CEV623_Chapter 7_Optimization

Case 1: Simplex Method for Standard Maximization Problem

Example 1 (Chocolate production)

PRODUCT
Material
REGULAR (x) PREMIUM (y) Capacity
Choco 1 kg 2 kg 12 kg
Sugar 3 kg 2 kg 24 kg
Profit (P) 10/kg 12/kg

Maximize profit, P = 10x + 12y subject to the constraints:

x + 2y ≤ 12
3x + 2y ≤ 24
x ≥ 0, y ≥ 0

Step 1:

Convert the problem to standard form

 Objective function to maximized

 All variables are non-negative
 Each linear constraint is in the form of ≤ to a non-negative constant

Step 2:
Convert to a system of equation by introducing slack variables to turn the constraints into equations,
and rewriting the objective function in standard form.

Rewriting the objective function: P – 10x – 12y = 0

Introducing slack variables to the constraints:

x + 2y + s1= 12
3x + 2y + s2 = 24
x ≥ 0, y ≥ 0, s1 ≥ 0, s2 ≥ 0

Step 3:
Write down the initial table:

P x y s1 s2 Solution Intercept
P 1 -10 -12 0 0 0
s1 0 1 2 1 0 12
s2 0 3 2 0 1 24
Step 4:
Select the pivot column: Choose the negative number with the largest magnitude in the top row
(excluding the rightmost (solution) entry). This column becomes the pivot column.
If there are 2 candidates, choose either one.
If all the numbers in the top are zero or positive (excluding the rightmost entry), then this is the
maximum objective function.

pivot
column
P x y s1 s2 Solution Intercept
P 1 -10 -12 0 0 0
s1 0 1 2 1 0 12
s2 0 3 2 0 1 24

Step 5:
Calculate the intercept column by dividing the solution column to the pivot column that has been
chosen. Do this to only those numbers that are non-negative in the solution column.

pivot
column
P x y s1 s2 Solution Intercept
P 1 -10 -12 0 0 0
s1 0 1 2 1 0 12 12/2 = 6
s2 0 3 2 0 1 24 24/2 = 12

Step 6:
Select the pivot row. This is the row that has the smallest positive number on the intercept column.

entering
variable
P x y s1 s2 Solution Intercept
P 1 -10 -12 0 0 0
leaving variable s1 0 1 2 1 0 12 6 Pivot row
s2 0 3 2 0 1 24 12

Then, the variable on top of the column is the entering variable and the variable on the left of the
pivot row is the leaving variable.

pivot
column
P x y s1 s2 Solution Intercept
row P P 1 -10 -12 0 0 0
row y y 0 1 2 1 0 12 6 Pivot row
row s2 s2 0 3 2 0 1 24 12
Use the pivot to clear the column in the normal manner (just like Gauss Elimination) and then relabel
the variable in the pivot row with the entering variable.

P x y s1 s2 Solution Intercept
row P' = row P - (-12)row y' P 1 -4 0 6 0 72
row y' = Row y/2 y 0 1/2 1 1/2 0 6
row s2' = Row s2 - (2)Row y' s2 0 2 0 -1 1 12

Step 7:
If all of the values in the first row are zero or positive, then stop. You have found the solution.
If this is not the case, you have to repeat steps 4 to 6 until all values in the first row or positive.

pivot
column
P x y s1 s2 Solution Intercept
P 1 -4 0 6 0 72
y 0 1/2 1 1/2 0 6
s2 0 2 0 -1 1 12
The pivot column is x.

pivot
column
P x y s1 s2 Solution Intercept
P 1 -4 0 6 0 72
y 0 1/2 1 1/2 0 6 6/(1/2) = 12
s2 0 2 0 -1 1 12 12/2 = 6
The pivot row is s2 (as is the smallest positive number in intercept column)
Entering variable = x and leaving variable is s2. Clearing the pivot.

entering
variable
P x y s1 s2 Solution Intercept
P 1 -4 0 6 0 72
y 0 1/2 1 1/2 0 6 12
pivot
s2 0 2 0 -1 1 12
leaving variable 6 row

pivot
column
P x y s1 s2 Solution Intercept
row P' P 1 -4 0 6 0 72
row y' y 0 1/2 1 1/2 0 6 12
pivot
x 0 2 0 -1 1 12
row x' 6 row
 P x y s1 s2 Solution Intercept
row P'' = row P' - (-4)row x'' P 1 0 0 4 2 96
row y''= Row y' - (1/2)row x'' y 0 0 1 3/4 -1/4 3
row x'' = row x'/2 x 0 1 0 - 1/2 1/2 6

P x y s1 s2 Solution Intercept
P 1 0 0 4 2 96
y 0 0 1 3/4 -1/4 3
x 0 1 0 - 1/2 1/2 6

All of the values in the first row are non-negative. Hence, the maximum objective has been reached.

The maximum value of P = 96

Values of x and y that give maximum P are x = 6, y = 3.
Case 2: Simplex Method for Standard Minimization Problem

Example 2: Production of Sweet Bread

PRODUCT
Material CAPACITY
PREMIUM (x) REGULAR (y)
Flour 60 kg/kg 60 kg/kg 300 kg
egg 12 kg/kg 6 kg/kg 36 kg
sugar 10 kg/kg 30 kg/kg 90 kg
Cost 0.12 0.15

Minimize production cost, P = 0.12x + 0.15y subject to the constraints:

60x + 60y ≥ 300

12x + 6y ≥ 36
10x + 30y ≥ 90
x ≥ 0, y ≥ 0

Step 1:

Convert the problem to standard form

 Objective function to minimized.

 All variables are non-negative
 Each linear constraint is in the form of ≥ to a non-negative constant

 The basic procedure used to solve the standard minimization problem is to convert it to a
standard maximization problem and then solved it using the techniques for standard
minimization problem.
 The transformed maximization problem from the minimization problem is called dual
maximization problem

Step 2:
First put the equations into a matrix form where the objective function coefficients is placed at the
bottom:
60 60 300 The first inequality constants
12 6 36 The second inequality constants
10 30 90 The third inequality constants
0.12 0.15 Objective function constants

Then, transpose this matrix

60 12 10 0.12 This becomes the new first inequality constants
60 6 30 0.15 This becomes the new second inequality constants
300 36 90 This becomes the objective function constants
Rewrite the new problem as maximization problem using the new coefficients above (call the dual
maximization problem) using new variables (u, v, w):

Maximize z = 300u + 36v +90w (called dual objective function)

Subject to the constraints (called dual constraints):

60u + 12v +10w ≤ 0.12
60u + 6v +30w ≤ 0.15
u ≥ 0, v ≥ 0, w ≥ 0

Solving the above dual maximization problem is equivalent to solving the standard minimization
problem.

Step 3:

Step 4:
Tabulated and solve the problem.

z u v w s1 s2 Solution Intercept
z 1 -300 -36 -90 0 0 0
s1 0 60 12 10 1 0 0.12 0.002
s2 0 60 6 30 0 1 0.15 0.0025

z u v w s1 s2 Solution Intercept
z 1 0 24 -40 5 0 0.6
u 0 1 0.2 0.166667 0.016667 0 0.002 0.012
s2 0 0 -6 20 -1 1 0.03 0.0015

z u v w s1 s2 Solution Intercept
z 1 0 12 0 3 2 0.66
w 0 1 0 0 0 0 0
s2 0 0 -0.3 1 -0.05 0.05 0.0015

Thus, the solution of the dual maximization problem is z = 0.66 which is also the minimize value for
P, in example minimum value of P is 0.66.
The values of x and y that give the minimum value of P are retrieved from the top row just below the
two slack variables which are x = 3 and y = 2.