PART – 1

Linear Programming Problems

Formulation of Linear Programming Problems

Example 1 : The following is a data for the Readymade Garment Production Unit.
Formulate the LPP from the data :

Profit (Rs.) 175 200 300 150

Products Shirt Trouser Jeans T-Shirt Machine
→ Capacity
(Min/Day)
M1 60 Min. 90 Min. 150 Min. 50 Min. 400
M2 125 Min. 160 Min. 180 Min. 75 Min. 300
M3 75 Min. 60 Min. 100 Min. 80 Min. 450

Solution :

We define the variables as follows :

Let x1 = Number of shirts produced

Let x2 = Number of trousers produced

Let x3 = Number of Jeans produced

Let x4 = Number of T-shirts produced

➢ Objective Function

Maximize Z = 175 x1 + 200 x2 + 300 x3 +150 x4

➢ Constraints :

• Machine 1 : 60 x1 + 90 x2 + 150 x3 + 50 x4 < 400

• Machine 2 : 125 x1 + 160 x2 + 180 x3 + 75 x4 < 300

• Machine 3 : 75 x1 + 60 x2 + 100 x3 + 80 x4 < 450

 ➢ Non-negativity Conditions : x1, x2, x3, x4  0

Hence the LPP is :

Maximize Z = 175 x1 + 200 x2 + 300 x3 +150 x4

Subject to : 60 x1 + 90 x2 + 150 x3 + 50 x4 < 400
125 x1 + 160 x2 + 180 x3 + 75 x4 < 300
75 x1 + 60 x2 + 100 x3 + 80 x4 < 450
Where x1, x2, x3, x4  0

Example 2 (Courtesy : Book : Operations Research : J K Sharma, 5th

Edition)
• A manufacturing company is engaged in producing three types of products :
A , B & C.

• The production department produces each day, components sufficient to

make 50 units of A, 25 units of B and 30 units of C. The management in
confronted with the problem of optimizing the daily production of the
products in the assembly department, where only 100 man-hours are
available daily for assembling the products. The following additional
information is available.

Type Profit distribution per Assembly time per

product unit of product(Rs) period (hours)

A 12 0.8
B 20 1.7
C 45 2.5
The company has a daily order commitment for 20 units of product A and a total
of 15 units of products B and C.

Formulate this problem as an LP model so as to maximize the total profit.

Solution :

Transportation Problem

Mumbai Delhi Kolkata Chennai Capacity

Kandla 5 10 24 22 400
Gudgaon 12 4 18 20 800
Kolkata 13 16 2 19 600
Bangalore 7 15 18 3 700
Demand 900 500 400 700

Transportation Problem – Defining variables :

Mumbai Delhi Kolkata Chennai Capacity

Kandla 5/x 10 / x 24 / x 22 / x 400
11 12 13 14
Gudgaon 12/ x 4/x 18/ x 20/ x 800
21 22 23 24
Kolkata 13/ x 16/ x 2/ x 19/ x 600
31 32 33 34
Bangalore 7/ x 15/ x 18/ x 3/ x 700
41 42 43 44
Demand 900 500 400 700

Diet Problem
• Diet is a very important aspect for the health of an individual. Specially for
the patients, diet plan has great importance. Suppose a doctor advises a
patient to have minimum of 3500 units of vitamins, 2800 units of calories
and 175 units of minerals. Suppose a person has three foods options having
different units of vitamins, calories and minerals, namely P, Q and R. One
 unit of P, Q and R cost Rs. 55, Rs. 90 and Rs. 120 respectively. One unit of P
has 160 units of vitamins, 100 units of calories and 40 units of minerals.
One unit of Q has 125 units of vitamins, 140 units of calories and 70 units of
minerals. One unit of R has 180 units of vitamins, 50 units of calories and
100 units of minerals. Formulate the diet model to have the minimum
expenses.

• Diet is very important aspect for the health of an individual. Specially for
the patients, diet plan has great importance. Suppose a doctor advises a
patient to have minimum of 3500 units of vitamins, 2800 units of calories
and 175 units of minerals. Suppose a person has three foods options having
different units of vitamins, calories and minerals, namely P, Q and R. One
unit of P, Q and R cost Rs. 55, Rs. 90 and Rs. 120 respectively. One unit of P
has 160 units of vitamins, 100 units of calories and 40 units of minerals.
One unit of Q has 125 units of vitamins, 140 units of calories and 70 units of
minerals. One unit of R has 180 units of vitamins, 50 units of calories and
100 units of minerals. Formulate the diet model to have the minimum
expenses.

Defining Variables :
Let

• x1 = Number of units of food P purchased

• x2 = Number of units of food Q purchased

• x3 = Number of units of food R purchased

Objective Function

Minimize Z = 55 x1 + 90 x2 + 120 x3

Constraints :

• Vitamins :

One unit of P has 160 units of vitamins,

 One unit of Q has 125 units of vitamins,

One unit of R has 180 units of vitamins.

A patient to have minimum of 3500 units of vitamins

So, vitamin constraint is : 160 x1 + 125 x2 + 180 x3  3500

• Similar for calories and minerals…..

Constraints :
• Vitamins : 160 x1 + 125 x2 + 180 x3  3500

• Calories : 100 x1 + 140 x2 + 50 x3  2800

• Minerals : 40 x1 + 70 x2 + 100 x3  175

Non-negativity Conditions

x1, x2, x3  0

Formulate the LPP :

• A company has two types of production units, namely Super (level) and
medium (level) units. The company wants to produce at least 5000 units
per day, having 8 working hours. Super unit can manufacture 100 units per
hour, but it produces 3 % defective items. Medium unit can manufacture 80
units per hour, but it produces 5 % defective items.

• The cost of affording a super unit is Rs 1800 per hour while that of a
medium unit is Rs 1200 per hour. Every defective item costs the company a
loss of Rs 300. There is an availability of only ten super units and fifteen
medium units. The company plans to assign work to the available units so
as to minimize the total daily cost of production. Formulation this problem
as an LP model.

Defining Variables :
Let

• x1 = Number of super production units

• x2 = Number of medium production units

For One Super Unit :

• Cost per hour = Rs. 1800 + Cost of loss occurred due to defective items

• But super unit produces defective items/hour = 100*0.03 = 3

• So, loss per hour = 3*300 = 900.

• Cost per hour = Rs. 1800 + 900 = 2700.

• Cost per DAY per super unit = 8(2700) = 17600.

For One Medium Unit :

• Cost per hour = Rs. 1200 + Cost of loss occurred due to defective items

• But super unit produces defective items/hour = 80*0.05 = 4

• So, loss per hour = 4*300 = 1200.

• Cost per hour = Rs. 1200 + 1200 = 2400.

• Cost per Day per medium unit = 8*2400 = 19200.

Objective Function
Minimize Z = 17600 x1 + 19200 x2

Constraints :

• Total units : (8*100) x1 + (8*80) x2  5000 i.e., 800 x1 + 640 x2  5000

• Super units : x1  10

• Medium units : x2  15
Non-negativity Conditions

x1, x2  0

Solution of LPPs : Graphical Method

• Solve :

• Maximize Z = 2x1 + 3x2

• Subject to the constraints :

x1 + x2  30

x1  20

x2  12

Where x1, x2  0.

Solution :

Consider x1 + x2 = 30

X1 30 0 15
X2 0 30 15

• Consider x1 = 20 :

X1 20 20 20
X2 15 0 30

• Consider x2 = 12 :

X1 30 0 15
X2 12 12 12
Calculation of Z at the vertices :

Vertex of feasible region Value of Z = 2x + 3x

1 2
at the vertex
(0, 0) 0
(20, 0) 40
(20, 10) 70
(18, 12) 72
(0, 12) 36

Optimal Feasible Solution is :

x1 = 18 x2 = 12 & Maximum Z = 72

Solution of LPPs : Graphical Method

• Solve : Maximize Z = 15x1 + 10x2

• Subject to the constraints :

2x1 + 3x2  180

 x1  60

x2  40 Where x1, x2  0.

Solution :

• Consider 2x1 + 3x2 = 180 :

• Consider x1 = 60 :

• Consider x2 = 40 :

Calculation of Z at the vertices :

Vertex of feasible region Value of Z = 15x + 10x

1 2
at the vertex
(0, 0) 0
(60, 0) 900
(60, 20) 1100
(30, 40) 850
(0, 40) 400

Optimal Feasible Solution is :

x1 = 60 ,x2 = 20 & Maximum Z = 1100

Solution of LPPs : Graphical Method

• Solve :

• Minimize Z = - x1 + 2x2

• Subject to the constraints :

- x1 + 3x2  10

x1 + x2  6

x1 - x2  2 Where x1, x2  0.
Optimal Feasible Solution is :

x1 = 2

x2 = 0

& Minimum Z = - 2

Solution of LPPs : Graphical Method

• Solve :

• Maximize Z = 2 x1 + x2

• Subject to the constraints :

x1 + 2x2  10

x1 + x2  6

x1 - x2  2

x1 - 2x2  1 Where x1, x2  0.

Calculation of Z at the vertices :

Vertex of feasible region Value of Z = 2x + x at the vertex

1 2
(0, 0) 0
(1, 0) 2
(3, 1) 7
(4, 2) 10
(2, 4) 8
(0, 5) 5
There are infinitely many optimal feasible solutions :

One solution is : x1 = 0, x2 = 2

Other solution is : x1 = 2, x2 = 3

& Minimum Z = - 2

In fact, every point on the line - 0.5 x1 + x2 = 2 which lie in the first quadrant is
optimal feasible solution.

Solution of LPPs : Graphical Method : Multiple Solutions

• Solve :

• Maximize Z = 10 x1 + 6x2

• Subject to the constraints :

5 x1 + 3 x2  30

x1 + 2x2  18

Where x1, x2  0.

Calculation of Z at the vertices :

Vertex of feasible region Value of Z = 10 x + 6x at the vertex

1 2
(0, 0) 2
(6,0) 60
(6/7, 60/7) 60
(0, 9) 54

There are infinitely many optimal feasible solutions :

One solution is : x1 = 6, x2 = 0

Other solution is : x1 = 6/7, x2 = 60/7 & Maximum Z = 60

In fact, every point on the line 5 x1 + 3 x2  30

which lie in the first quadrant is optimal feasible solution.

Solution of LPPs : Graphical Method : Unbounded Solutions

• Solve : Maximize Z = 3 x1 + 2x2

• Subject to the constraints :

x1 - x2  1

x1 + x2  3

Where x1, x2  0.

There is an unbounded solution.

Solution of LPPs : Graphical Method :

• Solve :Minimize Z = 4 x1 + 3 x2

• Subject to the constraints :

200x1 + 100x2  4000

x1 + 2 x2  50

x1 + x 2  35 Where x1, x2  0.

Calculation of Z at the vertices :

Vertex of feasible region Value of Z = 4x + 3x at the vertex

1 2
(0, 40) 120
(5, 30) 110
(20, 15) 125
(50, 0) 200

Optimal Feasible Solution is :

x1 = 5 ,x2 = 30 & Minimum Z = 110

Graphical Method : Infeasible Solution :

Solve :

• Maximize Z = 4 x1 + 3 x2

• Subject to the constraints :

-2 x1 + 3 x2  9

3 x1 - 2 x2  - 20

Where x1, x2  0.

• There is no common region in the 1st Quadrant of the graph.

• Hence there is an INFEASIBLE SOLUTION.

Test yourself : Graphical Method :

• Example 1 : Solve :

• Maximize Z = 2 x1 + 5 x2

• Subject to the constraints :

x1 + 4 x2  24

3 x1 + x2  21

x 1 + x2  9 Where x1, x2  0.

Test yourself : Graphical Method :

 • Example 2 : Solve :

• Maximize Z = 4 x1 + 3 x2

• Subject to the constraints :

x1 + 4 x2  2

3 x1 + 12 x2  15

Where x1, x2  0.

Test yourself : Graphical Method :

• Example 1 : Solve :

• Maximize Z = 2 x1 + 5 x2

• Subject to the constraints :

x1 + 4 x2  24, 3 x1 + x2  21, x1 + x2  9

Where x1, x2  0.

• Example 2 : Solve :Maximize Z = 4 x1 + 3 x2

• Subject to the constraints :

x1 + 4 x2  2, 3 x1 + 12 x2  15

Where x1, x2  0.

Solution : Example 1:

Vertex of feasible region Value of Z = 2x + 5x at the vertex

1 2
(0, 0) 0
(7, 0) 14
(6, 3) 27
(4, 5) 33
(0, 6) 30
Solution : Example 2 :

There is no common region.

Hence, there is no Solution.

Limitation of a Graphical method :

It is useful only when the LPP has two variables. If LPP has more than two
variables, this method is not applicable.

Simplex Method

• Used only when :

1. all constraints are of the form  type.

2. All right side numbers in the constraints are Non-negative.

3. Maximize Z = 40x1 + 50x2

Subject to 3x1 + x2 ≤ 9

x1 + 2x2 ≤ 8

Where x1 , x2 ≥ 0

Step 1 :Converting given maximization problem into minimization, we get

Minimize (- Z) = - 40x1 – 50x2

Step 2 :

• Forming a Canonical system by adding slack variables, we get

3x1 + x2 + s1 =9

x1 + 2x2 + s2 = 8

Where x1 , x2 , s1 , s2 ≥ 0

Step : 3 : Table 1 :
Basis X X X s s Ratio x/x
1 2 1 2 2

s 9 3 1 1 0 9
1

s 8 1 2 0 1 4
2

∆ -40 -50 0 0
j

Here leading element is 2.

So, s2 leaves the basis and x2 enters the basis.

Step : 4 : Table 2 :

Basis X X X s s Ratio x/x

1 2 1 2 2

s 5 5/2 0 1 -1/2 2
1

X 4 1/2 1 0 ½ 8
2

∆ - 15 0 0 25
j

Here leading element is 5/2.

So, s1 leaves the basis and x1 enters the basis.

Step 5 : Table 3 :

Basis X X X s s
1 2 1 2

X 2 1 0 2/5 -1/5
1

X 3 0 1 0 3/5
2

∆ 0 0 0 175
j
Here, all ∆j ≥ 0. Hence, the present solution is an optimal feasible solution.

Optimal feasible solution :

X1 = 2, X2 = 3

& Maximum Z = 230

Notes :

• While selecting the ratio, we take the minimum positive ratio. We don’t
consider negative ratio or zero. However, if ratio is 0 and that 0 is obtained
by dividing 0 by a positive number, then we shall select the ratio 0 to decide
the outgoing variables from the basis.

• While selecting the ratio, if all rations are negative ratio, then the solution
is unbounded.

• If there is a tie in ∆j row while selecting a column, any one column can be
selected. However if there is tie between columns of decision variable and
slack variable, a column of decision variable should be selected.

• If there is a tie in ratios while selecting a row, any one row can be selected.
However number of iterations can be reduced by the following calculation.

Example of a tie in the ratio

Basis X X X s s Ratio x/?

1 2 1 2

s 8 1 4 1 0 2
1

s 4 1 2 0 1 2
2

∆ -40 -50 0 0
j
Take selected column and slack variable columns

Basis X s s
2 1 2

s 4 1 0
1

s 2 0 1
2

-50 0 0

Divide slack variable columns by selected column

Basis X s s
2 1 2

s 4 1/4 = ¼ 0/4 = 0
1

s 2 0/2 = 0 1/2 = ½
2

-50

• Here comparing ratios from left to right column-wise, the minimum ratio is
0/2 = 0. It appears in S2 row. So, variable S2 is selected to leave the basis.

Example : 2

• Max Z = 3x1 + 2x2

Subject to

x1 + x2 ≤ 5

x1 + 3x2 ≤ 6

and x1, x2 ≥ 0

Optimal Feasible Solution :

x1 = 5, x2 = 0

& Max Z = 15
Example : 3

MAX Z = 2 x1 + 5 x2 + 7 x3

Subject to 3x1 + 2x2 + 4x3 ≤ 100

x1 + 4x2 + 2x3 ≤ 100

x1 + x2 + x3 ≤ 100

and x1, x2, x3 ≥ 0

Optimal Feasible Solution :

x1 = 0, x2 = 50/3 , x3 = 50/3

Max Z = 200

Practice Sums :

(1) Maximize Z = 2x1 + x2 + 3x3

Subject to 2x1 + x2 + 3x3 ≤ 59, 2x1 + 3x3 ≤ 75, 2x2 +6x3 ≤ 54

and x1,x2, x3 ≥ 0 [Ans: x1 = 51, x2 = 0, x3 = 8]

(2) Maximize Z = x1 + 2x2 + x3 - x4

Subject to x1 + x2 + 3x3 + 4x4 ≤ 60, x2 + 2x3 + 5x4 ≤ 50, 2x1 + 3x2 + 6x4 ≤ 72

and x1,x2, x3, x4 ≥ 0 [Ans: x1 = 0, x2 = 24, x3 = 12 , x4 = 0]

(3) MaxZ = 2x1 + 7x2 + 6x3 + 4x4

Subject to : x1 + x2 + 0.83x3 +0.5x4 ≤ 65, 1.2x1+ x2 + x3 + 1.2x4 ≤ 96,

0.5x1 + 0.7x2 + 1.2x3 +0.4x4 ≤ 80

and x1,x2, x3, x4 ≥ 0 Ans: x1= 0, x2 = 5.1601 x3 = 53.2015 , x4 =31.36

(4) Max Z = 2x1 + 3x2

Subject to : 0.25x1 + 0.5x2 ≤ 40, 0.4x1 + 0.2x2 ≤ 40, 0.8x2 ≤ 40

and x1, x2 ≥ 0 Ans: x1 = 80, x2 = 40

(5) Max Z = 3x1 + 5x2 + 4x3

Subject to : 2x1 + 3x2 ≤ 8, 2x1 + 5x2 ≤ 10, 3x1 + 2x2 + 4x3 ≤ 15

and x1,x2, x3 ≥ 0 Ans: x1 = 0, x2 = 2, x3 = 11/4

Big - M Method :
• Used when at least one constraint has ≥ or = sign with RHS number non-
negative.

• The objective function is revised to :

The original function + artificial variables with coefficient M,

where M is a very big positive number.

• Minimize Z = 8x1 + 6x2

Subject to 3x1 + 8x2 ≤ 96

2x1 + x2 ≥ 10
Where x1 , x2 ≥ 0

Step 1 :

• Forming a Canonical system by adding slack, surplus and artificial variables,

we get

3x1 + 8x2 + S1 = 96

2x1 7+ x2 – S2 + A1 = 10
 Where x1, x2, S1, S2, A1 ≥ 0

Also, the new objective function is :

Minimize Z = 8x1 + 6x2 + M (A1)

= 8x1 + 6x2 + M (10 - 2x1 - x2 + S2)

= 8x1 – 2Mx1 + 6x2 - Mx2 + M S2 + 10 M

= (8 – 2M) x1 + (6 – M) x2 + M S2 + 10 M

Step : 2 : Table 1 :

Basis X X X s s A Ratio
1 2 1 2 1
x/?
s 96 3 8 1 0 0
1

A 10 2 1 0 -1 1
1

∆ 8-2M 6-M 0 M 0
j

Please Note :

• While Observing the last row to select the most negative, if any number is
3−8𝑀 3−8𝑀 3
of the form , then express it as = − 4𝑀 , and then compare
2 2 2
last row elements.
• If artificial variable leaves the basis ,it’s column also leaves from the table.

Step : 2 : Table 1 :

Basis X X X s s A Ratio
1 2 1 2 1
 x/x
1

s 96 3 8 1 0 0 32
1

A 10 2 1 0 -1 1 5
1

∆ 8-2M 6-M 0 M 0
j

Step : 3 : Table 2 :

Basis X X X s s Ratio x/
1 2 1 2

s 81 0 13/2 1 3/2
1

x 5 1 1/2 0 -1/2
1

∆ 0 2 0 4
j

Optimal feasible solution is :

x1 = 5, x2 = 0

Minimum Z = 40

Example 2 : Minimize Z = x1 + x2 Subject to 2x1 + x2 ≥ 4, x1 + 7x2 ≥ 7, and x1, x2 ≥ 0

Solution :

x1 = 21/13 , x2 = 10/13

Minimum Z = 31/13

Big - M Method : Example 3 :

Maximize Z = 6x1 + 4x2

Subject to

2x1 + 3x2 ≤ 30
 3x1 + 2x2 ≤ 24

x1 + x2 ≥ 3 and x1, x2 ≥ 0.

Hint : First convert into minimization, then proceed.

Solution :

x1 = 8 , x2 = 0

Maximum Z = 48

• Note : Look at the last table of the solution of the previous problem. Here,
we reached the optimal solution, but at the same time, we had in the last
row and x2 column, we had ‘zero’, but x2 not being in the basis. This
suggests that there are multiple solutions to this LPP. Next solution can be
obtained by selecting x2 column and proceeding from the last table.

Example 4 :

• Maximize Z = 3x1 + 4x2

Subject to

2x1 + x2 ≤ 600

x1 + x2 ≤ 225

5x1 + 4x2 ≤ 1000

x1 + 2x2 ≥ 150 and x1, x2 ≥ 0.

Solution :

x1 = 75 , x2 = 150

Maximum Z = 825

Two Phase Method :

 • Used when at least one constraint has ≥ or = sign with RHS number non-
negative.

• The objective function is revised to :

Sum of Artificial variables with coefficient 1.

Example 1 :

Minimize Z = x1 + x2

Subject to 2x1 + x2 ≥ 4

x1 + 7x2 ≥ 7

and x1, x2 ≥ 0

Step 1 :

• Forming a Canonical system by adding slack, surplus and artificial variables,

we get

2 x1 + x2 - S1 + A1 =4

x1 + 7x2 – S2 + A2 =7

Where x1, x2, S1, S2, A1, A1 ≥ 0

Also, the new objective function is :

Minimize Z = A1

= 10 - 2x1 - x2 + S2

• Minimize Z = 8x1 + 6x2

Subject to 3x1 + 8x2 ≤ 96

 2x1 + x2 ≥ 10
Where x1 , x2 ≥ 0

Step 1 :

• Forming a Canonical system by adding slack, surplus and artificial variables,

we get

3x1 + 8x2 + S1 = 96

2x1 + x2 – S2 + A1 = 10

Where x1, x2, S1, S2, A1 ≥ 0

Also, the new objective function is :

Minimize Z = A1

= 10 - 2x1 - x2 + S2

Step : 2 : Phase 1 : Table 1 :

Basis X X X s s A Ratio
1 2 1 2 1
x/?
s 96 3 8 1 0 0 32
1

A 10 2 1 0 -1 1 5
1
’
-2 -1 0 1 0
∆
j
 ∆ 8 6 0 0 0
j

Please Note :

• If artificial variable leaves the basis, it’s column also leaves from the table.

Step : 2 : Phase 1 : Table 2 :

Basis X X X s s Ratio
1 2 1 2

s 81 0 13/2 1 3/2
1

x 5 1 1/2 0 -1/2
1
’
0 0 0 0
∆
j

∆ 0 2 0 4
j

Here all ∆j’ ≥ 0. So, Phase I is over.

Step : 3 : Phase 2 : Table 1 :

Basis X X X s s Ratio
1 2 1 2

s 81 0 13/2 1 3/2
1

x 5 1 1/2 0 -1/2
1

∆ 0 2 0 4
j

Here all ∆j ≥ 0. So, Phase II is over.

Optimal feasible solution is :

x1 = 5, x2 = 0

Minimum Z = 40

Duality

in

Linear Programming Problems

Standard Form :

• Remember : Convert the primal problem into standard form (Maximization

with  or Minimization with ≥) before finding the dual.

• Here, it is not necessary that RHS of the constraint must be a non-negative

number.

Duality in Linear Programming Problems

• Example 1 : Maximize z = 2x1 + x2,

subject to: subject to: 4 x1 + 5 x2 ≤ 4, 3 x1 − 7 x2 ≤ 2, x1 ≥ 0, x2 ≥ 0.

• Dual : Minimize z* = 4y1 + 2y2,

subject to: 4 y1 + 3 y2 ≥ 2, 5 y1 − 7 y2 ≥ 1, y1 ≥ 0, y2 ≥ 0.

Example 2 : Find the dual of the given LPP :

• Maximize Z = 6x1 + 10x2

• Subject to constraints,

2x1 + 8x2 ≤ 60 .......................(i)

3x1 + 5x2 ≤ 45 .......................(ii)
5x1 - 6x2 ≤ 10 .......................(iii)
0 x1 + x2 ≤ 40 .......................(iv)
where x1, x2 ≥ 0

Solution :

• Minimize W = 60y1 + 45y2 + 10y3 + 40y4

• Subject to constraints,

2y1+ 3y2+ 5y3+ 0 y4 ≥ 6

8y1+ 5y2 - 6y3 + y4 ≥ 10
where y1, y2, y3, y4 ≥ 0

Example 3 :

• Minimize Z = 6x1– 4x2+ 4x3

• Subject to constraints,

6x1– 10x2 + 4x3 ≥ 14 ..................(i)

6x1+ 2x2 + 6x3 ≥ 10 ..................(ii)
7x1– 2x2 + 5x3 ≤ 20 ..................(iii)
x1– 4x2 + 5x3 ≥ 3 ..................(iv)
4x1+ 7x2– 4x3 ≥ 20 ..................(v)
where x1, x2, x3≥ 0

• Minimize Z = 6x1– 4x2+ 4x3

• Subject to constraints,

6x1– 10x2 + 4x3 ≥ 14 ..................(i)

6x1+ 2x2 + 6x3 ≥ 10 ..................(ii)
7x1– 2x2 + 5x3 ≤ 20 ..................(iii)
x1– 4x2 + 5x3 ≥ 3 ..................(iv)
4x1+ 7x2– 4x3 ≥ 20 ..................(v)
where x1, x2, x3≥ 0

Example 3 : Standard Form :

 • Minimize Z = 6x1– 4x2+ 4x3

• Subject to constraints,

6x1 – 10x2 + 4x3 ≥ 14 ..................(i)

6x1 + 2x2 + 6x3 ≥ 10 ..................(ii)
-7x1 + 2x2 - 5x3 ≥ -20 ..................(iii)
x1 – 4x2 + 5x3 ≥ 3 ..................(iv)
4x1 + 7x2– 4x3 ≥ 20 ..................(v)
where x1, x2, x3 ≥ 0

Solution :

• Maximize W = 14y1+10y2+20y3+3y4+20y5

• Subject to constraints,

6y1+ 6y2 – 7y3+ y4 + 4y5 ≤ 6

10y1+ 2y2 + 2y3 – 4y4 +7y5 ≤ – 4
4y1+ 6y2– 5y3 + 5y4 – 4y5 ≤ 4
where y1, y2, y3, y4 and y5≥ 0

Duality in Linear Programming Problems

• Example 1 : Maximize z = 2x1 + x2,

subject to: 1 x1 + 3 x2 ≤ 4,

1 x1 − 1 x2 = 2,

x1 ≥ 0, x2 ≥ 0.

• Dual : Minimize z* = 4y1 + 2y2,

subject to: 1 y1 + 1 y2 ≥ 2,

3 y1 − 1 y2 ≥ 1,

y1 ≥ 0, y2 is unrestricted.
 • Example 1 : Maximize z = 2x1 + x2,

subject to: 1 x1 + 1 x2 ≤ 4,

3 x1 − 1 x2 ≤ 2,

x1 ≥ 0, x2 unrestricted.

• Dual : Minimize z* = 4y1 + 2y2,

subject to: 1 y1 + 3 y2 ≥ 2,

1 y1 − 1 y2 = 1,

y1 ≥ 0, y2 ≥ 0.

Advantage of Duals :

• Minimize Z = 6x1+ 4x2+ 7x3

• Subject to constraints,

6x1 – 10x2 + 4x3 ≥ 14 ..................(i)

6x1 + 2x2 + 6x3 ≥ 10 ..................(ii)
-7x1 + 2x2 - 5x3 ≥ -20 ..................(iii)
x1 – 4x2 + 5x3 ≥ 3 ..................(iv)
4x1 + 7x2– 4x3 ≥ 20 ..................(v)
where x1, x2, x3 ≥ 0

Solving by Big – M :

• Given variables : 5

• Surplus variables : 5

• Artificial variables : 5

• Total number of variables : 15

• & complications of Big-M !!!!!!!!!

Solution : The dual is :

• Maximize W = 14y1+10y2+20y3+3y4+20y5

• Subject to constraints,

6y1+ 6y2 – 7y3+ y4 + 4y5 ≤ 6

10y1+ 2y2 + 2y3 – 4y4 +7y5 ≤ 4
4y1+ 6y2– 5y3 + 5y4 – 4y5 ≤ 7
where y1, y2, y3, y4 and y5≥ 0

• Total number of variables (including slack) : 8

Dual Simplex Method :

• Maximize Z = - 3 x1 – 2x2

• Subject to constraints,

x1 + x2 ≥ 1
x1 + x2 7

x1 + 2x2 ≥ 10

x2  3, where x1, x2 ≥ 0

Step 1 :

• Converting given maximization problem into minimization, we get

Minimize (- Z) = 3 x1 + 2 x2

Step 2 :

Converting all constraints into ‘’ type, we get

- x1 - x2  - 1
x1 + x2  7

- x1 - 2x2  - 10
 x2  3, where x1, x2 ≥ 0

Note : Negative numbers in RHS allowed here.

Step 3 :

• Forming a Canonical system by adding slack variables, we get

- x1 - x2 + s1 = - 01
x1 + x2 + s2 = 07

- x1 - 2x2 + s3 = - 10

x2 + s4 = 03,

Where x1 , x2 , s1 , s2 , s3 , s4 ≥ 0

Step 4 : Table : 1

Basis X X X s s s s
1 2 1 2 3 4

s -1 -1 -1 1 0 0 0
1
s 7 1 1 0 1 0 0
2
s -10 -1 -2 0 0 1 0
3
s 3 0 1 0 0 0 1
4
∆ 3 2 0 0 0 0
j

Step 4 : Table : 1

Basis X X X s s s s
1 2 1 2 3 4

s -1 -1 -1 1 0 0 0
1
s 7 1 1 0 1 0 0
2
s -10 -1 -2 0 0 1 0
3
s 3 0 1 0 0 0 1
4
∆ 3 2 0 0 0 0
j
Ratio -3 -1 - - 0 -
∆ /s
j 3

Step 5 : Table : 2

Basis X X X s s s s
1 2 1 2 3 4

s 4 -1/2 0 1 0 -1/2 0
1
s 2 1/2 0 0 1 1/2 0
2
X 5 1/2 1 0 0 -1/2 0
2

s -2 -1/2 0 0 0 1/2 1
4
∆ 2 0 0 0 1 0
j
Ratio

Step 6 : Table 3 :

Basis X X X s s s s
1 2 1 2 3 4

s 6 0 0 1 0 -1 -1
1
s 0 0 0 0 1 1 1
2
X 3 0 1 0 0 0 1
2

X 4 1 0 0 0 -1 -2
1

∆ 0 0 0 0 3 4
j

• Here, all X ≥ 0. Hence, the present solution is an optimal feasible solution. It

is given by :

X1 = 4, X2 = 3

& Maximum Z = - 18