. Attempt any one part of the following : (10 × 1 = 10)

a. Explain the different conditions of getting union of two
existing binomial heaps. Also write algorithm for union of
two binomial heaps. What is its complexity ?
Design and Analysis of Algorithms SP–3 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Insert the elements 8, 20, 11, 14, 9, 4, 12 in a Red-Black tree
and delete 12, 4, 9, 14 respectively.

5. Attempt any one part of the following : (10 × 1 = 10)

a. When do Dijkstra and the Bellman Ford algorithm both fail
to find a shortest path ? Can Bellman Ford detect all
negative weight cycles in a graph ? Apply Bellman Ford
algorithm on the following graph :
Source
vertex 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G
2 C 11 1
D E
Fig. 1.

b. Given an integer x and a positive number n, use divide and

conquer approach to write a function that computes xn with
time complexity O(log n).

6. Attempt any one part of the following : (10 × 1 = 10)

a. Solve the subset sum problem using backtracking, where
n = 4, m = 18, w[4] = {5, 10, 8, 13}.

b. Give Floyd-Warshall algorithm to find the shortest path

for all pairs of vertices in a graph. Give the complexity of
the algorithm. Explain with example.

7. Attempt any one part of the following : (10 × 1 = 10)

a. What is the application of Fast Fourier Transform (FFT) ?
Also write the recursive algorithm for FFT.

b. Give a linear time algorithm to determine if a text T is a

cycle rotation of another string T. For example : RAJA and
JARA are cyclic rotations of each other.


Solved Paper (2018-19) SP–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
SOLUTION OF PAPER (2018-19)

Note : 1. Attempt all sections. If require any missing data; then choose
suitable.
2. Any special paper specific instruction.

Section-A

1. Attempt all questions in brief : (2 × 10 = 20)

a. Rank the following by growth rate :
n, 2 log n , log n, log (log n), log2 n, (log n)log n, 4, (3/2)n, n!
Ans. Rank in increasing order of growth rate is given as :
n
 3
4, log n, log (log n), log2 n, (log n)log n, log n, 2 log n , n, n!,  
 2

b. Prove that if n > = 1, then for any n-key B-tree of height h

and minimum degree t > = 2, h < = logt ((n + 1)/2).
Ans. Proof :
1. The root contains at least one key.
2. All other nodes contain at least t – 1 keys.
3. There are at least 2 nodes at depth 1, at least 2t nodes at depth 2, at
least 2ti – 1 nodes at depth i and 2th – 1 nodes at depth h.
h
 th  1
n  1 + (t – 1)  2t
i1
i 1
= 1 + 2(t – 1) 
 t  1 
= 2th – 1

4. So th  (n + 1)/2 as required.
Taking log both sides we get,
h  logt (n + 1)/2

c. Define principle of optimality. When and how dynamic

programming is applicable ?
Ans. Principle of optimality : Principle of optimality states that in an
optimal sequence of decisions or choices, each subsequence must
also be optimal.
Dynamic programming is mainly applicable where the solution of
one sub-problem is needed repeatedly. In this procedure, the
solutions of sub-problems are stored in a table, so that there is no
need to re-compute the sub-problems and can be directly accessed
from the table if required.

d. Explain application of graph colouring problem.

Ans. Application of graph colouring problem :
Design and Analysis of Algorithms SP–5 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1. Sudoku : Sudoku is a variation of graph colouring problem where
every cell represents a vertex. There is an edge between two vertices
if they are in same row or same column or same block.
2. Register allocation : In compiler optimization, register allocation
is the process of assigning a large number of target program variables
onto a small number of CPU registers. This problem is also a graph
colouring problem.
3. Bipartite graphs : We can check if a graph is bipartite or not by
colouring the graph using two colours. If a given graph is
2-colourable, then it is bipartite, otherwise not.
4. Map colouring : Geographical maps of countries or states where
no two adjacent cities cannot be assigned same colour.

e. Compare adjacency matrix and linked adjacency lists

representation of graph with suitable example / diagram.
Ans.
S. No. Adjacency matrix Linked adjacency list

1. An adjacency matrix is a Linke d adjace ncy list is a

square matrix used to collection of unordered lists used
represent a finite graph. to represent a finite graph.
2. The elements of the matrix Each list describes the set of
indicate whether pairs of adjacent vertices in the graph.
vertices are adjacent or not in
the graph.
3. Space complexity in the worst Space complexity in the worst
case is O(|V|2). case is O(|V|+|E|).

For example : Consider the graph :

0 1
2
4 3
Using adjacency matrix : Using adjacency list :
0 1 2 3 4
0 1 4 /
0 0 1 0 0 1
1 0 / 2 /
1 1 0 1 0 0
2 0 1 0 1 0 2 1 3 /

3 0 0 1 0 1 3 4 / 2 /

4 1 0 0 1 0 4 3 / 0 /

f. What are approximation algorithms ? What is meant by

p(n) approximation algorithms ?
Solved Paper (2018-19) SP–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans. Approximation algorithm : An approximation algorithm is a
way of dealing with NP-completeness for optimization problem.
This technique does not guarantee the best solution.
p(n) approximation algorithm : A is a p(n) approximate
algorithm if and only if for every instance of size n, the algorithm
achieves an approximation ratio of p(n). It is applied to both
maximization (0 < C(i)  C*(i)) and minimization (0 < C* (i)  C(i))
problem because of the maximization factor and costs are positive.
p(n) is always greater than 1.

g. What do you mean by stability of a sorting algorithm ?

Explain its application.
Ans. Stability of a sorting algorithm : Let A be an array, and let < be
a strict weak ordering on the elements of A.
Sorting algorithm is stable if :
i < j and A[i]  A[j] i.e., A[i] comes before A[j].
Stability means that equivalent elements retain their relative
positions, after sorting.
Application : One application for stable sorting algorithms is sorting
a list using a primary and secondary key. For example, suppose we
wish to sort a hand of cards such that the suits are in the order
clubs, diamonds, hearts, spades and within each suit, the cards are
sorted by rank. This can be done by first sorting the cards by rank
(using any sort), and then doing a stable sort by suit.

SECTION-B

2. Attempt any three of the following : (7 × 3 = 21)

a. Use a recursion tree to give an asymptotically tight solution
to the recurrence T(n) = T(n) + T((1 – )n) + cn, where  is
a constant in the range 0 <  <1 and c > 0 is also a constant.
Ans. T(n) =T(n) + T((1 – )n) + cn
Recursion tree :
Cost
cn cn

c n (1 – )cn cn

c 2 n c(1 –  ) n c(1 –  ) n c(1 –  ) n cn

cn
• • • • • •
• • • • • •
• • • • • •
•
T(1) T(1) T(1) T(1) (k + 1) times
Assuming T(1) = 1
So c(1 – )k n = 1
k
1  1 
cn = k
 
(1  ) 1  
Design and Analysis of Algorithms SP–7 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
 1 
log (n) = k log 
 1   
k= log cn
 log 1 (cn)
 1 
log  
1 

1   
So, Total cost = cn + cn + .... (k + 1) times = cn(k + 1)
log 1 (cn)
= cn ×
1 


Time complexity = O n log 1
1  
(cn) = O n log 1
1
n

b. Define BNP, NP hard and NP complete problems. Prove that
travelling salesman problem is NP complete.
Ans. P (BNP) : Class P are the problems which can be solved in
polynomial time, which take time like O(n), O(n2), O(n3).
Example : Finding maximum element in an array or to check
whether a string is palindrome or not. So, there are many problems
which can be solved in polynomial time.
NP : Class NP are the problems which cannot be solved in
polynomial time like TSP (travelling salesman problem).
Example : Subset sum problem is best example of NP in which
given a set of numbers, does there exist a subset whose sum is
zero, but NP problems are checkable in polynomial time means
that given a solution of a problem, we can check that whether the
solution is correct or not in polynomial time.
NP-complete : The group of problems which are both in NP and
NP-hard are known as NP-complete problem.
Now suppose we have a NP-complete problem R and it is reducible
to Q then Q is at least as hard as R and since R is an NP-hard
problem, therefore Q will also be at least NP-hard, it may be NP-
complete also.
NP hard and NP complete :
1. We say that a decision problem Pi is NP-hard if every problem in NP
is polynomial time reducible to Pi.
2. In symbols,
Poly
Pi is NP-hard if, for every Pj  NP, Pj   Pi .
3. This does not require Pi to be in NP.
4. Highly informally, it means that Pi is ‘as hard as’ all the problem in
NP.
5. If Pi can be solved in polynomial time, then all problems in NP.
6. Existence of a polynomial time algorithm for an NP-hard problem
implies the existence of polynomial solution for every problem in
NP.
Solved Paper (2018-19) SP–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
NP-complete problem :
1. There are many problems for which no polynomial time algorithms
is known.
2. Some of these problems are travelling salesman problem, optimal
graph colouring, the Knapsack problem, Hamiltonian cycles, integer
programming, finding the longest simple path in a graph, and
satisfying a Boolean formula.
3. These problems belongs to an interesting class of problems called
the ‘‘NP-complete’’ problems, whose status is unknown.
4. The NP-complete problems are traceable i.e., require a super
polynomial time.
Proof :
Part 1 : TSP is in NP.
Proof :
1. Let a hint S be a sequence of vertices V = v1,..., vn.
2. We then check two things :
a. First we check that every edge traversed by adjacent vertices is an
edge in G, such that the sum of these edge weights is less than or
equal to k.
b. Secondly we check that every vertex in G is in V, which assures
that every node has been traversed.
3. We accept S if and only if S satisfies these two questions, otherwise
reject.
4. Both of these checks are clearly polynomial, thus our algorithm
forms a verifier with hint S, and TSP is consequently in NP.
Part 2 : TSP is NP-Hard.
Proof :
1. To show that TSP is NP-Hard, we must show that every problem y
in NP reduces to TSP in polynomial time.
2. To do this, consider the decision version of Hamiltonian Cycle (HC).
3. Take G = (V, E), set all edge weights equal to 1, and let k = |V|= n,
that is, k equals the number of nodes in G.
4. Any edge not originally in G then receives a weight of 2 (traditionally
TSP is on a complete graph, so we need to add in these extra edges).
5. Then pass this modified graph into TSP, asking if there exists a tour
on G with cost at most k. If the answer to TSP is YES, then HC is
YES. Likewise if TSP is NO, then HC is NO.
First direction : HC has a YES answer => TSP has a YES answer.
Proof :
1. If HC has a YES answer, then there exists a simple cycle C that
visits every node exactly once, thus C has n edges.
2. Since every edge has weight 1 in the corresponding TSP instance
for the edges that are in the HC graph, there is a Tour of weight n.
Since k = n, and given that there is a tour of weight n, it follows that
TSP has a YES answer.
Second direction : HC has a NO answer => TSP has a NO answer.
Design and Analysis of Algorithms SP–9 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Proof :
1. If HC has a NO answer, then there does not exist a simple cycle C
in G that visits every vertex exactly once. Now suppose TSP has a
YES answer.
2. Then there is a tour that visits every vertex once with weight at
most k.
3. Since the tour requires every node be traversed, there are n edges,
and since k = n, every edge traversed must have weight 1, implying
that these edges are in the HC graph. Then take this tour and
traverse the same edges in the HC instance. This forms a
Hamiltonian Cycle, a contradiction.
This concludes Part 2. Since we have shown that TSP is both in NP
and NP-Hard, we have that TSP is NP-Complete.

c. Consider the weight and values of item listed below. Note

that there is only one unit of each item. The task is to pick
a subset of these items such that their total weight is no
more than 11 kgs and their total value is maximized.
Moreover, no item may be split. The total value of items
picked by an optimal algorithm is denoted by Vopt. A greedy
algorithm sorts the items by their value-to-weight rations
in descending order and packs them greedily, starting from
the first item in the ordered list. The total value of items
picked by the greedy algorithm is denoted by Vgreedy. Find
the value of Vopt – Vgreedy.
Item I1 I2 I3 I4
W 10 7 4 2
V 60 28 20 24
Ans. For Vgreedy :
Item W V
I1 10 60
I2 7 28
I3 4 20
I4 2 24
Arrange the items by V/W ratio in descending order :
Item W V V/W
I4 2 24 12
I1 10 60 6
I3 4 20 5
I2 7 28 4
Solved Paper (2018-19) SP–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Total weight W = 11 kg
I4 is picked so W = 11 – 2 = 9 kg
I1 cannot picked 10 > 9
I3 is picked, W = 9 – 4 = 5 kg
I2 cannot be picked 7 > 5
I4 and I3 are picked so
Vgreedy = V(I4) + V(I3) = 24 + 20 = 44
For Vopt : For calculating Vopt we use 0/1 knapsack problem, so only
item 1 is picked. Hence, Vopt = 60
So, Vopt – Vgreedy = 60 – 44 = 16

d. Insert the following key in a 2-3-4 B-tree :

40, 35, 22, 90, 12, 45, 58, 78, 67, 60 and then delete key 35 and 22
one after other.
Ans. In 2-3-4 B-trees, non-leaf node can have minimum 2 keys and
maximum 4 keys so the order of tree is 5.
Insert 40, 35, 22, 90 :
22 35 40 90
Insert 12 :
35

12 22 40 90
Insert 45, 58 :
35

12 22 40 45 58 90
Insert 78 :
35 58

12 22 40 45 78 90
Insert 67, 60 :
35 58

12 22 40 45 60 67 78 90
Delete 35 :
40 58

12 22 45 60 67 78 90
Delete 22 :
40 58

12 25 60 67 78 90
Design and Analysis of Algorithms SP–11 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
e. Prove that the weights on the edge of the connected
undirected graph are distinct then there is a unique
minimum spanning tree. Give an example in this regard.
Also discuss prim’s minimum spanning tree algorithm in
detail.
Ans. Proof :
1. Let we have an algorithm that finds an MST (which we will call A)
based on the structure of the graph and the order of the edges
when ordered by weight.
2. Assume MST A is not unique.
3. There is another spanning tree with equal weight, say MST B.
4. Let e1 be an edge that is in A but not in B.
5. Then, B should include at least one edge e2 that is not in A.
6. Assume the weight of e1 is less than that of e2.
7. As B is a MST, {e1} B must contain a cycle.
8. Replace e2 with e1 in B yields the spanning tree {e1} B – {e2}
which has a smaller weight compared to B.
9. This contradicts that B is not a MST.
So, MST of undirected graph with distinct edge is unique.
Example :
1
1 7

2 2 4

4 3 3
8 5

5 6
6
Fig. 1.
Step 1 : Arrange the edge of graph according to weight in ascending
order.
Edges Weight Edge Weight
12 1 14 7
13 2 35 8
46 3
25 4
36 5
56 6
Step 2 : Now draw the vertices as given in graph,
Solved Paper (2018-19) SP–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1

2 4
3

5 6
Now draw the edge according to the ascending order of weight. If
any edge forms cycle, leave that edge.
Step 3 :

1 1

2 4
3

5 6
Step 4 :

1 1

2 2 4
3

5 6
Step 5 :
1 1

2 2 4
3
3

5 6
Step 6 :

1 1

2 2 4

4 3
3

5 6
Design and Analysis of Algorithms SP–13 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Step 7 :

1 1

2 2 4

4 3
3
5
5 6
All the remaining edges, such as : 14, 35, 56 are rejected because
they form cycle.
All the vertices are covered in this tree. So, the final tree with
minimum cost of given graph is

1 1

2 2 4

4 3
3
5
5 6

Prim’s algorithm :
First it chooses a vertex and then chooses an edge with smallest
weight incident on that vertex. The algorithm involves following
steps :
Step 1 : Choose any vertex V1 of G.
Step 2 : Choose an edge e1 =V1V2 of G such that V2  V1 and e1 has
smallest weight among the edge e of G incident with V1.
Step 3 :If edges e1, e2, ......., ei have been chosen involving end
points V1, V2, ............., Vi+1, choose an edge ei+1 = VjVk with Vj = {V1
....... Vi+1} and Vk  {V1 ...........Vi+1} such that ei+1 has smallest
weight among the edges of G with precisely one end in {V1 ...............
Vi+1}.
Step 4 :Stop after n – 1 edges have been chosen. Otherwise goto
step 3.

SECTION-C

3. Attempt any one part of the following : (10 × 1 = 10)

a. The recurrence T(n) = 7T(n/3) + n2 describes the running
time of an algorithm A. Another competing algorithm B
has a running time of S(n) = a S(n/9) + n2. What is the
smallest value of a such that B is asymptotically faster
than A ?
Ans. Given that :
 n
T(n) = 7T    n2 ...(1)
 3
Solved Paper (2018-19) SP–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
 n 2
S(n) = aS    n ...(2)
9
Here, eq. (1) defines the running time for algorithm A and eq. (2) defines
the running time for algorithm B. Then for finding value of a for which B
is asymptotically faster than A we find asymptotic notation for the
recurrence by using Master’s method.
 n
Now, compare eq. ) with T(n) = aT    f (n)
 b
we get, a = 7, b = 3
f(n) = n2
a 7
log
nlog b = n 3 = n2.81
Now, apply cases of Master’s, theorem as :
7
Case 3 : f(n) = O (nlog 3 
)
 f(n) = O (n1.77 + )
 f(n) = O (n1.77 + 0.23)
 f(n) = O (n2)
Hence, case 3 of Master’s theorem is satisfied.
Thus, T(n) =  f(n)
 T(n) =  (n2)
Since recurrence (1) is asymptotically bounded by -notation which
is used to show optimum time we have to show that recurrence
given by eq. (2) is bounded by -notation which shows minimum
time (best case).
For the use satisfy the case 2 of Master theorem, Guess a = 81
n 2
S(n) = f(n) = 81 S    n
9
 a = 81, b = 9
f(n) = nlog 9 81
f(n) =  (nlog b a ) = (n2)
Hence, case 2 of Master’s theorem is satisfied.
 T(n) =  ( nlog 9 81 log n)
 T(n) =  (n2 log n)
Therefore, this shows that B is asymptotically faster than A when
a = 81.

b. How will you sort following array A of element using heap

sort : A = (23, 9, 18, 45, 5, 9, 1, 17, 6).
Ans. Given array : 23 9 18 45 5 9 1 17 6
First we call Build-Max heap
Design and Analysis of Algorithms SP–15 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
heap size [A] = 9
23

9 18
i=4
45 5 9 1

l=8 r=9
17 6
so i = 4 to 1 call MAX HEAPIFY (A, i)
i.e., first we call MAX HEAPIFY (A, 4)
A[l] = 7, A[i] = A[4] = 45, A[r] = 6
l  left [4] = 2 × 4 = 8
r  right[4] = 2 × 4 + 1 = 9
8  9 and A[8] = 17 < 45 (False)
Then, largest  4.
Similarly for i = 3, 2, 1 we get the following heap tree :

i 23
23 9 18
i=3 l r
9 18
l=6 r=7 45 5 9 1
45 5 9 1
17 6
17 6

45
i
23 23 18
r
i
45 l 18
9 1
5 9
l
9 5 9 1 r
17 6
17 6

So, final tree after Build-Max heap is

45
23 18

17 5 9 1

9 6

Now i = 9 down to 2 and size = 10 – 1 = 9 and call MAX HEAPIFY

(A, 1) each time
Solved Paper (2018-19) SP–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Exchanging A[1]  A[9]

45 6
23 18 23 18

17 5 9 1 17 5 9 1

9 6 9 45

45

Now call MAX HEAPIFY (A, 1) and

Exchange A[1] and A[8], size = 9 – 1 = 8
23 6
17 18 17 18

9 5 9 1 9 5 9 1

6 23
23 45
Now, call MAX HEAPIFY (A, 1), exchange A[1] and A[4] and size =
5–1=4

9 9

5 6 5 6

1 1

9 9 17 18 23 45
Now, call MAX HEAPIFY (A, 1), exchange A[1] and A[3] and size =
4–1=3

9 9

1 6 1 6

6 9 9 17 18 23 45
Exchange A[1] and A[2] size 3 – 1 = 2
5 5
1 1 1

5 6 9 9 17 18 23 45
Design and Analysis of Algorithms SP–17 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
The sorted array
1 5 6 9 9 17 18 23 45
Now call MAX HEAPIFY (A, 1) and exchange A[1] and A[7]
18 1
9 17 9 17
6 5 9 1 6 5 9 18
Now call MAX HEAPIFY (A, 1) and size = 7 – 1 = 6 exchange A[1]
and A[6]
17 17
6 9 6 9

1 5 9 1 5 9
Exchange A[1] and A[5] and size = 6 – 1 = 5
9 5
6 9 6 9

1 5 1 9

4. Attempt any one part of the following : (10 × 1 = 10)

a. Explain the different conditions of getting union of two
existing binomial heaps. Also write algorithm for union of
two binomial heaps. What is its complexity ?
Ans. There are four cases/conditions that occur while
performing union on binomial heaps.
Case 1 : When degree[x]  degree[next-x] = degree [sibling[next-
x]], then pointers moves one position further down the root list.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a b c d

Bk BL Case1 Bk BL
(a) (b )
Fig. 2.
Case 2 : It occurs when x is the first of three roots of equal degree,
that is, degree[x] = degree[next-x] = degree[sibling[next-x]], then
again pointer move one position further down the list, and next
iteration executes either case 3 or case 4.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a b c d

Bk Bk Bk Case 2 Bk Bk Bk
(a) (b )
Fig. 3.
Solved Paper (2018-19) SP–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Case 3 : If degree[x] = degree[next-x]  degree [sibling[next-x]]
and key[x]  key[next-x], we remove next-x from the root list and
link it to x, creating Bk+1 tree.

sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a b d

Bk Bk BL Case 3
Bk BL
c
key [x]  key [next-x]
(a) Bk
Bk+1
(b )

Fig. 4.
Case 4 : degree[x] = degree[next-x]  degree[sibling[next-x] and
key[next-x]  key x, we remove x from the root list and link it to
next-x, again creating a Bk+1 tree.
sibling
prev-x x next-x [next-x] prev-x x next-x
a b c d a c d

Bk Bk BL Case 4
b Bk BL
key [x]  key [next-x]
Bk
(a) Bk+1
(b )
Fig. 5.
Algorithm for union of binomial heap :
1. The BINOMIAL-HEAP-UNION procedure repeatedly links binomial
trees where roots have the same degree.
2. The following procedure links the Bk-1 tree rooted at node to the
Bk-1 tree rooted at node z, that is, it makes z the parent of y. Node z
thus becomes the root of a Bk tree.
BINOMIAL-LINK (y, z)
i p[y]  z
ii. sibling [y]  child[z]
iii. child[z]  y
iv. degree[z]  degree[z] + 1
3. The BINOMIAL-HEAP-UNION procedure has two phases :
a. The first phase, performed by the call of BINOMIAL-HEAP-
MERGE, merges the root lists of binomial heaps H1 and H2
Design and Analysis of Algorithms SP–19 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
into a single linked list H that is sorted by degree into
monotonically increasing order.
b. The second phase links root of equal degree until at most one
root remains of each degree. Because the linked list H is sorted
by degree, we can perform all the like operations quickly.
BINOMIAL-HEAP-UNION(H1, H2)
1. H  MAKE-BINOMIAL-HEAP ( )
2. head[H]  BINOMIAL-HEAP-MERGE(H1, H2)
3. Free the objects H1 and H2 but not the lists they point to
4. if head[H] = NIL
5. then return H
6. prev-x  NIL
7. x  head[H]
8. next-x  sibling[x]
9. while next-x  NIL
10. do if (degree[x]  degree[next-x]) or
(sibling[next-x]  NIL and degree[sibling[next-x]] = degree[x])
11. then prev-x  x  case 1 and 2
12. x  next-x  case 1 and 2
13. else if key[x]  key[next-x]
14. then sibling[x]  sibling[next-x]  case 3
15. BINOMIAL-LINK(next-x, x)  case 3
16. else if prev-x = NIL
 case 4
17. then head[H]  next-x
 case 4
18. else sibling[prev-x]  next-x  case 4
19. BINOMIAL-LINK(x, next-x)  case 4
20. x  next-x  case 4
21. next-x  sibling[x]
22. return H
BINOMIAL-HEAP-MERGE(H1, H2)
1. a  head[H1]
2. b  head[H2]
3. head[H1]  min-degree (a, b)
4. if head[H1] = NIL
5. return
6. if head[H1] = b
7. then b  a
8. a  head[H1]
9. while b  NIL
10. do if sibling[a] = NIL
11. then sibling[a]  b
12. return
13. else if degree [sibling[a]] < degree[b]
14. then a  sibling[a]
15. else c  sibling[b]
Solved Paper (2018-19) SP–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
16. sibling[b]  sibling[a]
17. sibling[a]  b
18. a  sibling[a]
19. bc
Time complexity of union of two binomial heap is O(log n).

b. Insert the elements 8, 20, 11, 14, 9, 4, 12 in a Red-Black tree

and delete 12, 4, 9, 14 respectively.
Ans. Insert 8 :
B
8
Insert 20 :
B
8 R
20
Insert 11 : Since, parent of node 11 is red. Check the colour of
uncle of node 11. Since uncle of node 11 is nil than do rotation and
recolouring.

B B
8 8 Left B
R rotation 11
20 R R
R 11 R 8 20
11
R
20

Insert 14 : Uncle of node 14 is red. Recolour the parent of node 14

i.e., 20 and uncle of node 14 i.e., 8. No rotation is required.
B B
11 11
B R B B
8 20 8 20

R 14 R 14

Insert 9 : Parent of node 9 is black. So no rotation and no

recolouring.
B B
11 11
B B B B
8 20 8 20
R R
R 9 14 R 9 14

Insert 4 : Parent of node 4 is black. So no rotation and no

recolouring.
Design and Analysis of Algorithms SP–21 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B B
11 11
B B B B
8 20 8 20
R R
R 4 R 9 14 R 4 R 9 14

Insert 12 : Parent of node 12 is red. Check the colour of uncle of

node 12, which is nil. So do rotation and recolouring.

B B
11 11
B Right B
B B
8 20 rotation 8 14
R
R 4 R 9 14 R 4 R 9 12 R 20 R

12 R

Delete 12 : Node 12 is red and leaf node. So simply delete node 12.

B B
11 11
B B B B
8 14 8 14

R 4 R 9 12 R 20 R R 4 R 9 20 R

Delete 4 : Node 4 is red and leaf node. So simply delete node 4.

B B
11 11
B B B B
8 14 8 14

R 4 9 R 20 R R 9 20 R

Delete 9 : Node 9 is red and leaf node. So simply delete node 9.

B B
11 11
B B B B
8 14 Case 1 8 14

9 R 20 R 20 R
Delete 14 : Node 14 is internal node replace node 14 with node 20
and do not change the colour.
Solved Paper (2018-19) SP–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B B
11 11
B B B B
8 14 8 20
R
20

5. Attempt any one part of the following : (10 × 1 = 10)

a. When do Dijkstra and the Bellman Ford algorithm both fail
to find a shortest path ? Can Bellman Ford detect all
negative weight cycles in a graph ? Apply Bellman Ford
algorithm on the following graph :
Source
vertex 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G
2 C 11 1
D E

Ans. Dijkstra algorithm fails to find a shortest path when the graph
contains negative edges.
Bellman Ford algorithm fails to find a shortest path when the graph
contain negative weight cycle.
No, Bellman Ford cannot detect all negative weight cycle in a graph.
Numerical :
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 
2 C 11
 D E 1
7 
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 
1
2 C 11
 D E 1
7 
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 
2 C 11
D E 1
7 12
Design and Analysis of Algorithms SP–23 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
  
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 13
2 C 11
D E 1
7 12
 15 
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 13
2 C 11
D E 1
7 12
 15 18
0 2 2 3
B F H
A
1 1
5 10 2
4 9 3 G 13
2 C 11
D E 1
7

b. Given an integer x and a positive number n, use divide and

conquer approach to write a function that computes xn with
time complexity O(log n).
Ans. Function to calculate xn with time complexity O(log n) :
int power(int x, unsigned int y)
{
int temp;
if(y == 0)
return 1;
temp = power(x, y/2);
if (y%2 == 0)
return temp * temp;
else
return x * temp * temp;
}

6. Attempt any one part of the following : (10 × 1 = 10)

a. Solve the subset sum problem using backtracking, where
n = 4, m = 18, w[4] = {5, 10, 8, 13}.
Ans. n=4
m = 18
w{4} = {5, 10, 8, 13}
Sorted order : w{4} = {5, 8, 10, 13}
Now we construct state-space tree.
Solved Paper (2018-19) SP–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
I = Include
I1 0 E1 E = Exclude

5 0
I2 I2 E2
E2 8 E3 0
15
I3 ×
I3
5 8
23 18
E3 ×
× E3 I3 
15 5
× I4 E4
15
I4 E4 18 5
28 15  ×
× ×
First Subset S1 = {5, 13}
Similarly, S2 = {8, 10}

b. Give Floyd-Warshall algorithm to find the shortest path

for all pairs of vertices in a graph. Give the complexity of
the algorithm. Explain with example.
Ans. Floyd-Warshall algorithm :
1. Floyd-Warshall algorithm is a graph analysis algorithm for finding
shortest paths in a weighted, directed graph.
2. A single execution of the algorithm will find the shortest path
between all pairs of vertices.
3. The algorithm considers the “intermediate” vertices of a shortest
path, where an intermediate vertex of a simple path p = (v1, v2, ...,
vm) is any vertex of p other than v1 or vm, that is, any vertex in the
set {v2, v3, ..., vm–1}.
4. Let the vertices of G be V = {1, 2, ..., n}, and consider a subset {1, 2,
..., k} of vertices for some k.
5. For any pair of vertices i, j  V, consider all paths from i to j whose
intermediate vertices are all drawn from {1, 2, ..., k}, and let p be a
minimum-weight path from among them.
6. Let dij( k) be the weight of a shortest path from vertex i to vertex j
with all intermediate vertices in the set {1, 2, ..., k}.
A recursive definition is given by
 wij if k  0
dij( k) =  ( k –1) ( k –1) ( k –1)
min( dij , dik  dkj ) if k  1
Floyd-Warshall (W) :
1. n  rows [W]
2. D(0)  W
3. for k  1 to n
4. do for i  1 to n
Design and Analysis of Algorithms SP–25 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
5. do for j  1 to n
( k –1)
6. do dij( k)  min( dij , dik( k–1)  dkj( k –1) )
7. return D(n)
Time complexity of Floyd-Warshall algorithm is O(n3).
Example :
2
3 4
8
1 2 3
–4 1 –5
7
5 4
6
Fig. 6.
dij(k) = min[ dij(k–1) , dik
(k –1) ( k –1)
 dkj ]
 (ijk–1) if dij(k –1)  dik
( k –1) ( k 1)
 dkj
 (ijk) = 
( k1) ( k –1) ( k1) ( k1)
 kj if dij  dik  dkj

0 3 8   4  NIL 1 1 NIL 1 
 0  1 7 NIL NIL NIL 2 2 
  (0) 
D(0) =  4 0   ;    NIL 3 NIL NIL NIL 
   
 2   5 0  4 NIL 4 NIL NIL 
    6 0 NIL
 NIL NIL 5 NIL 
0 3 8   4 NIL 1 1 NIL 1 
 0  1 7 NIL NIL NIL 2 2 
  (1) 
D(1) =  4 0   ;   NIL 3 NIL NIL NIL 
   
 2 5  5 0  2 4 1 4 NIL 1 
    6 
0 NIL NIL NIL 5 NIL 

0 3 8 4  4 NIL 1 1 2 1 
 0  1 7 NIL NIL NIL 2 2 
  (2) 
D(2) =  4 0 5 11 ;   NIL 3 NIL 2 2 
   
 2 5  5 0  2 4 1 4 NIL 1 
    6 0 NIL NIL NIL 5 NIL 

0 3 8 4  4  NIL 1 1 2 1 
 0  1 7  NIL NIL NIL 2 2 
  (3) 
D(3) =  4 0 5 11 ;    NIL 3 NIL 2 2 
   
 2 1  5 0  2 4 3 4 NIL 1 
    6 0  NIL NIL NIL 5 NIL 

Solved Paper (2018-19) SP–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
0 3  1 4  4  NIL 1 4 2 1 
3 0  4 1  1 4 NIL 4 2 1 
  (4) 
D(4) = 7 4 0 5 3 ;    4 3 NIL 2 1 
   
 2  1  5 0  2 4 3 4 NIL 1 
 8 5 1 6 
0 4 3 4 5 NIL 

0 1  3 2  4  NIL 3 4 5 1 
3 0  4 1 1  4 NIL 4 2 1 
  (5) 
D(5) = 7 4 0 5 3 ;    4 3 NIL 2 1 
   
 2  1  5 0  2 4 3 4 NIL 1 
 8 5 1 6 0 4 3 4 5 NIL 


7. Attempt any one part of the following : (10 × 1 = 10)

a. What is the application of Fast Fourier Transform (FFT) ?
Also write the recursive algorithm for FFT.
Ans. Application of Fast Fourier Transform :
1. Signal processing.
2. Image processing.
3. Fast multiplication of large integers.
4. Solving Poisson’s equation nearly optimally.
Recursive algorithm :
1. The Fast Fourier Transform (FFT) is a algorithm that computes a
Discrete Fourier Transform (DFT) of n-length vector in O(n log n)
time.
2. In the FFT algorithm, we apply the divide and conquer approach
to polynomial evaluation by observing that if n is even, we can
divide a degree (n – 1) polynomial.
A(x) =a0 + a1x + a2x2 +... + an–1 xn–1
n 
into two degree  – 1 polynomials.
2 
A[0](x) =a0 + a2x + a4x2 + ... + an–2 xn/2–1
A[1](x) =a1 + a3x + a5x2 + ... + an–1 xn/2–1
Where A[0] contains all the even index coefficients of A and A[1]
contains all the odd index coefficients and we can combine these
two polynomials into A, using the equation,
A(x) =A[0] (x2) + xA [1] (x2) ...(1)
FFT (a, w) :
1. n  length [a] n is a power of 2.
2. if n = 1
3. then return a
4. n e2i/n
5. x  0 x will store powers of  initially x = 1.
Design and Analysis of Algorithms SP–27 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
6. a[0]  (a0, a2,...an–2)
7. a[1]  (a1, a3, ... an–1)
8. y[0]  FFT(a[0], 2) Recursive calls with 2 as (n/2)th root of unity.
9. y[1]  FFT(a[0], 2)
10. for k  0 to (n/2) – 1
11. do yk  y[0] [1]
k + x yk
12. yk+(n/2)  y[0] – xy[1]
k k
13. x  xn
14. return y
Line 2-3 represents the basis of recursion; the DFT of one element
is the element itself. Since in this case
y0 = a0 10 = a0 1 = a0
Line 6-7 defines the recursive coefficient vectors for the polynomials
A[0] and A[1].  = kn
Line 8-9 perform the recursive DFTn/2 computations setting for
n
k = 0, 1, 2, ..., –1 i.e.,
2
y[0] [0]
k = A ( n ) ,
2k
y[1] [1] 2k
k = A ( n )
Lines 11-12 combine the results o f the recursive DFT n/2
calculations.
For y0, y2, ... y(n/2)–1, line 11 yields.
yk = y[0] k [1]
k   n yk

= A[0](2nk )   kn A[1]( 2nk ) = A( kn ) using equation (1)

For yn/2, y(n/2)+1 ... yn–1, line 12 yields.
yk+(n/2) = y[0] k [1] [0] k ( n /2) [1]
k –  n yk = yk   n yk [  kn ( n / 2)  –  kn ]
= A[0]( 2nk )   kn (n / 2) A[1]( 2nk )
= A[0]( 2nk  nn )   nk  (n /2) A[1]( 2nk  nn ) [  nn  1]
= A[0]( 2nk  n )   kn ( n/ 2) A[1] ( 2nk n )

= A( nk (n / 2) ) using equation (1)

each k = 0, 1, 2, ...,(n/2) – 1.
Thus, the vector y returned by the FFT algorithm will store the
values of A(x) at each of the roots of unity.

b. Give a linear time algorithm to determine if a text T is a

cycle rotation of another string T. For example : RAJA
and JARA are cyclic rotations of each other.
Ans. Knuth-Morris-Pratt algorithm is used to determine if a text T is a
cycle rotation of another string T.
Knuth-Morris-Pratt algorithm for string matching :
Solved Paper (2018-19) SP–28 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
COMPUTE-PREFIX-FUNCTION (P)
1. m  length [P]
2. [1]  0
3. k0
4. for q  2 to m
5. do while k > 0 and P [k + 1]  P[q]
6. do k  [k]
7. if P[k + 1] = P [q]
8. then k  k + 1
9. [q]  k
10. return 
KMP-MATCHER calls the auxiliary procedure COMPUTE-PREFIX-
FUNCTION to compute .
KMP-MATCHER (T, p)
1. n  length [T]
2. m  length [P]
3.  COMPUTE-PREFIX-FUNCTION (P)
4. q0
5. for i  1 to n
6. do while q > 0 and P [q + 1]  T[i]
7. do q  [q]
8. if P [q +1] = T[i]
9. then q  q + 1
10. if q = m
11. then print “pattern occurs with shift” i – m
12. q  [q]


Design and Analysis of Algorithms SP–1 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B. Tech.
(SEM. V) ODD SEMESTER THEORY
EXAMINATION, 2019-20
DESIGN AND ANALYSIS OF ALGORITHMS

Time : 3 Hours Max. Marks : 100

Note : Attempt all sections. If require any missing data; then choose
suitably.
Section-A

1. Attempt all questions in brief. (2 × 7 = 14)

a. How do you compare the performance of various
algorithms ?

b. Take the following list of functions and arrange them in

ascending order of growth rate. That is, if function g(n)
immediately follows function f(n) in your list, then it should
be the case that f(n) is O(g(n)).
f1(n) = n2.5, f2(n) = 2n, f3(n) = n + 10, f4(n) = 10n, f5(n) = 100n,
and f6(n) = n2 log n

c. What is advantage of binary search over linear search?

Also, state limitations of binary search.

d. What are greedy algorithms ? Explain their characteristics ?

e. Explain applications of FFT.

f. Define feasible and optimal solution.

g. What do you mean by polynomial time reduction ?

Section-B

2. Attempt any three of the following : (7 × 3 = 21)

a. i. Solve the recurrence T (n) = 2T(n/2) + n2 + 2n + 1
ii. Prove that worst case running time of any comparison sort
is  (n log n).

b. Insert the following element in an initially empty RB-Tree.

12, 9, 81, 76, 23, 43, 65, 88, 76, 32, 54. Now delete 23 and 81.
Solved Paper (2019-20) SP–2 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
c. Define spanning tree. Write Kruskal’s algorithm or finding
minimum cost spanning tree. Describe how Kruskal’s
algorithm is different from Prim’s algorithm for finding
minimum cost spanning tree.

d. What is dynamic programming ? How is this approach

different from recursion? Explain with example.

e. Define NP-hard and NP-complete problems. What are the

steps involved in proving a problem NP-complete ? Specify
the problems already proved to be NP-complete.

Section-C

3. Attempt any one part of the following : (7 × 1 = 7)

a. Among Merge sort, Insertion sort and quick sort which
sorting technique is the best in worst case. Apply the best
one among these algorithms to sort the list E, X, A, M, P, L,
E in alphabetic order.

b. Solve the recurrence using recursion tree method :

T (n) = T (n/2) + T (n/4) + T (n/8) + n

4. Attempt any one part of the following : (7 × 1 = 7)

a. Using minimum degree ‘t’ as 3, insert following sequence of
integers 10, 25, 20, 35, 30, 55, 40, 45, 50, 55, 60, 75, 70, 65, 80, 85
and 90 in an initially empty B-Tree. Give the number of
nodes splitting operations that take place.

b. Explain the algorithm to delete a given element in a binomial

heap. Give an example for the same.

5. Attempt any one part of the following : (7 × 1 = 7)

a. Compare the various programming paradigms such as
divide-and-conquer, dynamic programming and greedy
approach.

b. What do you mean by convex hull ? Describe an algorithm

that solves the convex hull problem. Find the time
complexity of the algorithm.

6. Attempt any one part of the following :

a. Solve the following 0/1 knapsack problem using dynamic
programming P = {11, 21, 31, 33} w = {2, 11, 22, 15} c = 40,
n = 4.
Design and Analysis of Algorithms SP–3 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Define Floyd Warshall algorithm for all pair shortest path
and apply the same on following graph :
3
1 2
4
6 3 5
1
1 2
3 4
2
Fig. 1.

7. Attempt any one part of the following : (7 × 1 = 7)

a. Describe in detail Knuth-Morris-Pratt string matching
algorithm. Compute the prefix function  for the pattern
ababbabbabbababbabb when the alphabet is  = {a, b}.

b. What is an approximation algorithm ? What is meant by

p(n) approximation algorithms ? Discuss approximation
algorithm for Travelling Salesman Problem.


Solved Paper (2019-20) SP–4 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
SOLUTION OF PAPER (2019-20)

Note : Attempt all sections. If require any missing data; then choose
suitably.
Section-A

1. Attempt all questions in brief. (2 × 7 = 14)

a. How do you compare the performance of various
algorithms ?
Ans. To compare the performance of various algorithms first we measure
its performance which depends on the time taken and the size of
the problem. For this we measure the time and space complexity of
various algorithms which is divided into different cases such as
worst case, average case and best case.

b. Take the following list of functions and arrange them in

ascending order of growth rate. That is, if function g(n)
immediately follows function f(n) in your list, then it should
be the case that f(n) is O(g(n)).
f1(n) = n2.5, f2(n) = 2n, f3(n) = n + 10, f4(n) = 10n, f5(n) = 100n,
and f6(n) = n2 log n
Ans. f3(n) = f4(n) = f5(n) < f2(n) < f6(n) < f1(n)

c. What is advantage of binary search over linear search?

Also, state limitations of binary search.
Ans. Advantages of binary search over linear search :
1. Input data needs to be sorted in binary search but not in linear
search.
2. Linear search does the sequential access whereas binary search
access data randomly.
3. Time complexity of linear search is O(n) where binary search has
time complexity O(log n).
Limitation of binary search :
1. List must be sorted.
2. It is more complicated to implement and test.

d. What are greedy algorithms ? Explain their characteristics ?

Ans. Greedy algorithms : Greedy algorithms are shortsighted in their
approach in the sense that they take decisions on the basis of
information at hand without worrying about the effect these
decisions may have in the future.
Characteristics of greedy algorithm :
1. Greedy algorithms are most efficient.
2. For every instance of input greedy algorithms makes a decision
and continues to process further set of input.
Design and Analysis of Algorithms SP–5 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
3. The other input values at the instance of decision are not used in
further processing.

e. Explain applications of FFT.

Ans. Application of Fast Fourier Transform :
1. Signal processing.
2. Image processing.
3. Fast multiplication of large integers.
4. Solving Poisson’s equation nearly optimally.

f. Define feasible and optimal solution.

Ans. Feasible solution : A feasible solution is a set of values for the
decision variables that satisfies all of the constraints in an
optimization problem. The set of all feasible solutions defines the
feasible region of the problem.
Optimal solution : An optimal solution is a feasible solution where
the objective function reaches its maximum (or minimum) value.

g. What do you mean by polynomial time reduction ?

Ans. A polynomial time reduction is a method for solving one problem
using another. For example, if a hypothetical subroutine solving
the second problem exists, then the first problem can be solved by
transforming or reducing it to inputs for the second problem and
calling the subroutine one or more times.

Section-B

2. Attempt any three of the following : (7 × 3 = 21)

a. i. Solve the recurrence T (n) = 2T(n/2) + n2 + 2n + 1
ii. Prove that worst case running time of any comparison sort
is  (n log n).
Ans.
 n
i. T (n) = 2T(n/2) + n2 + 2n + 1  2T    n2
 2
 n
Compare it with T(n) = aT    f (n)
 b
we have, a = 2, b = 2, f(n) = n2
Now, we apply cases for Master’s theorem.
nlog b a = nlog 2 2 = n
This satisfies case 3 of Master’s theorem.

 f(n) =   nlog b a  E     n1  E 
=  (n1 + 1) where E = 1
=  (n2)
Solved Paper (2019-20) SP–6 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
 n2 
Again 2 f    c f(n2) ...(1)
 2
eq. (1) is true for c = 2
 T(n) =  (f(n))
 T(n) =  f(n2)
ii. Let T(n) be the time taken by merge sort to sort any array of n
elements.
  n    n 
Therefore, T(n) = T         g(n)
  2    2 
where g(n)  (n)
This recurrence, which becomes :
 n
T(n) = 2T    g (n)
 2
when n is even is a special case of our general analysis for divide-
and conquer algorithms.
 n
Compare the above given recurrence with T(n) = aT    f (n)
 b
we get a= 2
b=2
f(n) = g(n)
log a log 2
Now we find, n b = n 2  n1  n
 f(n) = (n)
i.e., case 2 of Master’s theorem applied then

T(n) =  nlog b a log n 
 T(n) =  (n log n)
Hence, the worst case running time of merge sort is (n log n).

b. Insert the following element in an initially empty RB-Tree.

12, 9, 81, 76, 23, 43, 65, 88, 76, 32, 54. Now delete 23 and 81.
Ans.
B
Insert 12 : 12

B
Insert 9 : 12

R
9
B
Insert 81 : 12
R
R
9 81
Design and Analysis of Algorithms SP–7 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
B B
12 12
Insert 76 : R B
R Case 1 B
9 81 9 81
R
76 76 R
Insert 23 : B B
12 12
B B
B Case 3 B
9 81 9 76
R
76 23 81
R R R
23

Insert 43 : B B
12 12
B B
B Case 1 R
9 76 9 76
R R B B
23 81 23 81

R 43 43 R

Insert 65 : B B
12 12
B B
R Case 3 R
9 76 9 76
R B B
23 81 B 43 81
R
R
43 23 65 R

R
65

Insert 88 : B
12
B
R
9 76
B
43 81 B
R
R R
23 65 88
Solved Paper (2019-20) SP–8 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 76 : B
12
B
R
9 76 Case 2
B B
43 81
R
R R
23 65 76 R 88

76 B

R
12 81 B

B 9 B 43 76 R 88 R

R 23 R 65

Insert 32 : B
76
R B
12 81
B
B R
9 43 76 R 88

R
23 R 65

R
32

76 B

R
12 81 B

B 9 R 43 76 R 88 R

B 23 B 65

32 R
Design and Analysis of Algorithms SP–9 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Insert 54 : B
76

R 12 81 B

R R
B 9 R 43 76 88

B 23 65 B

R R
32 54

Delete 23 : B B
76 76

R 12 81 B R 12 81 B

R R R R
B 9 R 43 76 88 B 9 R 43 76 88

B 23 65 B B 32 65 B

R R R
32 54 54

Delete 81 : B B
76 76

R 12 81 B R 12 88 B

R R R
B 9 R 43 76 88 B 9 R 43 76

B 32 65 B B 32 65 B

R R
54 54

c. Define spanning tree. Write Kruskal’s algorithm or finding

minimum cost spanning tree. Describe how Kruskal’s
algorithm is different from Prim’s algorithm for finding
minimum cost spanning tree.
Ans. Spanning tree :
1. A spanning tree of a graph is a subgraph that contains all the
vertices and is a tree.
Solved Paper (2019-20) SP–10 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. A spanning tree of a connected graph G contains all the vertices
and has the edges which connect all the vertices. So, the number of
edges will be 1 less the number of nodes.
3. If graph is not connected, i.e., a graph with n vertices has edges less
than n – 1 then no spanning tree is possible.
4. A graph may have many spanning trees.
Kruskal’s algorithm :
i. In this algorithm, we choose an edge of G which has smallest weight
among the edges of G which are not loops.
ii. This algorithm gives an acyclic subgraph T of G and the theorem
given below proves that T is minimal spanning tree of G. Following
steps are required :
Step 1 :Choose e1, an edge of G, such that weight of e1, w(e1) is as
small as possible and e1 is not a loop.
Step 2 :If edges e1, e2, ........., ei have been selected then choose an
edge ei+1 not already chosen such that
i. the induced subgraph

G[{e1 ..... ei+1}] is acyclic and

ii. w(ei+1) is as small as possible
Step 3 :If G has n vertices, stop after n – 1 edges have been chosen.
Otherwise repeat step 2.
If G be a weighted connected graph in which the weight of the
edges are all non-negative numbers, let T be a subgraph of G
obtained by Kruskal’s algorithm then, T is minimal spanning tree.
Difference :
S. No. Kruskal’s algorithm Prim’s algorithm
1. Kruskal’s algo rithm Prim’s algorithm initializes with a
initiates with an edge. node.
2. Kruskal’s algorithm selects Prim’s algorithms span from one
the edges in a way that the node to another.
position of the edge is not
based on the last step.
3. Kruskal’s can be used on In Prim’s algorithm, graph must be
disconnected graphs. a connected graph.
4. Kruskal’s time complexity Prim’s algo rithm has a time
in worst case is O(E log E). co mple xity in wo rst case o f
O(E log V).

d. What is dynamic programming ? How is this approach

different from recursion? Explain with example.
Ans.
Design and Analysis of Algorithms SP–11 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1. Dynamic programming is a stage-wise search method suitable for
optimization problems whose solutions may be viewed as the result
of a sequence of decisions.
2. It is used when the sub-problems are not independent.
3. Dynamic programming takes advantage of the duplication and
arranges to solve each sub-problem only once, saving the solution
(in table or something) for later use.
4. Dynamic programming can be thought of as being the reverse of
recursion. Recursion is a top-down mechanism i.e., we take a
problem, split it up, and solve the smaller problems that are created.
Dynamic programming is a bottom-up mechanism i.e., we solve all
possible small problems and then combine them to obtain solutions
for bigger problems.
Difference :
1. In recursion, sub-problems are solved multiple times but in dynamic
programming sub-problems are solved only one time.
2. Recursion is slower than dynamic programming.
For example :
Consider the example of calculating nth Fibonacci number.
fibo(n) = fibo(n – 1) + fibo(n – 2)
fibo(n – 1) = fibo(n – 2) + fibo(n – 3)
fibo(n – 2) = fibo(n – 3) + fibo(n – 4)
.................................
................................
................................
fibo(2) = fibo(1) + fibo(0)
In the first three steps, it can be clearly seen that fibo(n – 3) is
calculated twice. If we use recursion, we calculate the same sub-
problems again and again but with dynamic programming we
calculate the sub-problems only once.

e. Define NP-hard and NP-complete problems. What are the

steps involved in proving a problem NP-complete ? Specify
the problems already proved to be NP-complete.
Ans. NP-hard problem :
1. We say that a decision problem Pi is NP-hard if every problem in NP
is polynomial time reducible to Pi.
2. In symbols,
Poly
Pi is NP-hard if, for every Pj  NP, Pj   Pi .
3. This does not require Pi to be in NP.
4. Highly informally, it means that Pi is ‘as hard as’ all the problem in
NP.
5. If Pi can be solved in polynomial time, then all problems in NP.
6. Existence of a polynomial time algorithm for an NP-hard problem
implies the existence of polynomial solution for every problem in
NP.
Solved Paper (2019-20) SP–12 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
NP-complete problem :
1. There are many problems for which no polynomial time algorithms
is known.
2. Some of these problems are travelling salesman problem, optimal
graph colouring, the Knapsack problem, Hamiltonian cycles, integer
programming, finding the longest simple path in a graph, and
satisfying a Boolean formula.
3. These problems belongs to an interesting class of problems called
the ‘‘NP-complete’’ problems, whose status is unknown.
4. The NP-complete problems are traceable i.e., require a super
polynomial time.
Steps involved in proving a problem is NP-complete :
Step 1 : Given a problem U, show that U is NP.
Step 2 : Select a known NP-complete problem V.
Step 3 : Construct a reduction from V to U.
Step 4 : Show that the reduction requires polynomial time.
List of NP-complete problems :
1. 3SAT
2. Vertex-cover
3. Clique
4. Hamiltonian circuits.

Section-C

3. Attempt any one part of the following : (7 × 1 = 7)

a. Among Merge sort, Insertion sort and quick sort which
sorting technique is the best in worst case. Apply the best
one among these algorithms to sort the list E, X, A, M, P, L,
E in alphabetic order.
Ans. Merge sort technique is best in worst case because of its time
complexity O(n log n).
Numerical :
Given : E, X, A, M, P, L, E
Pass 1 : Merge each pair of element to obtain sorted list :
E X A M P L E
After sorting each pair, we get
E X A M L P E
Pass 2 : Merge each pair to obtain the list :
A E M X E L P
Pass 3 : Again merge the two sub arrays to obtain the list :
A E E L M P X
b. Solve the recurrence using recursion tree method :
T (n) = T (n/2) + T (n/4) + T (n/8) + n
Ans.
Design and Analysis of Algorithms SP–13 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
7n 72 n
T(n) = n   2 + ........+ log n times
8 8
=  (n log n)
n n

n n n 7n
2 4 8 8
log n

2
n n n n n n n n n 7n
2
4 8 16 8 16 32 16 32 64 8

4. Attempt any one part of the following : (7 × 1 = 7)

a. Using minimum degree ‘t’ as 3, insert following sequence of
integers 10, 25, 20, 35, 30, 55, 40, 45, 50, 55, 60, 75, 70, 65, 80, 85
and 90 in an initially empty B-Tree. Give the number of
nodes splitting operations that take place.
Ans. Insert 10, 25, 20, 35 :
20

Split
10 20 25 35  10 25 35
Insert 30 :

20

10 25 30 35
Insert 55 :
20 20 30

Split
10 25 30 35 55  10 25 35 55
Insert 40, 45 :

20 30 20 30 40

Split
10 25 35 40 45 55  10 25 35 45 55
Insert 50, 55 :
Solved Paper (2019-20) SP–14 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in

20 30 40 20 30 40 50
Split

10 25 35 45 50 55 55 10 25 35 45 55 55


Split
30

20 40 50

10 25 35 45 55 55

Insert 60, 75 :

30

20 40 50 Split

10 25 35 45 55 55 60 75

30

20 40 50 55

10 25 35 45 55 60 75

Insert 70, 65 :
Design and Analysis of Algorithms SP–15 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
30

20 40 50 55 Split


10 25 35 45 55 60 65 70 75

30

40 50 55 65 Split
20


10 25 35 45 55 60 70 75

30 50

20 40 55 65

10 25 35 45 55 60 70 75
Insert 80, 85 :
30 50

20 40 55 65
Split

10 25 35 45 55 60 70 75 80 85

30 50

20 40 55 65 75

10 25 35 45 55 60 70 80 85
Insert 90 :
30 50

20 40 55 65 75

10 25 35 45 55 60 70 80 85 90
Number of nodes splitting operations = 9.
Solved Paper (2019-20) SP–16 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Explain the algorithm to delete a given element in a binomial
heap. Give an example for the same.
Ans. Deletion of key from binomial heap :
The operation BINOMIAL-HEAP-DECREASE (H, x, k) assigns a
new key ‘k’ to a node ‘x’ in a binomial heap H.
BINOMIAL-HEAP-DECREASE-KEY (H, x, k)
1. if k > key [x] then
2. Message “error new key is greater than current key”
3. key [x]  k
4. y  x
5. z  P [y]
6. While (z  NIL) and key [y] < key [z]
7. do exchange key [y]  key [z]
9. y  z
10. z  P [y]
Deleting a key : The operation BINOMIAL-HEAP-DELETE
(H, x) is used to delete a node x’s key from the given binomial heap
H. The following implementation assumes that no node currently
in the binomial heap has a key of – .
BINOMIAL-HEAP-DELETE (H, x)
1. BINOMIAL-HEAP-DECREASE-KEY (H, x, – )
2. BINOMIAL-HEAP-EXTRACT-MIN(H)
For example : Operation of Binomial-Heap-Decrease (H, x, k) on
the following given binomial heap :
Suppose a binomial heap H is as follows :

Head[H] 37 10 1

41 28 13 6 16 12 25

77 8 14 29 26 23 18

11 17 38 42

27

The root x with minimum key is 1. x is removed from the root list of
H. i.e.,
Design and Analysis of Algorithms SP–17 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
x
Head[H] 37 10 1

41 28 13 6 16 12 25

77 8 14 29 26 23 18

11 17 38 42

27
Now, the linked list of x’s children is reversed and set head[H] to
point to the head of the resulting list, i.e., another binomial heap H.

Head[H ] 25 12 16 6

18 26 23 8 14 29

42 11 17 38

27

Now, call BINOMIAL-HEAP-UNION (H, H) to uniting the two

binomial heaps H and H. The resulting binomial heap is
Head[H]
25 12 6

37 18 10 8 14 29

41 16 28 13 11 17 38

26 23 77 27

42

5. Attempt any one part of the following : (7 × 1 = 7)

a. Compare the various programming paradigms such as
divide-and-conquer, dynamic programming and greedy
approach.
Solved Paper (2019-20) SP–18 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Ans.
S. No. Divide and Dynamic Greedy
conquer programming approach
approach approach
1. Optimizes by Same as Divide and Optimizes by
breaking down a Conquer, but optimizes making the best
subpro blem into by caching the answers choice at the
simpler versions of to each subproblem as moment.
itself and using multi- no t to repeat the
thre ading and calculation twice.
recursion to solve.
2. Always finds the Always finds the optimal Does not always
optimal solution, but solutio n, but canno t find the optimal
is slower than work on small datasets. solution, but is
Greedy. very fast.
3. Re quires some Requires a lot of memory Requires almost
me mo ry to for tabulation. no memory.
remember recursive
calls.

b. What do you mean by convex hull ? Describe an algorithm

that solves the convex hull problem. Find the time
complexity of the algorithm.
Ans.
1. The convex hull of a set S of points in the plane is defined as the
smallest convex polygon containing all the points of S.
2. The vertices of the convex hull of a set S of points form a (not
necessarily proper) subset of S.
3. To check whether a particular point p S is extreme, see each
possible triplet of points and check whether p lies in the triangle
formed by these three points.

Fig. 1.
4. If p lies in any triangle then it is not extreme, otherwise it is.
5. We denote the convex hull of S by CH(S). Convex hull is a convex
set because the intersection of convex sets is convex and convex
hull is also a convex closure.
Design and Analysis of Algorithms SP–19 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Graham-Scan algorithm :
The procedure GRAHAM-SCAN takes as input a set Q of points,
where |Q|  3. It calls the functions Top(S), which return the point
on top of stack S without changing S, and to NEXT-TO-TOP(S),
which returns the point one entry below the top of stack S without
changing S.
GRAHAM-SCAN(Q)
1. Let p0 be the point in Q with the minimum y-coordinate, or the
leftmost such point in case of a tie.
2. Let <p1, p2, ...., pm> be the remaining points in Q, sorted by polar
angle in counter clockwise order around po (if more than one point
has the same angle remove all but the one that is farthest from po).
3. Top [S]  0
4. PUSH (p0, S)
5. PUSH (p1, S)
6. PUSH (p2, S)
7. for i  3 to m
8. do while the angle formed by points NEXT-To-TOP(S), Top(S), and
pi makes a non left turn.
9. do POP(S)
10. PUSH (pi, S)
11. return S
Time complexity :
The worst case running time of GRAHAM-SCAN is
T(n) = O(n) + O(n log n) + O(1) + O(n) = O(n log n)
where n = |Q|
Graham’s scan running time depends only on the size of the input
it is independent of the size of output.

6. Attempt any one part of the following :

a. Solve the following 0/1 knapsack problem using dynamic
programming P = {11, 21, 31, 33} w = {2, 11, 22, 15} c = 40,
n = 4.
Ans. Numerical :
w = {2, 11, 22, 15}
c = 40
p = {11, 21, 31, 33}
Initially,
Solved Paper (2019-20) SP–20 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
Item wi pi
I1 2 11
I2 11 21
I3 22 31
I4 15 33
Taking value per weight ratio, i.e., pi / wi
Item wi vi / wi pi
I1 2 11 22
I2 11 21 232
I3 22 31 682
I4 15 33 495

Now, arrange the value of pi in decreasing order.

Item wi pi pi
I3 22 31 682
I4 15 33 495
I2 11 21 232
I1 2 11 22

Now, fill the knapsack according to decreasing value of pi.

First we choose item I3 whose weight is 22, then choose item I4
whose weight is 15. Now the total weight in knapsack is 22 + 15 =
37. Now, next item is I2 and its weight is 11 and then again I1. So, we
choose fractional part of it, i.e.,

22
15
40
3
The value of fractional part of I1 is,
232
= × 3 = 63
11
Thus, the maximum value is,
= 682 + 495 + 63 = 1190
Design and Analysis of Algorithms SP–21 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
b. Define Floyd Warshall algorithm for all pair shortest path
and apply the same on following graph :
3
1 2
4
6 3 5
1
1 2
3 4
2
Fig. 2.
Ans. Floyd Warshall algorithm :
1. Floyd-Warshall algorithm is a graph analysis algorithm for finding
shortest paths in a weighted, directed graph.
2. A single execution of the algorithm will find the shortest path
between all pairs of vertices.
3. It does so in (V3) time, where V is the number of vertices in the
graph.
4. Negative-weight edges may be present, but we shall assume that
there are no negative-weight cycles.
5. The algorithm considers the “intermediate” vertices of a shortest
path, where an intermediate vertex of a simple path p = (v1, v2, ...,
vm) is any vertex of p other than v1 or vm, that is, any vertex in the
set {v2, v3, ..., vm–1}.
6. Let the vertices of G be V = {1, 2, ..., n}, and consider a subset {1, 2,
..., k} of vertices for some k.
7. For any pair of vertices i, j  V, consider all paths from i to j whose
intermediate vertices are all drawn from {1, 2, ..., k}, and let p be a
minimum-weight path from among them.
8. Let dij( k) be the weight of a shortest path from vertex i to vertex j
with all intermediate vertices in the set {1, 2, ..., k}.
A recursive definition is given by
 wij if k  0
dij( k) =  ( k –1) ( k –1) ( k –1)
 min( dij , dik  dkj ) if k  1
Numerical :
dij(k) = min[ dij(k–1) , dik
(k –1) ( k –1)
 dkj ]

 (ijk–1) if dij(k –1)  dik

( k –1) ( k 1)
 dkj
 (ijk) = 
( k1) ( k –1) ( k1) ( k1)
 kj if dij  dik  dkj
Solved Paper (2019-20) SP–22 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
1 2 3 4 5
1 0  6 3 

2 3 0   
D(0) =
3   0 2 

4  1 1 0 

5  4  2 0

1 2 3 4 5
1 0 4 6 3 

2 3 0 9 6 
D(1) = 
3 6 4 0 2

4 4 1 1 0 

5 7 4 3 2 0

1 2 3 4 5
1 0 4 6 3 

2 3 0 7 6 
D(2) = 3 6 3 0 2 

4 4 1 1 0 

5 6 3 3 2 0

Now, if we find D(3), D(4) and D(5) there will be no change in the
entries.

7. Attempt any one part of the following : (7 × 1 = 7)

a. Describe in detail Knuth-Morris-Pratt string matching
algorithm. Compute the prefix function  for the pattern
ababbabbabbababbabb when the alphabet is  = {a, b}.
Ans. Knuth-Morris-Pratt algorithm for string matching :
COMPUTE-PREFIX-FUNCTION (P)
1. m  length [P]
2. [1]  0
3. k  0
Design and Analysis of Algorithms SP–23 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
4. for q  2 to m
5. do while k > 0 and P [k + 1]  P[q]
6. do k  [k]
7. if P[k + 1] = P [q]
8. then k  k + 1
9. [q]  k
10. return 
KMP-MATCHER calls the auxiliary procedure COMPUTE-PREFIX-
FUNCTION to compute .
KMP-MATCHER (T, p)
1. n  length [T]
2. m  length [P]
3.  COMPUTE-PREFIX-FUNCTION (P)
4. q  0
5. for i  1 to n
6. do while q > 0 and P [q + 1]  T[i]
7. do q  [q]
8. if P [q +1] = T[i]
9. then q  q + 1
10. if q = m
11. then print “pattern occurs with shift” i – m
12. q  [q]
Numerical :
pattern = ababbabbabbababbabb
length 19
Initially, (1) = 0 and k = 0
For q  2 to 9
For q = 2 and k / 0
P[0 + 1]  P[2].
[2] = 0
For q= 3
P[0 + 1] = P[3]
k= k+1=1
[3] = 1
For q= 4k>0
P[1 + 1] = P[4]
P[2] = P[4]
k= k+1=2
[4] = 2
For q= 5k>0
P[2 + 1]  P[5]
[5] = 0
For q= 6
P[0 + 1] = P[6]
Solved Paper (2019-20) SP–24 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
k= k+1=1
[6] = 1
For q= 7k>0
P[1 + 1] = P[7]
P[2] = P[7]
k= k+1=2
[7] = 0
For q= 8k>0
P[2 + 1] = P[8]
P[3]  P[8]
[8] = 0
For q= 9
P[0 + 1] = P[9]
k= k+1
[9] = 2
For q= 10 k > 0
P[2 + 1]  P[10]
[10] = 0
For q= 11 k > 0
P[0 + 1]  P[11]
[11] = 0
For q= 12 k > 0
P[0 + 1]  P[12]
k= k+1
[12] = 2
For q= 13 k > 0
P[2 + 1]  P[13]
[13] = 0
For q= 14
P[0 + 1] = P[14]
k= k+1
[14] = 2
For q= 15 k > 0
P[2 + 1]  P[15]
[15] = 0
For q= 16 k > 0
P[0 + 1]  P[16]
[16] = 0
For q= 17
P[0 + 1]  P[17]
k= k+1
[17] = 2
For q= 18 k > 0
P[2 + 1]  P[18]
[18] = 0
For q= 19
P[0 + 1]  P[19]
Design and Analysis of Algorithms SP–25 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
[19] = 0
String a b a b b a b b a b b a b a b b a b b
P[i] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
[i] 0 0 1 2 0 1 2 0 2 0 0 2 0 2 0 0 2 0 0

b. What is an approximation algorithm ? What is meant by

p(n) approximation algorithms ? Discuss approximation
algorithm for Travelling Salesman Problem.
Ans. Approximation algorithm :
1. An approximation algorithm is a way of dealing with NP-
completeness for optimization problem. This technique does not
guarantee the best solution.
2. The best of an approximation algorithm is to come as close as possible
to the optimum value in a reasonable amount of time which is at
most polynomial time.
3. Let c(i) be the cost of solution produced by approximate algorithm
and c*(i) be the cost of optimal solution for some optimization
problem instance i.
4. For minimization and maximization problem, we are interested in
finding a solution of a given instance i in the set of feasible solutions,
such that c(i) / c*(i) and c*(i) / c(i) be as small as possible respectively.
5. We say that an approximation algorithm for the given problem
instance i, has a ratio bound of p(n) if for any input of size n, the
cost c of the solution produced by the approximation algorithm is
within a factor of p(n) of the cost c* of an optimal solution. That is
max(c(i) / c*(i), c*(i) / c(i))  p(n)
p(n) approximation algorithm : A is a p(n) approximate
algorithm if and only if for every instance of size n, the algorithm
achieves an approximation ratio of p(n). It is applied to both
maximization (0 < C(i)  C*(i)) and minimization (0 < C* (i)  C(i))
problem because of the maximization factor and costs are positive.
p(n) is always greater than 1.
Approximation algorithm for Travelling Salesman Problem
(TSP) :
1. The key to designing approximation algorithm is to obtain a bound
on the optimal value (OPT).
2. In the case of TSP, the minimum spanning tree gives a lower bound
on OPT.
3. The cost of a minimum spanning tree is not greater than the cost of
an optimal tour.
The algorithm is as follows :
1. Find a minimum spanning tree of G.
Solved Paper (2019-20) SP–26 B (CS/IT-Sem-5)
More pdf : www.motivationbank.in
2. Duplicate each edge in the minimum spanning tree to obtain a
Eulerian graph.
3. Find a Eulerian tour (J) of the Eulerian graph.
4. Convert J to a tour T by going through the vertices in the same
order of T, skipping vertices that were already visited.


 More pdf
ST :Uwww.motivationbank.in
D YZONE A D I T YA .COM
www.studyzoneaditya.com

ENGINEERING ALL CONTENTS

Books, Quantum, Typing Notes, Handwritten Notes,
Practical File's And All Engineering Study Materials
Available Here.

For more details visit and follow touch to icons below

studyzoneaditya.com B.TECH STUDY ZONE

STUDY ZONE B.TECH STUDY ZONE