Chapter

LINEAR
PROGRAMMING
Simplex Method

Simplex Maximization Problems 4.1

Simplex Minimization Problems 4.2

Objectives:
At the end of the chapter, the students should be able to:

Convert the LP model to standard form

Define and differentiate slack, surplus and artificial variables in
LP models
Apply each step in solving LPP by simplex method
Compare the graphic method and simplex method
Appreciate the use of algebra in solving LP problems by simplex
method
Recognize infeasibility and unboundedness in simplex method
of solving LPP

69
 SIMPLEX METHOD

The graphic method of solving linear programming problems (LPP)

becomes impractical when the problem involves more than two variables ; it is
being used if the problem is limited only to two variables. The useful method
applied for LPP involving two or more variables is the simplex method or tubular
method.

Definition
The simplex method of solving linear programming problems is a set
of mathematical steps (with the use of tables) that provides a direct and
economical way of solving linear programming problems.

4.1 Simplex Maximization

Steps in Solving Maximization Problems
Set up the objective function and the constraints of the problem.

Convert the objective function and the explicit constraints to standard

form.
Follow the following rules in converting the LP model to standard form:

Symbol in the To put into Standard Form

constraint
≤ Add a slack variable (Si) that represents the amount
by which the right-hand side of the constraint exceeds
its left-hand side.
= Add an artificial variable (Ai) or extraneous variable,
which serves as a starting basic variable. This variable
has no physical meaning in the problem hence it is
called “artificial “.
≥ Subtract a surplus variable (Si) which represents the
excess amount of the left-hand side of the constraint
over the right-hand side, and add an artificial variable
(Ai )
The artificial variables in the optimum or final solution should be zero provided
a feasible solution exists. The reasonable way to do this is to penalize these

70
variables in the objective function by using, a very small number coefficient,
M, (the letter M means million), or 100 if the numerical coefficients found in
the constraints and objective function are all less than or equal to 10, let it
equal to – 1000 if the numerical coefficients are less than or equal to 100, let it
equal to – 10,000, if the numerical coefficients are less than or equal to 1000,
and so on.
Set up the initial tableau.
The number of rows = no. of constraints + 4
The number of columns = no. of decision variables + no. of slack/surplus
variables + no. of artificial variables + 4

Fill up the first and second rows of the initial tableau.

The variable row (second row) contains the variables of the

objective function. List the variables in order, beginning with the
decision variables, followed by the slack variables, and then the
artificial variables (follow the subscripts of these variables).
The C1, row (first row) takes the numerical coefficients of the
variables in the variable row.

Put the entries in the first three columns of the initial table.

The second column is the solution column (also called basis or

variable column). Initially, the variables to be listed here are the
artificial variables of the constraints because the variables in this
column should be replaced by the decision variables or slack
variables as the final solution is obtained. The artificial variables
should be forced to zero when the optimum solution is attained
since it has no meaning in the perspective of the original problem;
it is just use as the starting basic variable in the solution. If the
constraints do not contain the artificial variables, use the slack
variables.
The first column is the contribution to profit column (C1), which
takes the coefficients of the variables in the solution column.
The third column is the quantity column or the constant column,
which takes the constant terms or the right-hand values of the
constraints.

Fill up the succeeding rows with the corresponding constant coefficients of

the constraints.

Determine the second to the last row, Zj row, by summing up the product
of C1 by each of the coefficient in the constraints including the constant
terms. (The variable Z is commonly used for the objective variable).

7
1
 Determine the last row, (C1 – Zj) row.

The last column, Q1, initially has zero entries.

Set up the next tableau (iteration 1). The entries are the same as in the initial
table except for the column Q1 column.

Determine the optimum column, column with the most positive value (zero
is not included) in the C1- Z1 row of the initial table. Highlight this column.
The variable in this column is called the entering variable (variable which
contributes the most value to the objective function).

Change the Q1 column by dividing each entry in the quantity column by

the corresponding non-zero and non-negative entry in the optimum
column. No value should appear if the divisor is zero or negative.

Determine the incoming row or pivotal row by choosing the smallest value
in the Q1, column. Highlight this row. The variable in this row, which is to
be replaced by the entering variable is called the leaving variable. This
variable creates the bottleneck to the optimal solution. In case that there
is no possible value to select in the Q1 column, then there will be no
pivotal row and the solution is final. If this case arises, the problem is said
to be unbounded.

Set up the third table (Iteration 2)

Identify the pivot, intersection of the optimum column and the pivotal row
in iteration 1.

Make the value of pivot equal to one (1) by dividing all entries in the pivotal
row of iteration 1 (now the replacing row) by the pivot. Do this operation
from quantity column to the last column before the Q1 column. Replace
the entry in the solution column by the entering variable and quantity
column by the corresponding constant coefficient of the entering variable.

Compute the new entries for the remaining rows by reducing the optimum
column entries to zero (entries in the constraint rows only) by using
algebraic manipulations.

New row = row to be replaced – intersection of row with pivot x new pivot row

d. Solve the new Z1 and ( C1 – Z1 ) rows.

72
 If the (C1 – Z1) has still positive entry, go back to step 4 and step 5 (for
iteration 3,4…).

The optimum or final solution is obtained if the entries in the (C1 – Z1) row are
all zeros or /and negative values. At this point, if the obtained Z value is negative
and one or more of the basic variables (found in the solution column) are
artificial, then the solution is said to be infeasible.

4.1.1 Writing LP Model to Equation Form

The following examples will illustrate how to write the objective function
and its constraints to a standard form.

Examples: Write the following LP models to standard form

1. Maximize Z = 5x + 12y 2. Maximize Z = 120x + 80y

Subject to 2x + 3y ≤ 80 Subject to 2x + y ≤ 12
3x + 4y ≤ 100 2x + 9y ≥ 36
x,y ≥ 0 2x + 3y = 24
x,y ≥ 0

1.) Since the inequalities involve less than, simply add slack variable.
For constraint 1, 2x + 3y ≤ 80, add S1. The subscript 1 indicates that the
slack variable is added to the left side of the 1st constraint.

Thus, constraint 1 becomes

2x + 3y + S1 = 80

Doing the same process to constraint 2, then it becomes

3x + 4y + S1 = 100

The objective function should also contain the slack variables added in the
constraints. But since the slack variable has no contribution to the
objective function, its numerical coefficient must be zero. Thus the
objective function becomes

Max Z = 5x + 12y + 0S1 + 0S2

73
 2.) Max Z=120x + 80y + 0S1 + 0S2 – MA2 – MA3
Subject to 2x + y + S1 = 12 ? Do you know how the
subscripts were used?
2x + 9y -S2+A2 = 36
2x + 3y + A3 = 24

4.1.2 Obtaining the Final Solution of LPP

The following examples illustrate the step-by-step solution of LPP.

Examples : Solve the following linear programming problems

Maximize Z = 90x + 110y

Subject to 6x + 4y ≤ 120 } explicit constraints
3x + 10y ≤ 180
X ≥ 0,y ≥ 0 implicit constraint
New program with slack variables (simplified)

Maximize Z = 90x + 110y + 0S1 + 0S2

Subject to 3x + 2y + S1 = 60
3x + 10y +S2 = 180
x,y,S1,S2 ≥0
Initial Table:

Contribution to profit column

Variable column
Table 4.1 constant column
Cj 90 110 0 0 Q1
C1 Solution Quantity X Y S1 S2 Objective
coef. row
0 S1 60 3 2 1 0 0 Variable row
0 S2 180 3 10 0 1 0 } constraint
Zj 0 0 0 0 0 coef.row
Cj – Zj 90 110 0 0

Initial tableau in the succeeding examples and in exercises will not be

shown anymore. Initial tableau and iteration 1 will be in one table only. The first
table is considered as the first iteration.

74
 Iteration 1:

i.) Determine the optimum column and the pivotal row, and then highlight then.

ii.) The intersection of these column and row is the pivot element
iii.) Determine the Q1 column.

60÷2=30
180 ÷ 10 = 18
Table 4.2
Cj 90 110 0 0 Q1
C1 Solution Quantity X Y S1 S2
0 S1 60 3 2 1 0 30
0 S2 180 3 10 0 1 18
Zj 0 0 0 0 0
Cj- Zj 90 110 0 0

Iteration 2:
i.) Pivotal row: Second constraint coefficient row.

Change the outgoing variable by the entering variable (reflect

corresponding coefficients in the C1 column). Pivot element in the replacing pivot
row should be equal to one (1), thus divide each entry in this row by the pivot
element (10).
180÷10=18
3 ÷ 10 = 3/10
10 ÷ 10 = 1 New pivot element
0 ÷ 10 = 0
1 ÷ 10 = 1/10

Table 4.3
Cj 90 110 0 0 Q1
C1 Solution Quantity x y S1 S2

110 Y 18 3/10 1 0 1/10

Zj
Cj – Zj

7
5
ii.) Remaining rows

1.) First constraint coefficient row

The intersection with the optimum column should be zero. There are
several algebraic manipulations that can be used to make it zero, one is to
apply Formula 4.1.

-22 R22 + R11 = R22 Intersection with the optimum column

-2 (18)+60 = 24
-2( 3/10 ) +3 = 12/5
-2 (1)+2 =0
-2 (0)+1 =1
-2(1/10)+ = -1/5

2.) Z1 row

Z1 = C1(R21)+C2(R22)
Z1 = 0 (24) + 110 (18) = 1980 The Cj – Zj may be computed from the
0 (12/5) + 110 (3/10) = 33 table
0 (0) + 110 (1) = 110
0 (1) + 110 (0) =0
0 (-1/5) + 110 (1/10) = 1

Table 4.4 (Complete entry except column Q1)

Cj 90 110 0 0 Q1
C1 Solution Quantity x y S1 S2
0 S1 24 12/5 0 1 -1/5 R21
110 y 18 3/10 1 0 1/10 10 R22
Z1 1980 33 110 0 11 60
C1–Z1 57 0 0 -11

Table 4.5 (New optimum column and New pivotal row)

Cj 90 110 0 0 Q1
C1 Solution Quantity x y S1 S2
0 S1 24 12/5 0 1 -1/5 R21
110 y 18 3/10 1 0 1/10 10 R22
Zj 1980 33 110 0 11 60
Cj – Zj 57 0 0 -11

76
Solution for Q1:
24 ÷ 12/5 = 10
18 ÷ 3/10 = 60

Zj = 0 ( 5 ) – M ( 8 ) + 0 ( 10 ) = - 8M
0(1)–M920+0(1)=-2M
0(1)–M(1)+0(2)=-M
0(3)–M(1)+0(2)=-M
0(1)–M(0)+0(0)=0
0(0)–M(1)+0(0)=-M
0(0)–M(0)=0(1)=0

The solution is not yet final at this point since there is still positive entry in
the (Cj – Zj) row.

Iteration 3:

i.) Pivotal row: R21

Change the 0 and S1 (outgoing variable) by 90 and x (entering variable)
respectively.

Pivot element this time is 12/5. Again it should be change to one (1), thus
multiply each entry in this row by 5/12 to make it one.

R31 = 5/12 R21 -5/12 ( 24 ) = 10

-5/12 ( 12/5 ) = 1 New pivot element
5/12(0)=0
5/12 ( 1 ) = 5/12
5/12 ( -1/5 ) = -1/12

ii.) Remaining row

1.) For R22 the intersection with the optimum column should be equal to zero.

R32 3/10 R21 + R22

3/10 (10 ) + 18 =15
-3/10 (1 ) = 3/10 = 0 Intersection with the optimum
column
-3/10(0)+1=1
-3/10( 3/12 ) + 0 = -1/8
-3/10(-1/12)+1/10= 1/8

77
 2.) Z1 row

Zj= 90 (10)+ 110 (15)=2550

90 ( 1 ) + 110 (0)=90
90 ( 0 ) + 110 (1)=110
90 (5/12)+ 110 (-1/8) = 95/4
90 (-1/12)+ 110 (1/8) = 25/4

Table 4.6 (Final Tableau)

Cj 90 110 0 0
C1 Solution Quantity x y S1 S2
90 x 10 1 0 5/12 -1/12 R31
110 y 15 0 1 -1/8 1/8 R32
Zj 2550 90 110 95/4 25/4
Cj – Zj 0 0 -95/4 -25/4

Since the entries in the Cj-Zj are zeros and negative numbers, then the
last iteration gives the optimum solution. Thus optimum values of x = 10, y =
15,andZ = 2550.

Example 2. Maximize Z = 25x + 30y + 20z

Subject to : x + y + 3z ≤ 5
2x + y + z = 8
X + 2y + 2z ≤ 10
X, y, z ≥ 0

New program:
Maximize Z = 25x + 30y + 20z + 0S1 – MA2 +
0S3 Subject to x + y + 3z + S1 = 5
2x + y + z + A2 = 8

Do you know how the slack and artificial variables were added?
x + 2y + 2z + S3 = 10
x, y, z, S1,S3,A2 ≥ 0

The initial tableau may be skipped since it has the same entries with
iteration 1 except for Cj column, Zj and Cj – Zj rows

7
8
Iteration 1
Below is the set of solutions for Zj row Solution for Q1
entries:
5÷1=5
Zj = 0(5)–M(8)+0(10)=-8M 8÷2=4
0(1)–M920+0(1)=-2M 10÷1=10
0(1)–M(1)+0(2)=-M
0(3)–M(1)+0(2)=-M
0(1)–M(0)+0(0)=0
0(0)–M(1)+0(0)=-M
0(0)–M(0)=0(1)=0 From the Cj – Zj row, the
most positive is ( 25 + 2M ),
thus the optimum column is the x row (highlighted). From the Q1 column, the
smallest number is 4, thus the pivotal row is A2 row or R12 row (highlighted).
Table 4.7
Cj 25 30 20 0 -M 0 Qj
C1 Solution Quantity X Y Z S1 A2 S3
0 S1 5 1 1 3 1 0 0 5
-M A2 8 2 1 1 0 1 0 4
0 S3 10 1 2 2 0 0 1 10 R11
Zj -8M -2M -M -M 0 -M 0 R12
Cj-Zj 25+2M 30+M 20+M 0 0 0 R13

Entering variable: ___x____ Leaving variable: ___A 2 _____

*Do you know how the entering and leaving variables were determined?
Iteration 2

a.) Pivotal row (R12): Divide each entry by 2 to make pivot equal to one.

R22=R21÷2= 8÷2=4
2 ÷2=1 New pivot
1 ÷2=½
1÷2=½
0÷2=0
1÷2=½
0÷2=0

7
9
 b.) Remaining rows

1.) R21 and R23

R23 -R22 + R11 R23 -R22 + R11

-(4)+5=1 -(4)+10=6
-(1)+1=0 Intersection with the -(1)+1=0
optimum column.
-(1/2)+1=½ (1/2) + 2 = 3/2
-(1/2)+3=5/2 (1/2)+2=3/2
-(0)+1=1 -(0)+0=0
-(½)+0=-1/2 (½)+0=-1/2
-(0)+0=0 -(0)+1=1

1.) Zj row
Zj = 0(1)+2594)+0(6)=100
0(0)+2591)+0(0)=25
0 (½)+25(½)+0(3/2)=25/2
0 (5/2)+25(½)+0(3/2)=25/2
0(1)+25(0)+0(0)=0
0 (-1/2)+25(½)+0(-1/2)=25/2
0(0)+25(0)+)(1)=0

3.) Cj – Zj row may be computed from the table.

Table 4.8
Cj 25 30 20 0 -M 0 Qj
C1 Solution Quantity X y Z S1 A2 S3
R21
0 S1 1 0 1/2 5/2 1 -1/2 0 2
25 x 4 1 1/2 1/2 0 1/2 0 8 R22

0 S3 6 0 3/2 3/2 0 1/2 1 4 R23

Zj 100 25 25/2 25/2 0 25/2 0
Cj-Zj 0 35/2 15/2 0 -M-25/2 0

Entering variable: __y__ _ Leaving variable: __S 1 ___

80
Iteration 3

a.) Pivotal row (R21): Multiply each entry by 2 to

make pivot equal to 1.

R31 = 2R21 2(1) =2

2(0) = 0
2(½) =1 New pivot
2 (5/2) = 1
2(1) = 2
2 ( 9-1/2 ) = -1
2(0) = 0

b.) Remaining row:

1.) R32 and R33

R33 -3 R21 + R23
R32 -R21 + R22
-3(1) +6 = 3
-(1) +4 = 3
-3(0) +0 = 0
-(0) + 1 = 1 Intersection with the
-3(½) +3/2= 0
-(½) +½= 0 optimum column.
-3(5/2) +3/2=-6
-(5/2) +½=-2
-3(1) + 0 = -3
-(1) + 0 = -1
-3(-1/2)+½ = 1
-(-1/2)+½= 1
-3(0) +0 = 1
-(0) + 0 = 0

2.) Zj row
Zj = 30 (2) +25(3) +0(3) =135
30(0) +25(1) +0(0) = 25
30(1) +25(0) +0(0) = 30
30(5) +25(-2)+0(-6)=100
30(2) +25(-1)+0(-3) = 35
30(-1)+25(1) +0(1) = -5
30(0) +25(0) +0(1) = 0

81
Table 4.9 (Final Tableau)

Cj 25 30 20 0 -M 0
C1 Solution Quantity x y Z S1 A2 S3
30 y 2 0 1 5 2 -1 0 R31
R32
25 x 3 1 0 -2 -1 1 0 R33
0 S3 3 0 0 -6 -3 1 1
Zj 135 25 30 100 35 5 0
Cj-Zj 0 0 -80 -35 -M+5 0

Since the entries in the Cj-Zj of the last table are zeros and negative numbers,
then the solution is optimum and Table 3.9 is the final tableau. Hence the
optimum values of x = 3, y = 2, z = 0 (the value is 0 if the variable is not in the
final tableau) and the optimum value of Z = 135.

82
Solve the following LPP:

New Program:
Max Z = 65x1 + 70x2 + 60x3
Subject to 2x1 + 3x2 + 4x3 ≤ 34 Subject to:
x1 + 5x2 + 2x3 ≤ 48
2x1 + 3x2 ≤ 20
x1, x2,x3 ≥ 0

Iteration 1:
Cj
C1 Solution Quantity X1 X2 X3

Zj
Cj-Zj

Entering variable: Leaving variable:

Sol’n for Qj: New sol’n for X2:

Iteration 2:
Cj
C1Solution Quantity X1 X2 X3

Zj
Cj-Zj

Entering variable: Leaving variable:

83
 Sol’n for S1: Sol’n for S2: Sol’n for Qj: New sol’n for X3:

Iteration 3:
Cj
C1Solution Quantity X1 X2 X3

Zj
Cj-Zj

Entering variable: Leaving variable:

Sol’n for S2: Sol’n for X2: Sol’n for Qj: New sol’n for X1:

Final Tableau:
Cj
C1Solution Quantity X1 X2 X3

Zj
Cj-Zj

X1 = X2 = X3 = Max value of Z =

84
 Sol’n for S2: Sol’n for X3:

New Program:
2. Max Z = 120x1 + 80x2
Subject to 2x1 +x2 ≤ 12 Subject to:
2x1 + 9x2 ≥ 36
2x1 + 3x2 = 24
x1, x2 ≥ 0
Iteration 1:
Cj
C1 Solution Quantity X1 X2

Zj
Cj-Zj

Entering variable: Leaving variable:

Sol’n for Qj: New sol’n for X2:

Iteration 2:
Cj
C1Solution Quantity X1 X2

Zj
Cj-Zj

85
 Entering variable: Leaving variable:

Sol’n for S1: Sol’n for A3: Sol’n for Qj: New sol’n for X1:

Iteration 3:
Cj
C1Solution Quantity X1 X2

Zj
Cj-Zj

Entering variable: Leaving variable:

Sol’n for X2: Sol’n for A3: Sol’n for Qj: New sol’n for S2:

Final Tableau:
Cj
C1Solution Quantity X1 X2

Zj
Cj-Zj

86
X1 = x2 = Maximum value of Z =

Sol’n for X1: Sol’n for X2:

New Program:
4. Max Z = 8x + 7y
5.
Subject to: x +y ≤ 10 Subject to:
x+y≥2
x≤5
y≤6
x, y ≥ 0

Iteration 1:
Cj
C1 Solution Quantity X Y

Zj
Cj-Zj

Entering variable: Leaving variable:

Sol’n for Qj: New sol’n for x:

87
Iteration 2:
Cj
C1Solution Quantity X y

Zj
Cj-Zj

Entering variable: Leaving variable:

Sol’n for S1: Sol’n for A3: Sol’n for Qj: New sol’n for S2:

Iteration 3:
Cj
C1 Solution Quantity x y

Zj
Cj-Zj

Entering variable: Leaving variable:

88
 Sol’n for S1: Sol’n for x: Sol’n for Qj: New sol’n for y:

Final Tableau:
Cj
C1 Solution QuantityX Y

Zj
Cj-Zj

Sol’n for S2: Sol’n for S4:

Decision: x = y= Max value of Z =

89
 4.2 Simplex Minimization
Steps in Solving Minimization Problems

The steps in solving minimization problems by simplex method are the same as in
maximization except for steps 2, 4, 1, 6 and 7.
2. Follow the following rules in converting the LP model to standard form:
Symbol in the constraint To put into Standard Form
≤ Add a slack variable ( S1 ).

≥ Subtract a surplus variable ( S1 ) and add an artificial variable ( A1 )

= Add an artificial variable ( A1 )

The coefficient of A1 in the objective function is a very big number, let it equal to M, ( the letter M
means million ), or 100 if the numerical coefficients found in the constraints and objective function
are less than or equal to 10, let it equal to 1000 if the numerical coefficients are less than or equal
to 100, let it equal to 10000, if the numerical coefficients are less than or equal to 1000, and so on.
4.1 Determine the optimum column with the most negative value in the Cj – Zj row of the
initial table.

If the Cj – Zj has still negative entry, go back to step 4 and step 5 ( for iteration 3,4….)

The optimum or final solution is obtained if the entries in the Cj – Zj row are all zero or positive
values.
Examples : Write the following LP models to standard form

New Program
Minimize Z = 6x + 4y Minimize Z = 6x + 4y + MA1 + 0S2 + 0S3 + MA3
Subject to x + 2y = 10 Subject to x + 2y +A1 = 10
y≤4 y + S2 =4
3x + 2y ≥ 18 3x + 2y -S3+A3 =18
x,y ≥ 0 x, y, S2, A1, A3 ≥ 0

Iteration 1:
Cj 6 4 M 0 0 M Q1
C1 Solution Quantity x y A1 S2 S3 A3
M A1 10 1 2 1 0 0 0 5 R11
R12
0 S2 4 0 1 0 1 0 0 4 R13
M A3 18 3 2 0 0 -1 1 9
Zj 28M 4M 4M M 0 -M M
Cj-Zj 6-4M 4-4M 0 0 M 0

90
Entering variable _______y_____ Leaving variable ____S 2___________
Solution for iteration 1:
Zj = M(10)+0(4)+M(18)=28 Zj = M(0)+0(1)+M(0)=0
M M(0)+0(0)+M(-1)=-M
M(1)+0(0)+M(3)=4M M(0)+0(0)+M(1)=M
M(2)+0(1)+M(2)=4M
M(1)+0(0)+M(0)=M

Iteration 2:
Cj 6 4 M 0 0 M Q1
C1 Solution Quantity x y A1 S2 S3 A3
R21
M A1 2 1 0 1 -2 0 0 2
R22
4 Y 4 0 1 0 1 0 0 R23
M A3 10 3 0 0 -2 -1 1 3.33
Zj 16+12M 4M 4 M 4-4M -M M
Cj-Zj 6-4M 0 0 -4 M 0
+4M

Entering variable ___x______ Leaving variable ____A1__________

Solutions for iteration 2:

Pivotal row: The pivotal row R12 – R22, since pivot is already equal to one.

Remaining rows:

R21 -2 R12 + R11 R23 -2 R12 + R13

-2(4)+10=2 -2(4)+18=10
-2(0)+1=1 -2(0)+3=3
Intersection with the optimum
-2(1)+2=0 -2(1)+2=0
column
-2(0)+1=1 -2(0)+0=0
-2(1)+0=-2 -2(1)+0=-2
-2(0)+0=0 -2(0)+1=-1
-2(0)+0=0 -2(0)+1=1

Zj = M(2)+4(4)+M(10)=16 Zj = M(2)+4(1)+M(-2)=4–4M
+12 M M(0)+4(0)+M(-1)=-M
M(1)+4(0)+M(3)=4M M(0)+4(0)+M(1)=M
M(0)+4(1)+M(0)=4
M(1)+4(0)+M(0)=M
91
Iteration 3:
Cj 6 4 M 0 0 M Q1
C1 Solution Quantity x y A1 S2 S3 A3
6 X 2 1 0 1 -2 0 0 --- R31
R32
4 Y 4 0 1 0 1 0 0 4 R33
M A3 4 0 0 -3 4 -1 1 1
Zj 28+4M 6 4 6-3M -8 -M M
+4M
Cj-Zj 0 0 4M-6 8-4M M 0

Entering variable ____S2_____ Leaving variable ____A3__________

Solutions for iteration 3:

Pivotal row: The pivotal row R31 = R21, since pivot is already equal to one.

Remaining rows: R32 = R22, since the intersection with the optimum column is already equal
to zero.

R33-3 R21 + R23

-3(2)+10=4
-3(1)+3=0 Intersection with the optimum column
3(0)+0=0
3(1)+0=-3
3(-2)+(-2)=4 -
3(0)+(-1) =-1 -
3(0)+1=1

Zj row:
Zj = 6 ( 2) + 4 ( 4 ) +M ( 4 ) = 28 + 4M
6(1)+4(0)+M(0)=6
6(0)+4(1)+M(0)=4
6(1)+4(0)+M(-3)=6–3M
6(-2)+4(1)+M4)=-8+4M
6(0)+4(0)+M(-1)=-M
6(0)+4(0)+M(1)=M

92
Final Tableau:
Cj 6 4 M 0 0 M
C1 Solution Quantity x y A1 S2 S3 A3
6 X 4 1 0 -1/2 0 -1/2 1/2 R41
4 Y 3 0 1 1/4 0 1/4 -1/4 R42
0 S2 1 0 0 -1/4 1 -1/4 1/4 R43
Zj 36 6 4 0 0 -2 2
Cj-Zj 0 0 M 0 2 M-2

Solutions for the final tableau:

Pivotal row:
R43 R33+4
4+4 =1
0+4 =0
0+4 =0
-3+4 = -1/4
4+4 =1 New pivot
-1+4 = -1/4
1+4 = 1/4

Remaining rows:
R42 -1/4 R33 + R32
R41 ½ R33 + R31
-1/4(4)+4 =3
½(4)+2 =4
-1/4(0)+0 =0
½(0)+1 =1
½(0)+0 =0 -1/4(0)+1 =1
-1/4(-3)+0=-1/4
½(-3)+1=-1/2
½(4)+2 =0 Intersection with the optimum -1/4(4)+1 =0
column -1/4(-1)+0=1/4
½(-1)+0 = -1/2
-1/4(1)+0 = -1/4
½(1)+0 = 1/2

Zj = 6(4)+4(3)+0(4)=36 From the last table, the Cj – Zj has zeros

6 (1)+4 (0)+.0(0)=6 and positive entries, thus it gives the
6 (0)+4 (1)+0(0)=4 optimum solution. The solutions are x =
6 (-1/2)+4(3/4)+0(-3)=0 4, y = 3 and the minimum value of Z =
6 (0)+4 (0)+0(4)=0 36.
6 (-1/2)+4(1/4)+0 (-1)=-2 In the previous problem, if 1000
6 (1/2)+4(-1/4)+0 (1)=M is used instead of the number M, the
following are the solutions. Note that we
can easily determine the most negative and positive entry in the last row,Cj-Zj, of each tableau.
Solutions for each tableau are left to the reader.
93
 New Program
Minimize Z = 6x + 4y Minimize Z = 6x + 4y + 1000A 1 + 0S 2 + 0S3 + 1000A 3
Subject to x +2y = 10 Subject to: x + 2y + A1 = 10
y≤4 y+ S2 =4
3x +2y ≥ 18 x + 2y -S3+A3 = 18
x,y ≥ 0 x,y,S2,S3,A1,A3 ≥ 0

Iteration 1:
Cj 6 4 1000 0 0 1000 Q1
C1 Solution Quantity x Y A1 S2 S3 A3
R11
1000 A1 10 1 2 1 0 0 0 5
R12
0 S2 4 0 1 0 1 0 0 4 R13
1000 A3 18 3 2 0 0 -1 1 9
Zj 28000 4000 4000 1000 0 -1000 1000
Cj-Zj -3994 -3996 0 0 1000 0

Entering variable ____y_____ Leaving variable ____S2__________

Iteration 2:
Cj 6 4 1000 0 0 1000 Q1
C1 Solution Quantity x Y A1 S2 S3 A3
R21
1000 A1 2 1 0 1 -2 0 0 2
R22
4 Y 4 0 1 0 1 0 0 --- R23
1000 A3 10 3 2 0 -2 -1 1 3.33
Zj 12016 4000 4 1000 -3996 -1000 1000
Cj-Zj -3994 0 0 3996 1000 0

Entering variable ____x_____ Leaving variable ____A 1________

Iteration 3:
Cj 6 4 1000 0 0 1000 Q1
C1 Solution Quantity x y A1 S2 S3 A3
1000 A1 2 1 0 1 -2 0 0 R31
R32
4 y 4 0 1 0 1 0 0 4 R33
1000 A3 4 0 0 -3 4 -1 1 1
Zj 4028 6 4 -2994 3992 -1000 1000
Cj-Zj 0 0 3994 -3992 1000 0

Entering variable S2 Leaving variable: A3

94
Final Tableau:
Cj 6 4 1000 0 0 1000 Q1
C1 Solution Quantity x y A1 S2 S3 A3
R41
6 x 4 1 0 -1/2 0 -1/2 1/2 0
R42
4 y 3 0 1 3/4 0 1/4 -1/4 0 R43
0 S2 1 0 0 -3/4 1 -1/4 1/4 0
Zj 36 6 4 0 0 -2 2
Cj-Zj 0 0 1000 0 2 998

Decision : x = 4 , y= 3 and the Minimum Z = 36 .

Solve the following LPP:

New Program
Min Z = 40x + 24y + 32z _______________________________________
Subject to y – z ≤ 0 Subject to:__________________________
X + y + z ≥ 25 _________________________________
Z≤5 _________________________________
x,y,z ≥ 0 _________________________________

Iteration 1:
Cj
C1 Solution Quantity x y z

Zj
Cj-Zj

Entering variable: Leaving variable: .

Iteration 2:
Cj
C1 Solution Quantity x y z

Zj
Cj-Zj

Decision: x = y= z= Minimum Z =

95