ME 324: Industrial Engineering and Operations Research

Assignment 7 solution (Section A)

Q. 1: Obtain the dual form of the following problem:
Minimize z = 40y1 + 24y2 + 10y3
s.t.
y1 + 3y2 + y3  2,
3y1 + y2 + y3  5,
y1  0, y2  0, y3  0.
By solving the dual form using simplex method, find out the minimum value of z and corresponding y1,
y2 and y3. Which are the active constraints of primal problem? Find out the unit worth of active
constraints.
Solution:
Dual form:
Maximize w = 2x1 + 5x2
s.t
x1 + 3x2  40,
3x1 + x2  24,
x1 + x2  10
x1  0, x2  0.
We introduce slack variables to convert the dual form into standard form.
Maximize w = 2x1 + 5x2 + 0x3 + 0x4 + 0x5
s.t
x1 + 3x2 + x3 = 40,
3x1 + x2 + x4 = 24,
x1 + x2 + x5 = 10,
x1, x2, x3, x4, x5  0.
The starting tableau is:
Basic x1 x2 x3 x4 x5 Solution
w –2 –5 0 0 0 0
x3 1 3 1 0 0 40
x4 3 1 0 1 0 24
x5 1 1 0 0 1 10
The most negative coefficient in the w-row is –5 (corresponding to x2), so x2 enters the basis and x5
leaves the basis.

Basic x1 x2 x3 x4 x5 Solution
w 3 0 0 0 5 50
x3 –2 0 1 0 –3 10
x4 2 0 0 1 –1 14
x2 1 1 0 0 1 10
Perform row operations to make the coefficients of basic variables x2 equal to zero in w, x3, and x4-row.
Since there are no negative values in the w -row, we have reached the optimal solution.

1
The optimal solution is x1 = 0 and x2 = 10. Maximum w = 5×10 = 50.
At the optimum, the objective value of dual and corresponding primal problem is equal. Hence,
minimum value of z = 50.
In the dual problem, the optimal values of the slack/surplus variables in the objective function (w-row)
correspond to the values of the primal variables i.e., y1 = 0, y2 = 0, and y3 = 5.
The active constraints in the primal problem correspond to the dual variable values in the dual solution.

The primal constraint y1 + 3y2 + y3  2 (corresponding to dual variable x1) is inactive (since x1 = 0).

The primal constraint 3y1 + y2 + y3  5 (corresponding to dual variable x2) is active (since x2 = 10).

The unit worth of active constraint 3y1 + y2 + y3  5 is x2 = 10 (corresponding dual variable value).

Q.3: Consider the following problem:

Minimize z  x1 + 3x2
Subject to
x1+x2 = 4,
x1  0, x2  0.
Obtain the dual form of this problem. Solve the dual form graphically as well as by simplex method
and by that solution only tell the maximum value of z and the unit worth of resource of constraint
x1+x2=4.
Solution:
Dual form:

Maximize w  4y
Subject to
y  1,
y  3,
y is unrestricted.
Using graphical method:

Fig. 1. Feasible region of the dual problem

2
Since y is unrestricted, it can take both positive and negative values. Therefore, the feasible region
extends infinitely in the negative direction. The feasible region for y is bounded by y ≤ 1, as it is stricter
than y ≤ 3. Therefore, the feasible region includes all values of y ≤ 1, extending infinitely in the negative
direction.
The objective function is w = 4y. To maximize w, we need to take the largest possible value of y within
the feasible region (y = 1). Hence, maximum w = 4×1 = 4.
Using simplex method
We introduce slack variables to convert the dual form into standard form. Since, y is unrestricted in
sign, we split it into two variables i.e., y = y+ – y

Maximize w = 4y+4y + 0s1 + 0s2

Subject to
y+– y– + s1 = 1,
y+ – y– + s2 = 3,
y–, y+, s1, s2  0.

The starting tableau is:

Basic y+ y s1 s2 Solution
w –4 4 0 0 0
s1 1 –1 1 0 1
s2 1 –1 0 1 3
The most negative coefficient in the w-row is –4 (corresponding to y+), so y+ enters the basis and s1
leaves the basis (using minimum ratio test).

Basic y+ y s1 s2 Solution
w 0 0 4 0 4
y+ 1 –1 1 0 1
s2 0 0 –1 1 2
Perform row operations to make the coefficients of basic variable y+ equal to zero in w, and s2-row.
Since there are no negative values in the w -row, we have reached the optimal solution.

From the tableau: y+ = 1, and y= 0, which gives y = y+ y–= 1  0 =1

Therefore, the optimal solution is: y =1. Maximum w = 4×1 + 4×0 = 4.
At the optimum, the objective value of dual and corresponding primal problem is equal. Hence,
maximum value of z = 4.
Unit worth of resource of constraint x1 + x2 = 4 is y =1.

Q. 5: Consider the following linear programming problem:

Maximize z = 2x + y
subject to
x – y ≤ 0,
2x + 3y ≤ 30,
x, y  0.

3
Solve this linear programming problem by simplex method. From the final simplex tableau, find out
the values of decision variables x & y, the objective function and unit worth of resources of the
constraints.
Solution:
Maximize z = 2x + y
subject to
x – y + s1 ≤ 0,
2x + 3y + s2 ≤ 30,
x, y, s1, s2  0.
The starting tableau is:

Basic x y s1 s2 Solution
z –2 –1 0 0 0
s1 1 –1 1 0 0
s2 2 3 0 1 30
The most negative coefficient in the z-row is –2 (corresponding to x), so x enters the basis and x5 leaves
the basis.
s1: (0/1) = 0, s2 = (30/2) = 15. By minimum ratio rule, s1 is leaving.

Basic x y s1 s2 Solution
z 0 –3 2 0 0
x 1 –1 1 0 0
s2 0 5 –2 1 30
Now y becomes the entering variable and s2 becomes the leaving variable.

Basic x y s1 s2 Solution
z 0 0 4/5 3/5 18
x 1 0 3/5 1/5 6
y 0 1 –2/5 1/5 6

The optimal solution is x = 6, and y = 6. Maximum z = 18. Unit worth of resource for first constraint
corresponds to coefficient of s1 in z-row i.e., 4/5, and unit worth of resource for second constraint
corresponds to coefficient of s2 in z-row i.e., 3/5.

Q. 7: Carry out first phase operation of two phase Simplex method for the following problem:
Minimize z = 4x1 + x2 + x3
Subject to
2x1 + x2 + 2x3 = 4,
4x1 + 2x2 + 4x3 = 7,
x1, x2, x3  0.
By the Simplex table obtained in the first phase, show that there is no feasible solution to this problem.
Solution:
Phase I: Minimizing the Artificial Variable
Define the objective function: r = R1 + R2
subject to
2x1 + x2 + 2x3 + R1 = 4

4
4x1 + 2x2 + 4x3 + R2 = 7
x1, x2, x3, R1, R2  0
Initial Simplex Tableau for Phase 1
Basic x1 x2 x3 R1 R2 Solution
r 0 0 0 −1 −1 0
R1 2 1 2 1 0 4
R2 4 2 4 0 1 7
Make the r-row consistent by adding second and fourth row in r-row. Apply the simplex table, you shall
get the following final table:
Basic x1 x2 x3 R1 R2 Solution
r 6 3 6 0 0 11
R1 2 1 2 1 0 4
R2 4 2 4 0 1 7
Considering x1 (with the most positive r-row coefficient) as entering variable and R2 as leaving variable
(using minimum ratio test), the following simplex table is obtained.
Basic x1 x2 x3 R1 R2 Solution
r 0 0 0 0 –3/2 1/2
R1 0 0 0 1 –1/2 1/2
x1 1 1/2 1 0 1/4 7/4
Perform row operations to make the coefficients of basic variable x1 equal to zero in R1 and r-row.
The iterations of phase I are complete. Optimum iteration shows that the artificial variable R 1 in the
base is positive (1/2), meaning that the LP is infeasible.

Q.9: Use the dual simplex method, solve the following problem:
Maximize z = 10 – 3x1 – 2x2
subject to
x1 + x2 ≥ 1; x1 + x2 ≤ 7; x1 + 2x2 ≥ 10; x2 ≤ 3, and x1, x2 ≥ 0.
Solution:
Make proper LP structure; ignore constant term in the objective function. Multiply the first and third
inequalities by –1 to convert them to ≤ constraints (all constraints must be of the type ≤).
Convert the maximization problem to minimization problem for easier handling. Introduce slack
variables to convert into standard form.
Minimize z = 3x1 + 2x2 + 0x3 + 0x4 + 0x5 + 0x5
subject to
–x1 – x2 + x3 = –1,
x1 + x2 + x4 = 7,
–x1 – 2x2 + x5 = –10,
x2 + x6 = 3
x1, x2, x3, x4, x5, x6  0.

5
The starting tableau is:
Basis x1 x2 x3 x4 x5 x6 Solution
Z –3 –2 0 0 0 0 0
x3 –1 –1 1 0 0 0 –1
x4 1 1 0 1 0 0 7
x5 –1 –2 0 0 1 0 –10
x6 0 1 0 0 0 1 3
The tableau is optimum because all the reduced costs in Z-row are 0. The starting solution is infeasible
as at least one of the solutions is negative.
In dual simplex method, we focus on the most negative basic variable. The most negative basic variable
is –10 in the x5 row, so x5 will be our leaving variable.
In x5 row, coefficient of x1 and x2 are negative. Hence, we will consider them for calculating ratio
(|coefficient in Z-row/coefficient in x5 row|): x1: |(3/1)| = 3, x2 = (2/2) = 1. By minimum ratio rule,
x2 is entering.

Basic x1 x2 x3 x4 x5 x6 Solution
Z –2 0 0 0 –1 0 10
x3 –1/2 0 1 0 –1/2 0 4
x4 1/2 0 0 1 1/2 0 2
x2 1/2 1 0 0 –1/2 0 5
x6 –1/2 0 0 0 1/2 1 –2
Perform row operations to make the coefficients of basic variable x2 equal to zero in x3, x4 and x6-row.
Now x6 becomes the leaving variable (as it is most negative) and x1 becomes the entering variable
(only coefficient corresponding to x1 is negative; hence, no need to calculate ratio).
Basic x1 x2 x3 x4 x5 x6 Solution
Z 0 0 0 0 –3 –4 18
x3 0 0 1 0 –1 –1 6
x4 0 0 0 1 1 1 0
x2 0 1 0 0 0 1 3
x1 1 0 0 0 –1 –2 4
Now, this table is feasible as all the basic variables are positive.
Answer: x1 = 4 and x2 = 3. Z = 18. Hence, z = 10 – 18 = –8.