MEI

Mathematics in Education and Industry

MEI Structured Mathematics

Module Summary Sheets

Decision Mathematics 1
(Version B: reference to new book)
Topic 1: Algorithms
Topic 2: Graphs
Topic 3: Networks
Topic 4: Critical Path Analysis
Topic 5: Linear Programming
Topic 6: Simulation

Purchasers have the licence to make multiple copies for use

within a single establishment
MEI November 2004
MEI, Oak House, 9 Epsom Centre, White Horse Business Park, Trowbridge, Wiltshire. BA14 0XG.
Company No. 3265490 England and Wales Registered with the Charity Commission, number 1058911
Tel: 01225 776776. Fax: 01225 775755.

Summary D1 Topic 1:
References:
Chapter 1
Pages 1-7

Exercise 1B
Q. 2

Algorithms

Terminology

Examples

An Algorithm is a set of instructions for carrying out a task.

There are many ways in which an algorithm can be
expressed, including:
Written language
Pseudo-code (an abbreviated form of language, putting
on instruction per line and making use of specific words
such as repeat)
Flow chart
Computer program

Division by repeated subtraction (Pseudo

code)
Read A,B (A>B; A,B>0)
Counter=0
repeat
Set A to A-B
Increment Counter
until A<B
Write Answer Counter
Write Remainder A

Order of Complexity the order of an algorithm is a

measure of the amount of work (number of operations,
instructions) needed to complete it. It is expressed in terms
of n, where n is the size of the problem. It is a proportion,
not an exact measure. O(n) means that the algorithm takes
an amount of effort proportional to the square of its size (in
its worst case).
References:
Chapter 1
Pages 7-9

Heuristic Algorithms attempt to find a good solution to a

problem, but are not guaranteed to find the optimal or
complete solution.

Draw all connections between n points.

There are n points, so each can be connected to n-1 others. Therefore we could
make n(n-1) connections, but we will then
have done each one twice. Therefore the
total is n(n-1) = n - n
When n is large, n is considerably bigger than n so this dominates the expression. Since n is directly proportional to
n, we say that this algorithm is O(n)
For example Bin Packing algorithms
(see below)
Factorial: 10! = 10 9! = 10 9 8! =

Recursion defining something in terms of itself .

References:
Chapter 1
Pages 15-19

Standard Types of Algorithm

Searching
Finding a particular item in a list

Linear
Binary
Index

Exercise 1C
Q. 1

Sorting
Placing all items in a list in the correct order

Exchange Find the smallest and swap it with the first. Find the
next smallest and swap it with the second; etc.
Bubble
On the first pass compare adjacent items (1st with 2nd,
2nd with 3rd etc.), and swap if necessary. Repeat until
all items sorted.
Quick
Choose a pivot value and sort items into two sub-lists,
containing items smaller and larger than the pivot respectively. Repeat algorithm on each sub-list until all
sub-lists are of length 1 or 0.

Bin Packing
The process of allocating items to
locations, given a set of items of
specified sizes and a set of locations
with specified capacities.

First fit

Mathematical
Leading to the calculation of a required value

HCF, LCM, Factorial

References:
Chapter 1
Pages 20-24

Exercise 1D
Q. 1

Exercise 1E
Q. 4

Every data item is checked

For ordered data. Repeatedly examines the middle to
decide which half contains the target.
Search the index to find the page, then search the
page.

NOTE: Syllabus requires knowledge of properties of

algorithms in all topics.

Take each item in turn, placing it in the first available

slot.
First Fit Decreasing order the items in decreasing size then apply
First Fit.

Decision Mathematics 1
Version B: page 2
Competence statements A1, A2, A3, A4
MEI

Summary D1 Topic 2:
References:
Chapter 2
Pages 43-45

Graphs

Terminology
Examples
Graph collection of vertices & edges

Multiple edges

Sub graph any set of edges & vertices taken from a graph
is a sub-graph
References:
Chapter 2
Pages 47-50

vertices

Vertex/Node the dots in a graph (usually where 2 or more

edges meet, but not necessarily)

edge

Edge/arc a line between two vertices

Exercise 2A
Q. 4

loop

Isomorphic - two graphs are isomorphic if one can twisted

into the shape of the other ie: by relabelling the vertices of
one graph you can show that it is the same as the other
Weight the weight of an edge is a number representing a
real-world value, often distance or time
Multiple edges two or more edges joining the same pair of
vertices

These two graphs

are isomorphic.

Loop an edge starting and finishing at the same vertex

15

Simple graph a graph with no loops or multiple edges

Connected graph a graph in which a route can be found
between any pair of vertices (ie the graph is in one part)
Cycle a route starting and finishing at the same vertex

tree

Tree a graph with no cycles

Degree of a vertex the number of edges starting or finishing at that vertex
Handshake lemma the sum of all the degrees in a graph is
twice the number of edges (and therefore even); consequently there will always be an even number of odd vertices
Di-graph a graph in which the edges indicate direction

digraph

Incidence matrix a matrix representing the edges in a

graph
The topic of graphs is fundamental to much of Discrete Maths, and the theory
of graphs is a large topic in its own right. However, this course focuses on
ways of using graphs. Exam questions could test you on your knowledge of
graphs; they could also ask you to apply a graph to solve a practical problem.
Questions like this in Ex 2A, include:2,3,4,5,9,12,15,16,17
Other questions, including some taken from past examinations of other syllabuses, involve terminology and notation not specified in this syllabus. Tackling
these, and investigating the terms involved, will enhance your knowledge of
this subject, and your modelling skills.
These questions include:6,7,8,10,11,13,14

From
Fro

To
A B C
A 1 1
B 1 1
C 1 1 2

Incidence matrix and graph

Decision Mathematics 1
Version B: page 3
Competence statements g1, g2
MEI

Summary D1 Topic 3:

Networks

References:
Chapter 3
Pages 60-61

Terminology
Network - a graph with weighted arcs (or edges)

References:
Chapter 3
Pages 62-65

Minimum connector the smallest possible tree which leaves a

graph connected (spanning tree)

Example
B
2

Kruskals algorithm
1
Select the shortest edge in a network
2
Select the next shortest edge which does not create a cycle Kruskals
3
Repeat step 2 until all vertices have been connected

Exercise 3A
Q. 1

References:
Chapter 3
Pages 68-79

Prims algorithm
1
Select any vertex
2
Select the shortest edge connected to that vertex
3
Select the shortest edge which connects a previously
chosen vertex to a new vertex
4
Repeat step 3 until all vertices have been connected
Shortest path the shortest route between any two nodes in a
network

1
2
3

C
Prims
(starting from A)
1
AB
2
BD
3
DC

AB
CD
BD

Example using Dijkstra, from A D

B
D

Dijkstras algorithm
Exercise 3B
Q. 1

Step 1 Begin by giving the start vertex permanent label 0,

and order label 1.

2
Exercise 3C
Q. 2, 3

Exercise 3D
Q. 4

References:
Chapter 3
Pages 86-87

Put in temporary values for each vertex you can reach

directly from the start.
Step 2 Pick the vertex with the smallest temporary value, and
make that the permanent value. Put in the correct order label.
Then update the temporary values for all the vertices you can
reach directly from the vertex you've just made permanent.

D
A
C

Repeat step 2 (always looking at all the temporary values for

the smallest), until you have a permanent value at your target
vertex.

B
D
A
C

NB If at any stage you can choose between two equally good

options, pick either; you will always pick the other one next.
4

Exercise 3E
Q. 4

Step 3 Trace back to find the route - but write it down

forwards, giving the total distance covered.

D
A

NB If you find yourself with a choice of arcs when tracing

back, your network will have more than one optimum
solution.

C
B

Decision Mathematics 1
Version B: page 4
Competence statements N1, N2, N3, N4
MEI

D
A
C

Summary D1 Topic 4:
References:
Chapter 4
Pages 101-107

Critical Path Analysis

Terminology
Example
A Critical Path Analysis question will consist of a project to Make a cup of instant black coffee
be completed by carrying out a number of tasks.
Activity

Activity a task to be done which takes a defined amount of

time. Represented by an edge on a graph.
Event an instant in time when one or more activities start
or finish. Represented by a vertex.
Exercise 4A
Q. 4

Precedence Table lists each activity and those activities

which must immediately precede it.
Critical Path the set of edges (activities) which must be
carried out at a fixed time for the project to be completed in
the minimum time.

Duration (secs)

Preceding
activities

A Fill Kettle

20

B Boil Kettle

90

C Put water in mug

10

B, D, E

D Put coffee in
mug

20

E Put sugar in mug

15

The above precedence table leads to the following graph (activity network)

Event times each event has an earliest and a latest time:

these are the earliest that all incoming activities could
finish, and the latest that all outgoing activities could
commence, without affecting the minimum completion
time.
Float - the amount of time by which an activity (not on the
critical path) can be delayed or extended without affecting
the critical path timing.
independent float does not affect other activities
interfering float means that it is dependent on other
activities not using all of their float
References:
Chapter 4
Pages 111-115

Exercise 4B
Q. 3

Exercise 4C
Q. 9

Resourcing
Resourcing is the task of finding how many people are
required to complete the project in the minimum time.
1. Each task requires one person. In this case the critical
path will usually be completed by a single person. The
other tasks are scheduled (possibly making use of float
time to delay the start times) so that the least number of
people are required [there are links here to the concepts
of bin-packing from the algorithms section].

The events are shown as circles, the activities as

edges. All precedences are maintained; a
dummy activity is inserted after E so that each
activity is uniquely defined by (start event, finish
event).
The minimum completion time (assuming tasks
can be done in parallel) is 120 seconds. The critical activities for this are A, B and C, since a delay
in any of these would lead to a delay in the final
completion time.

The event at the end of E has an earliest time of

15 seconds (one cannot get there sooner), and a
2. Tasks require different numbers of people. A resource latest time of 110 seconds (since anything later
histogram should be drawn. Then, as above, start times than this would delay the start of C and hence the
are adjusted in an attempt to even out the requirments entire project).
this is known as resource levelling.
Activity D has float time of 90 seconds (since it
Dummy Activity - these are represented by dotted or faint
can start at 0, must finish by 110, and takes 20
lines and are added for two reasons:
seconds itself).
1. To avoid two activities sharing both start and finish
2.

events.
To ensure correct logic.

Example: if C is dependent on A and B, but D is only

dependent on B

Decision Mathematics 1
Version B: page 5
Competence statements X1, X2, X3, X4, X5
MEI

Summary D1 Topic 5:
References:
Chapter 5
Pages 138-145

Linear Programming

Terminology
Variable a named quantity which can take more than one
value.

Examples
x, y, Xab, Product1
maximise 2x + 3y

Objective Function an expression which is to be maximised or minimised, expressed as a linear function of the
variables.

P = 2x + 3y
x + 2y 12
x3
2x y = 4

Profit/Cost the value of the objective function, depending

on whether it is a maximisation or minimisation problem.
Constraint an equation or inequality involving at least one
variable which limits the value(s) the variables can take.
Feasible region that part of the graph which represents
solutions which satisfy all of the constraints
Optimum solution the solution offering the best value of
the objective function
References:
Chapter 5
Pages 141-142

Graphical Solutions
(assuming a two-variable maximisation problem)
Label axes for each variable.
For each constraint, draw the line representing the
equality. For inequalities, shade out the unwanted side
(above for <, below for >).
Shade out left of vertical and below horizontal axes.

Example 5.1
Page 141-145

Exercise 5A
Q. 2

References:
Chapter 5
Pages 148-156

The optimum solution will be at one (or two adjacent)

vertices of the feasible region. It can thus be found by
evaluating the Profit function at each vertex.
This point will correspond to the point where a line
equal in gradient to the profit function and moving
across the feasible region away from the origin, will
last touch the feasible region.
Integer problems
If the optimum solution is not integer, evaluate the objective function at the integer point(s) [usually near to
the optimum point] and select the best.
Minimisation problems
These are done the same way at this level.

Exercise 5B
Q. 3

Exercise 5C
Q. 3

Example
Maximise 2x + 3y subject to
2x + y 8, x + 2y 8
The gradient of P = 2x + 3y is 2/3

Multiple solutions
These occur when the objective function is parallel to a
constraint, and occasionally in integer problems.
Spare capacity
Any constraint which does not pass through the optimal solution will have spare capacity which can be
calculated. This can be done by substituting the values
for the variables and noting the difference across the
inequality sign.

Decision Mathematics 1
Version B: page 6
Competence statements L1, L2, L3, L4,L 5
MEI

The optimum solution occurs at the intersection

of the lines 2x + y 8,
x + 2y 8
which is x = 2, y = 3 with a profit of 14.
This can be estimated from the graph, or
calculated by solving the simultaneous
equations.
If Integer solutions are required, then the
nearest integer point in the feasible region is
(2, 3) so the solution is x = 2, y = 3 with a profit
of 13. However, it is important to check other
integer points nearby (e.g. 2, 4) to see if they
are also feasible and give a greater profit.
Multiple solutions would occur if the profit
function changed to say P = 2x + y: it would be
all points on the constraint line between (2, 3)
and (4,0).
At the integer solution the constraints have
spare capacity of 1 and 1/2 respectively.

Summary D1 Topic 6:
References:
Chapter 6
Pages 168-169

Exercise 6A
Q. 3
References:
Chapter 6
Pages 173-174

Exercise 6B
Q. 1
References:
Chapter 6
Page 174

Simulation

Terminology
Simulation: a mathematical model that can be used to
test what might happen in situations where an
experiment with real subjects may be too dangerous
or take too long.
Deterministic model: assumes that chance events do
not occur.
Stochastic model: allows for an element of chance.
Random variables: used in stochastic models to
simulate probabilities. Of two types
Uniformly distributed random variables
Non-uniform random variables
Rules of the simulation and Cumulative frequency
methods: allocating the random variables in line with
the probabilities
Inter arrival times: how often a new customer joins
the queue

including rejecting values where necessary

2. Make simplifying assumptions

All times are taken to the nearest minute

Any times which fall outside this range will be ignored.

4. Perform the simulation

Use the random numbers to obtain the inter-arrival
times, the actual arrival times and service times.
Inter arrival time random numbers: 13, 24, 88, 12

Inter-arrival time (mins)

0.2

0.5

0.3

Random Numbers

00-19

20-69

70-99

Service time (mins)

probability

probability
Random Numbers

/6

00-15

/2

16-63

/3

64-95

The simulation is usually set out in a table.

Service time random numbers: 58, 73, 11, 98, 39
(Ignore 98)

In the table, the beginning of the simulation is usually

counted as time 0 and subsequent arrival times are
given as time from the beginning of the simulation.
5. Interpret the results
How does this compare with reality?

Exercise 6D
Q. 4

coins or dice
random number tables, random number generated
from a calculator

When modelling queuing it is necessary to find out the

distributions for inter-arrival times and serving time.

100/6 = 16 to nearest whole number. Allocate random

numbers in multiples of 16 and reject any which are
left over.

References:
Chapter 6
Page 178

Queuing simulations are a common example

Queuing simulations
1. Decide on the important variables

3. Write the rules of the simulation.

Use the probabilities to allocate random numbers.
These are usually 2 digit numbers from 00 to 99.

Exercise 6C
Q. 3

Examples
What system of queuing will minimise customer waiting times in a bank?
How does an infectious disease spread through a
population?

6. Amend the model

Often the first run is to validate the model to check
that it models accurately the current or known
situation.

Decision Mathematics 1
Version B: page 7
Competence statements Z1, Z2, Z3, Z4, Z5
MEI

customer

RND

Arrival
time

Service
start

RND

Service
time

Service
ends

Queing
time

58

13

73

24

11

88

39

Average queuing time =

total queuing time

3
= min per customer
number of customers 4

Note: the above simulation starts at the arrival of the first

customer, otherwise an initial arrival time would be
needed.
A simulation would normally have a much longer run
than this.
A simulation should be performed several times to get a
more reliable result.