Worked Examples for Chapter 5

Example for Section 5.1

Consider the linear programming problem given in Table 6.1 as the dual problem for
the Wyndor Glass Co. example. This problem is
Minimize

W = 4 y1 + 12 y2 + 18 y3,

subject to
y1

+ 3 y3 3
2 y2 + 2 y3 5

and
y1 0, y2 0, y3 0.
(a) Identify the 10 sets of defining equations for this problem. For each one, solve
(if a solution exists) for the corresponding corner-point solution, and classify it as
a CPF solution or corner-point infeasible solution. For each corner-point solution,
give the corresponding basic solution and its set of nonbasic variables.
This problem has five constraints (two functional constraints and three nonnegativity
constraints), where the constraint boundary equations for these constraints are
obtained by replacing each by =. Since there are three decision variables, each
corner-point solution lies at the intersection of three of the constraint boundaries, so
the corresponding three constraint boundary equations become the defining equations
for that solution. There are 10 different sets of three constraint boundary equations
(defining equations).
The basic solutions are obtained from the corner-point solutions by using the
augmented form of the model. After introducing surplus variables y4 and y5, the
augmented form is
Minimize
subject to

W = 4 y1 + 12 y2 + 18 y3,
y1

+ 3 y3 - y4 3
2 y2 + 2 y3
- y5

and
y1 0, y2 0, y3 0, y4 0, y5 0.

Following are the 10 different ways of choosing three defining equations from
the five constraint boundary equations. Solving these three defining equations gives
the corner-point solution. The corresponding basic solution is obtained by then using
the augmented form of the model to calculate the values of the surplus variables. The
three variables that equal zero in the basic solution are the nonbasic variables.
(1) Defining Equations: y1 = 0, y2 = 0, y3 = 0
Corner-point solution: (0, 0, 0)
The corner-point solution (0, 0, 0) is infeasible.
The corresponding basic solution (y1, y2, y3, y4, y5) = (0, 0, 0, -3, -5)
Nonbasic variables: y1, y2, y3
(2) Defining Equations: y1 = 0, y2 = 0, y1 + 3 y3 = 3
Corner-point solution: (0, 0, 1)
The corner-point solution (0, 0, 1) is infeasible.
The corresponding basic solution (y1, y2, y3, y4, y5) = (0, 0, 1, 0, -3)
Nonbasic variables: y1, y2, y4
(3) Defining Equations: y1 = 0, y2 = 0, 2 y2 + 2 y3 = 5
Corner-point solution: (0, 0, 2.5)
The corner-point solution (0, 0, 2.5) is a CPF.
The corresponding basic solution (y1, y2, y3, y4, y5) = (0, 0, 2.5, 4.5, 0)
Nonbasic variables: y1, y2, y5
(4) Defining Equations: y1 = 0, y1 + 3 y3 = 3, y3 = 0
No solution exists.
(5) Defining Equations: y1 = 0, 2 y2 + 2 y3 = 5, y3 = 0
Corner-point solution: (0, 2.5, 0)
The corner-point solution (0, 2.5, 0) is infeasible.
The corresponding basic solution (y1, y2, y3, y4, y5) = (0, 2.5, 0, -3, 0)
Nonbasic variables: y1, y3, y5
(6) Defining Equations: y2 = 0, y1 + 3 y3 = 3, y3 = 0
Corner-point solution: (3, 0, 0)
The corner-point solution (3, 0, 0) is infeasible.

The corresponding basic solution (y1, y2, y3, y4, y5) = (3, 0, 0, 0, -5)
Nonbasic variables: y2, y3, y4
(7) Defining Equations: y2 = 0, 2 y2 + 2 y3 = 5, y3 = 0
No solution exists.
(8) Defining Equations: y2 = 0, y1 + 3 y3 = 3, 2 y2 + 2 y3 = 5
Corner-point solution: (-4.5, 0, 2.5)
The corner-point solution (-4.5, 0, 2.5) is infeasible.
The corresponding basic solution (y1, y2, y3, y4, y5) = (-4.5, 0, 2.5, 0, 0)
Nonbasic variables: y2, y4, y5
(9) Defining Equations: y3 = 0, y1 + 3 y3 = 3, 2 y2 + 2 y3 = 5
Corner-point solution: (3, 2.5, 0)
The corner-point solution (3, 2.5, 0) is a CPF.
The corresponding basic solution (y1, y2, y3, y4, y5) = (3, 2.5, 0, 0, 0)
Nonbasic variables: y3, y4, y5
(10) Defining Equations: y1 = 0, y1 + 3 y3 = 3, 2 y2 + 2 y3 = 5
Corner-point solution: (0, 1.5, 1)
The corner-point solution (0, 1.5, 1) is a CPF.
The corresponding basic solution (y1, y2, y3, y4, y5) = (0, 1.5, 1, 0, 0)
Nonbasic variables: y1, y4, y5

Example for Sections 5.2 and 5.4

Consider the following model.
Maximize

Z = 4 x1 + 3 x2 + 6 x3,

subject to
3 x1 +
x2 + 3 x3 30
2 x1 + 2 x2 + 3 x3 40
and
x1 0, x2 0, x3 0.

We introduce x4 and x5 as slack variables for the respective constraints. The

augmented form of the model then is
Maximize

Z = 4 x1 + 3 x2 + 6 x3,

subject to
3 x1 + x2 + 3 x3 + x4
2 x1 + 2 x2 + 3 x3

= 30
+ x5 = 40

and
x1 0, x2 0, x3 0, x4 0, x5 0.
Using matrix notation, we have

x1

x2

3 1 3 1 0
30
c = 4 3 6 0 0, A =
, b = , x = x3 .

2 2 3 0 1
40
x
4

x5

Now let us apply the revised simplex method step by step to solve this problem.
Iteration 0:
Since x4 and x5 are the initial basic variables,
x 4
xB = ,
x5

1 0
-1
B=
= B , so
0
1

x4
1 0 30
30
= B-1b =
= .

x5
0 1 40
40

cB = 0, 0
Iteration 1:
The coefficients in Eq. (0) are
cBB-1A - c = [-4

-3

cBB-1 = [0

6] for x1, x2, x3,

0]

for x4, x5.

Since -6 is the most negative coefficient, we choose x3 as the entering basic variable.
The coefficients of x3 in Eqs. (1) and (2) are
3 1 03 3
.
B-1
3 0 13 3

The right-hand side of these equations was identified in Iteration 0 as

30
x4
B-1b = , where this gives the value of xB = .

40
x5

Applying the minimum ratio test, since (30/3) < (40/3), we choose x4 as the leaving
basic variable. Thus,
x3
xB = .

x5

3
Since the coefficients of x3 in Eqs. (1) and (2) are and Eq. (1) yielded the
3

leaving basic variable,

1 / 3
1 / 3
= ,
=
3 / 3
1
1 / 3 0 1

B-1 =
1 1 0

so

B-1 becomes

1 / 3 0
0
=
.
1
1 1

Therefore,
x3
1 / 3 0 30
10
= B-1b =
= .

x5

1 1 40
10

cB = 6, 0 .
Iteration 2:

The coefficients in Eq. (0) now are

cBB-1A - c = [2
-1
cBB-1 = [2

0]
0]

for x1, x2, x3,

for x4, x5.

Since only x2 has a negative coefficient, it becomes the entering basic variable. The
coefficients of x2 in Eqs. (1) and (2) are
1
B-1 =
2

1 / 3 0 1
1 / 3

= .
1 1 2
1

The right-hand side of these equations was identified in Iteration 1 as

10
B-1b = , where this gives the value of xB =
10

x3
.

x5

Applying the minimum ratio test, since

10 10
,
1/ 3 1

we choose x5 as the leaving basic variable. Thus,

x3
xB = .

x2

1 / 3
Since the coefficients of x2 in Eqs. (1) and (2) are and Eq. (2) yielded the
1

leaving basic variable,

(1 / 3) / 1
1 / 3
=
,
=
1 / 1
1

B-1 =

so B-1 becomes

1 1 / 3 1 / 3 0
2 / 3 1 / 3

=
.
1 1 1
1
0
1

Therefore,

x3
2 / 3 1 / 3 30
20 / 3
= B-1b =
=

.
x2

1 40
1
10

cB = 6, 3 .
Optimality Test:
The coefficients in Eq. (0) now are
2 / 3 1 / 3 3

cBB-1A - c = [6 3]
1 2
1

= [1

cBB

-1

0]
= [1

1 3
- [4
2 3

3 6]

for x1, x2, x3,

1]

for x4, x5.

Since all these coefficients are nonnegative, the current solution is optimal. This
solution is
x3
20 / 3
,
xB = = B-1b =

x2

10

as found in Iteration 2. The other variables are nonbasic variables, so x1 = 0, x4 = 0,

and x5 = 0.
Z = 4(0) + 3(10) + 6(20/3) = 70.

Example for Section 5.3

Consider the following problem.
Maximize
Z = x1 - x2 + 2 x3,
subject to
x1 + x2 + 3x3
15
2x1

- x2 + x3

-x1 + x2 + x3
and
x1 0,

x2

0,

x3 0.

Let x4 , x5, and x6 denote the slack variables for the respective constraints. After the
simplex method is applied, a portion of the final simplex tableau is as follows:
Coefficient of:
Basic
Variable

x1

x2

x3

x4

x5

x6

Right

Eq

Side

(0)

3
2

1
2

x4

(1)

-1

-2

x3

(2)

1
2

1
2

x2

(3)

1
2

1
2

(a) Use the fundamental insight presented in Sec. 5.3 to identify the missing
numbers in the final simplex tableau. Show your calculations. Identify the
defining equations of the CPF solution corresponding to the optimal BF solution
in the final simplex tableau.
From the coefficients for (x4, x5, x6) in the final simplex tableau, we observe that

1 1 2
S* = 0 1 / 2 1 / 2
0 1 / 2 1 / 2

and y* = 0 3 / 2 1 / 2 .

Thus, by the fundamental insight presented in Sec. 5.3, the constraint coefficients for
(x1, x2, x3) in the final tableau are

0 0
1
1 1 2 1 1 3
S*A = 0 1 / 2 1 / 2 2 1 1 = 1 / 2 0 1 .
3 / 2 1 0
0 1 / 2 1 / 2 1 1 1

The coefficients in the objective function for (x1, x2, x3) in the final tableau are

1 1 3
yA c = 0 3 / 2 1 / 2 2 1 1 - 1 1 2 = 3 / 2 0 0. .
1 1 1
1 1 2 15 5

The final right-hand side is S*b = 0 1 / 2 1 / 2 2 = 3, and

0 1 / 2 1 / 2 4 1

Z* = y*b = [0

15

1/2] 2 = 5.

4

3/2

Therefore, the complete final simplex tableau is

Coefficient of:
Basic
Variable

x1

x2

x3

x4

x5

x6

Eq

Right
Side

(0)

x4
x3
x2

(1)
(2)
(3)

0
0
0

3/2
1
1/2
-3/2

3/2

1/2

0
0
1

0
1
0

1
0
0

-1
1/2
-1/2

-2
1/2
1/2

5
3
1

(b) Identify the defining equations for the CPF solution corresponding to the
optimal BF solution in the final simplex tableau.
Since the optimal BF solution is (x1, x2, x3, x4, x5, x6) = (0, 1, 3, 5, 0, 0), the
corresponding CPF solution is (x1, x2, x3) = (0, 1, 3). The nonbasic variables are x1, x5,
x6, and these are the indicating variables that indicate that the following constraints
hold with equality.
x1
2x1 - x2 + x3
-x1 + x2 +x3

0
2
4

Therefore, the defining equations for this CPF solution are

x1 = 0,
2x1 - x2 + x3
-x1 + x2 + x3

= 2,
= 4.