Unit 2 Optimization

Unit 2.2 Linear programming II: Simplex Method

Simplex method, Solution to maximization problems, solution to minimization problems, Big-M method, some
special cases in linear programming

Algorithm of General Form into Standard Form:

1. Check whether the objective function of LPP is maximized or minimized. If it is to be minimized then we convert
'
it into a problem of maximization by Max 𝑍 = - Min Z

RI
2. Check all the decision variables are ≥ 0 if any decision variables are unrestricted.

PA
2𝑥1 + 𝑥2 ≤ 4; 𝑥1≥0 and 𝑥2 is unrestricted.

( ' ''
)
2𝑥1 + 𝑥2 − 𝑥2 ≤ 4; 𝑥1, 𝑥2, 𝑥2 ≥ 0
' ''

A
3. Express the problem in standard form by introducing slack or surplus variable to convert the inequality constraints

K
into equation. When,

≤ Add slack variable

SH
≥ Subtract surplus variable

= Nothing
A

4. Check whether all 𝑏𝑖(𝑖 = 1, 2, ……, 𝑛) are positive or not. If any𝑏𝑖 is negative then multiply the in equation of
constraints by – 1.
K

Algorithm of Simplex Method:

A

1. Convert the given general LPP into standard LPP.

PR

(i) Objective function of LPP must be maximized. If it is to be minimized then we convert it into a problem of
'
maximization by Maximize 𝑍 = Maximize Z

(ii) Check all the decision variable is greater than zero.

Y

(iii) Express the problem in standard form by introducing slack or surplus variable to convert the inequality
JA

constraints into equation.

(iv) All the values of right hand side must be positive.

2. Write the values of initial basic feasible solution.

3. Write the standard form LPP into matrix form.

4. Construct the initial Simplex table.

5. Calculate the values of 𝑍𝑗 − 𝐶𝑗 and check the basic feasible solution for optimality.
 𝑍𝑗 − 𝐶𝑗 = 𝐶𝑗 × 𝐵𝑥 − 𝐶𝑗

(i) If all 𝑍𝑗 − 𝐶𝑗≥0, the optimal solution will obtained.

(ii) If at least one 𝑍𝑗 − 𝐶𝑗 is negative value then indicate by an arrow and this column is called key-column.

(iii) If more than one 𝑍𝑗 − 𝐶𝑗 is negative value then choose the most negative value of them and this column is
called key-column.

6. Calculate the minimum ratio.

RI
𝑆𝑜𝑙𝑢𝑡𝑖𝑜𝑛
Minimum ratio = 𝐶𝑘
, 𝐶𝑘= key-column > 0 but less than or equal to zero, then not find ratio.

PA
7. Construct the new Simplex table by entering incoming vector.

8. Repeat the step 5 and 6.

A
Artificial variable:

K
Example: Use the Simplex Method to solve the following LP problem.

Maximize Z = 3x1 + 5x2 + 4x3

SH
Subject to the constraints:

2x1 +3 x2 ≤ 8
A

2𝑥2 + 5𝑥3 ≤ 10
K

3x1 + 2x2 + 4x3 ≤ 15

A

𝑥1, 𝑥2, 𝑥3 ≥ 0
PR

General form of LP problem change in standard form of LPP, we have

Maximize Z = 3x1 + 5x2 + 4x3 + 0 s1 + 0 s2 + 0 s3

Subject to the constraints:

Y

2x1 + 3x2 +0x3 + s1 + 0s2 + 0s3 = 8

JA

0𝑥1 + 2𝑥2 + 5x3 + 0s1 + s2 + 0s3 = 10

3x1 + 2x2 + 4x3 + 0s1 + 0s2 + s3 = 15

𝑥1, 𝑥2, 𝑥3 ≥ 0
 Where, s1, s2 and s3 are slack variable (unused resources) with cost zero.

Simplex table I

Basic 𝐶𝑗 3 5 4 0 0 0 Solution Ratio

variable
coefficient
Basic variable 𝑥1 𝑥2 𝑥3 𝑠1 𝑠2 𝑠3
𝐶𝑗
Bx
0 𝑠1 2 3 0 1 0 0 8 8/3

RI
0 𝑠2 0 2 5 0 1 0 10 5
0 𝑠3 3 2 4 0 0 1 15 15/2

PA
𝑧𝑗 0 0 0 0 0 0
𝐶𝑗 − 𝑍𝑗 3 5 4 0 0 0

A
Where, for Maximize 𝐶𝑗 − 𝑍𝑗 ≤ 0 and Minimize 𝐶𝑗 − 𝑍𝑗 ≥ 0, key column, key row and key element, x3 is entering

K
variable and s1 is leaving variable.

𝑘𝑒𝑦 𝑟𝑜𝑤 𝑒𝑙𝑒𝑚𝑒𝑛𝑡 2 3 0 1 0 0 8

New R1 element = 𝑘𝑒𝑦 𝑒𝑙𝑒𝑚𝑒𝑛𝑡
= 3
, 3
, 3
, 3
, 3
, 3
, 3
= 2/3, 1, 0, 1/3, 0, 0, 8/3
SH
Old R2 element – key column element × New R1 Old R3 element – key column element × New R1
0 – 2 × 2/3 = -4/3 3 – 2 × 2/3 = 5/3
2–2×1=0 2–2×1=0
A

5–2×0=5 4–2×0=4
0 – 2 × 1/3 = -2/3 0 – 2× 1/3 = - 2/3
1–2×0=1 0 – 2× 0 = 0
K

0–2×0=0 1 – 2× 0 = 1
10 – 2 × 8/3= 14/3 15 – 2× 8/3 = 29/3
A

Iteration I
PR

Basic 𝐶𝑗 3 5 4 0 0 0 Solution Min Ratio

variable
Basic variable 𝑥1 𝑥2 𝑥3 𝑠1 𝑠2 𝑠3
coefficient
B
CB
5 𝑥2 2/3 1 0 1/3 0 0 8/3 ---
Y

0 𝑠2 -4/3 0 5 -2/3 1 0 14/3 14/15

JA

0 𝑠3 5/3 0 4 -2/3 0 1 29/3 29/12

𝑧𝑗 10/3 5 0 5/3 0 0

𝐶𝑗 − 𝑍𝑗 -1/3 0 4 -5/3 0 0

−4 −2 14
𝑘𝑒𝑦 𝑟𝑜𝑤 𝑒𝑙𝑒𝑚𝑒𝑛𝑡 0 5 1 0
New R2 element = 𝑘𝑒𝑦 𝑒𝑙𝑒𝑚𝑒𝑛𝑡
= 5
3
, 5
, 5
, 3
5
, 5
, 5
, 3
5
= -4/15, 0, 1, -2/15, 1/5, 0, 14/15

Old R1 element – key column element × New R2 Old R3 element – key column element × New R2
2/3 –0 × (-4/15) = 2/3 5/3 – 4× (-4/5) = 41/15
 1 – 0×0 = 1 0–4×0=0
0 – 0×1 = 0 4–4×1=0
1/3 – 0×(-2/15) = 1/3 -2/3– 4× (-2/15) = - 2/15
0 – 0×1/5 = 0 0 – 4× 1/5 = -4/5
0 – 0×0 = 0 1 – 4× 0 = 1
8/3 – 0×14/15 = 8/3 29/3 – 4× 14/15 = 89/15
Iteration II

Basic 𝐶𝑗 3 5 4 0 0 0 Solution Min Ratio

variable
coefficient Basic variable 𝑥1 𝑥2 𝑥3 𝑠1 𝑠2 𝑠3

RI
𝐶𝑗 B
5 𝑥2 2/3 1 0 1/3 0 0 8/3 4
4 𝑥3 -4/15 0 1 - 2/15 1/5 0 14/15 -2/7

PA
0 𝑠3 41/15 0 0 -2/15 -4/5 1 89/15 89/41
𝑧𝑗 34/15 5 4 17/15 4/5 0
11/15 0 0 -17/15

A
𝐶𝑗 − 𝑍𝑗 -4/5 0

K
𝑘𝑒𝑦 𝑟𝑜𝑤 𝑒𝑙𝑒𝑚𝑒𝑛𝑡 41 15 15 15 −2 15 −4 15 15 89 15
New R3 element = 𝑘𝑒𝑦 𝑒𝑙𝑒𝑚𝑒𝑛𝑡
= 15
× 41
, 0× 41
, 0× 41
, 15
× 41
, 5
× 41
, 1× 41
, 15
× 41
SH
= 1, 0, 0, -2/41, -12/41, 15/41, 89/41

Old R1 element – key column element × New R3 Old R2 element – key column element × New R3
2/3 –2/3 × 1= 0 -4/15 + 4/15× 1 = 0
A

1 – 2/3×0 = 1 0 + 4/15 × 0 = 0
0 – 2/3×0 = 0 1 + 4/15 × 0 = 1
1/3 – 2/3×(-2/41) = 15/41 -2/15 + 4/15× (-2/41) = - 6/41
K

0 – 2/3×(-12/41) = 8/41 1/5 + 4/15× (-12/41) = 5/41

0 – 2/3×15/41 = -10/41 0 + 4/15× 15/41 = 4/41
A

8/3 – 2/3×89/41 = 50/41 14/15 + 4/15× 89/41 = 62/41

Iteration III
PR

Basic 𝐶𝑗 3 5 4 0 0 0 Solution Min Ratio

variable
Basic variable 𝑥1 𝑥2 𝑥3 𝑠1 𝑠2 𝑠3
coefficient
B
𝐶𝑗
Y

5 𝑥2 0 1 0 15/41 8/41 -10/41 50/41

4 𝑥3 0 0 1 - 6/41 5/41 4/41 62/41
JA

3 𝑥1 1 0 0 -2/41 -12/41 15/41 89/15

𝑧𝑗 3 5 4 45/41 24/41 11/41
𝐶𝑗 − 𝑍𝑗 0 0 0 -45/41 -24/41 -11/41

In maximize 𝐶𝑗 − 𝑍𝑗 ≤ 0 follows then closed after, then solution is x1 = 89/41, x2 = 50/41, x3 = 62/41, then
maximum value of Z = 3 × 89/41 + 5×50/41+4×62/41 = 765/41.

Example: Solve the following LPP using Simplex method.

 Minimize Z = 2x1 – 3x2 + 6x3

Subject to:

3x1 – x2 + 2x3 ≤ 7

2x1 – 4x2 ≥ - 12

- 4x1 + 3x2 + 8x3 ≤ 10

𝑥1, 𝑥2, 𝑥3 ≥ 0

RI
Example 2: A company makes two kinds of leather belts, belt A and belt B. Belt A is a high quality belt and belt B is
of lower quality. The respective profits are Rs 4 and Rs 3 per belt. The production of each of type A requires twice

PA
as much time as a belt of type B, and if all belts were of type B, the company could make 1,000 belts per day. The
supply of leather is sufficient for only 800 belts per day (both A and B combined). Belt A requires a fancy buckle
and only 400 of these are available per day. There are only 700 buckles a day available for belt B.

A
What should be the daily production of each type of belt? Formulate this problem as an LP model and solve it using
the simplex method.

K
Solution:

Let x1 and x2 be the number of belts of type A and B, respectively, manufactured each day. Then the LP
SH
model would be as follows:

Maximize Z = 4x1 + 3x2

A

Subject to the constraints

K

Time availability: 2x1 + x2 ≤ 1,000

Supply of leather: x1 + x2 ≤ 800

A

Buckles availability x1 ≤ 400 and x2 ≤ 700

PR

And x1, x2 ≥ 0

Example 3: Solve the following LPP using Simplex method.

Y

Maximize Z = 16x1 + 17x2 + 10x3

Subject to:
JA

𝑥1 + 𝑥2 + 4𝑥3 ≤ 2000

2x1 + x2 + x3 ≤ 3600

𝑥1 + 2𝑥2 + 2𝑥3 ≤ 2400

𝑥1 ≤ 30
 𝑥1, 𝑥2, 𝑥3 ≥ 0

Big M Method Maximization Problem:

In Big M Method to convert to standard form

≤ Add slack variable

≥ Subtract surplus variable and add an artificial variable

= Add an artificial variable

RI
Artificial variable:

It is a fictitious and cannot have any physical meaning. It is merely a device to get the starting basic feasible

PA
solution, so that the Simplex procedure may be adopted as usual until the optimal solution is obtained. Cost of
artificial variable is – M. M is very-very large value and – M is very-very small value, i.e. – 10M is less than – 5M

A
Steps for Big M Method:

Solve the modified LPP by Simplex method until the any one of the three cases may arise.

K
(i) If no artificial variable appear in the basis and the optimality conditions are satisfied, then the current solution is
an optimal basic feasible solution.
SH
(ii) If at least one artificial variable is there in the basis at zero level and the optimality conditions is satisfied, then
the current solution is an optimal basic feasible solution.
A

(iii) If at least one artificial variable is appear in basis at positive level and the optimality conditions is satisfied, then
the original problem has no feasible solution.
K

Example: Charne’s Big M Method or Penalty Method to solve the following LP Problem.
A

Max Z = 𝑥1 − 𝑥2 + 3𝑥3
PR

Subject to:

𝑥1 + 𝑥2 ≤ 20

𝑥1 + 𝑥3 = 5
Y
JA

𝑥2 + 𝑥3 ≥ 20

𝑥1, 𝑥2, 𝑥3 ≥ 0

Solution: Given,
JA
Y
PR
A
K
A
SH
K
A
PA
RI