(CSE215) Operations Research and Optimisation Techniques

Sheet 3 – Solutions
Question 1
Show how the M-method will indicate that the following problem has no feasible
solution.
𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑧𝑧 = 2𝑥𝑥1 + 5𝑥𝑥2
Subject to
3𝑥𝑥1 + 2𝑥𝑥2 ≥ 6
2𝑥𝑥1 + 𝑥𝑥2 ≤ 2
𝑥𝑥1 ≥ 0, 𝑥𝑥2 ≥ 0

First, write out the constraints in the correct form:

𝑧𝑧 = 2𝑥𝑥1 + 5𝑥𝑥2 − 𝑀𝑀𝑀𝑀
3𝑥𝑥1 + 2𝑥𝑥2 − 𝑠𝑠1 + 𝑅𝑅 = 6
2𝑥𝑥1 + 𝑥𝑥2 + 𝑠𝑠2 = 2
𝑥𝑥1 , 𝑥𝑥2 , 𝑠𝑠1 , 𝑠𝑠2 , 𝑅𝑅 ≥ 0
𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠2 𝑅𝑅 Solution
1 -2 -5 0 0 M 0
0 3 2 -1 0 1 6
0 2 1 0 1 0 2

to get rid of the M in the z-row, perform 𝑧𝑧 = 𝑧𝑧 − 𝑀𝑀(𝑟𝑟𝑟𝑟𝑟𝑟 1)

𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠2 𝑅𝑅 Solution

1 -2-3M -5-2M M 0 0 -6M
0 3 2 -1 0 1 6
0 2 1 0 1 0 2

to decide which variables are to be the basic variables, look for variable columns with
only one non-zero elements, which must also be on separate rows.
Basic Variables: 𝑠𝑠2 , 𝑅𝑅 and 𝑧𝑧
Non-Basic Variables: 𝑥𝑥1 , 𝑥𝑥2 and 𝑠𝑠1
Thus, 𝑠𝑠2 = 2, 𝑅𝑅 = 6, 𝑥𝑥1 = 0, 𝑥𝑥2 = 0 and 𝑠𝑠1 = 0.
We then apply simplex iterations by looking for the most negative value in the z-row.
 Basic 𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠2 𝑅𝑅 Solution
𝑧𝑧 1 -2-3M -5-2M M 0 0 -6M
𝑅𝑅 0 3 2 -1 0 1 6
𝑠𝑠2 0 2 1 0 1 0 2

Basic 𝑥𝑥1 Solution Ratio

𝑅𝑅 3 6 6� = 2
3
𝑠𝑠2 2 2 2� = 1
2

new 𝑥𝑥1 -row = (0 1 0.5 0 0.5 0 1)

The leaving variable is 𝑠𝑠2 .
new 𝑧𝑧-row = (1 -2-3M -5-2M M 0 0 -6M) + (2+3M)(0 1 0.5 0 0.5 0 1)
= (1 0 -4-0.5M M 1+1.5M 0 2-3M)
new 𝑅𝑅-row = (0 3 2 -1 0 1 6) – (3)(0 1 0.5 0 0.5 0 1) = (0 0 0.5 -1 -1.5 1 3)
Basic 𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠2 𝑅𝑅 Solution
𝑧𝑧 1 0 -4-0.5M M 1+1.5M 0 2-3M
𝑅𝑅 0 0 0.5 -1 -1.5 1 3
𝑥𝑥1 0 1 0.5 0 0.5 0 1

Basic 𝑥𝑥2 Solution Ratio

𝑅𝑅 3�
0.5 = 6
0.5 3
𝑥𝑥1 1�
0.5 = 2
0.5 1

new 𝑥𝑥2 -row = (0 2 1 0 1 0 2)

The leaving variable is 𝑥𝑥1 .
new 𝑧𝑧-row = (1 0 -4-0.5M M 1+1.5M 0 2-3M) + (4+0.5M)(0 2 1 0 1 0 2)
= (1 8+M 0 M 5+2M 0 10-2M)
new 𝑠𝑠2 -row = (0 0 0.5 -1 -1.5 1 3) – (0.5)(0 2 1 0 1 0 2) = (0 -1 0 -1 -2 1 2)
Basic 𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠2 𝑅𝑅 Solution
𝑧𝑧 1 8+M 0 M 5+2M 0 10-2M
𝑅𝑅 0 -1 0 -1 -2 1 2
𝑥𝑥2 0 2 1 0 1 0 2

since R, the artificial variable, is a basic, non-zero variable at the optimal solution, then
the solution is infeasible.
Question 2
Consider the following set of constraints:
−2𝑥𝑥1 + 3𝑥𝑥2 = 3
4𝑥𝑥1 + 5𝑥𝑥2 ≥ 10
𝑥𝑥1 + 2𝑥𝑥2 ≤ 5
6𝑥𝑥1 + 7𝑥𝑥2 ≤ 3
4𝑥𝑥1 + 8𝑥𝑥2 ≥ 5
𝑥𝑥1 , 𝑥𝑥2 ≥ 0
For each of the following problems, develop the new 𝑧𝑧-row after substituting out the
artificial variables
(a) Maximise 𝑧𝑧 = 5𝑥𝑥1 + 6𝑥𝑥2 , subject to (1), (3) and (4).
(b) Maximise 𝑧𝑧 = 2𝑥𝑥1 − 7𝑥𝑥2 , subject to (1), (2), (4) and (5).
(c) Minimise 𝑧𝑧 = 3𝑥𝑥1 + 6𝑥𝑥2 , subject to (3), (4) and (5).
(d) Minimise 𝑧𝑧 = 4𝑥𝑥1 + 6𝑥𝑥2 , subject to (1), (2) and (5).
(e) Minimise 𝑧𝑧 = 3𝑥𝑥1 + 2𝑥𝑥2 , subject to (1) and (5).
(1) − 2𝑥𝑥1 + 3𝑥𝑥2 + 𝑅𝑅1 = 3
(2) 4𝑥𝑥1 + 5𝑥𝑥2 − 𝑠𝑠1 + 𝑅𝑅2 = 10
(3) 𝑥𝑥1 + 2𝑥𝑥2 + 𝑠𝑠2 = 5
(4) 6𝑥𝑥1 + 7𝑥𝑥2 + 𝑠𝑠3 = 3
(5) 4𝑥𝑥1 + 8𝑥𝑥2 − 𝑠𝑠4 + 𝑅𝑅3 = 5
𝑥𝑥1 , 𝑥𝑥2 , 𝑠𝑠1 , 𝑠𝑠2 , 𝑠𝑠3 , 𝑠𝑠4 , 𝑅𝑅1 , 𝑅𝑅2 , 𝑅𝑅3 ≥ 0
(a)
𝑧𝑧 − 5𝑥𝑥1 − 6𝑥𝑥2 + 𝑀𝑀𝑅𝑅1 = 0
𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠2 𝑠𝑠3 𝑅𝑅1 Solution
1 -5 -6 0 0 M 0
0 -2 3 0 0 1 3
0 1 2 1 0 0 5
0 6 7 0 1 0 3

𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠2 𝑠𝑠3 𝑅𝑅1 Solution

1 -5+2M -6-3M 0 0 0 -3M
0 -2 3 0 0 1 3
0 1 2 1 0 0 5
0 6 7 0 1 0 3

𝑧𝑧 + (−5 + 2𝑀𝑀)𝑥𝑥1 − (6 + 3𝑀𝑀)𝑥𝑥2 = −3𝑀𝑀

(b)
𝑧𝑧 − 2𝑥𝑥1 + 7𝑥𝑥2 + 𝑀𝑀𝑅𝑅1 + 𝑀𝑀𝑅𝑅2 + 𝑀𝑀𝑅𝑅3 = 0
𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠3 𝑠𝑠4 𝑅𝑅1 𝑅𝑅2 𝑅𝑅3 Solution
1 -2 7 0 0 0 M M M 0
0 -2 3 0 0 0 1 0 0 3
0 4 5 -1 0 0 0 1 0 10
0 6 7 0 1 0 0 0 0 3
0 4 8 0 0 -1 0 0 1 5

𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠3 𝑠𝑠4 𝑅𝑅1 𝑅𝑅2 𝑅𝑅3 Solution

1 -2-6M 7-16M M 0 M 0 0 0 -18M
0 -2 3 0 0 0 1 0 0 3
0 4 5 -1 0 0 0 1 0 10
0 6 7 0 1 0 0 0 0 3
0 4 8 0 0 -1 0 0 1 5

𝑧𝑧 − (2 + 6𝑀𝑀)𝑥𝑥1 + (7 − 16𝑀𝑀)𝑥𝑥2 + 𝑀𝑀𝑠𝑠1 + 𝑀𝑀𝑠𝑠4 = −18𝑀𝑀

(c)
𝑧𝑧 − 3𝑥𝑥1 − 6𝑥𝑥2 − 𝑀𝑀𝑅𝑅3 = 0
𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠2 𝑠𝑠3 𝑠𝑠4 𝑅𝑅3 Solution
1 -3 -6 0 0 0 -M 0
0 1 2 1 0 0 0 5
0 6 7 0 1 0 0 3
0 4 8 0 0 -1 1 5

𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠2 𝑠𝑠3 𝑠𝑠4 𝑅𝑅3 Solution

1 -3+4M -6+8M 0 0 -M 0 5M
0 1 2 1 0 0 0 5
0 6 7 0 1 0 0 3
0 4 8 0 0 -1 1 5

𝑧𝑧 − (3 − 4𝑀𝑀)𝑥𝑥1 − (6 − 8𝑀𝑀)𝑥𝑥2 − 𝑀𝑀𝑠𝑠4 = 5𝑀𝑀

(d)
𝑧𝑧 − 4𝑥𝑥1 − 6𝑥𝑥2 − 𝑀𝑀𝑅𝑅1 − 𝑀𝑀𝑅𝑅2 − 𝑀𝑀𝑅𝑅3 = 0
𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠4 𝑅𝑅1 𝑅𝑅2 𝑅𝑅3 Solution
1 -4 -6 0 0 -M -M -M 0
0 -2 3 0 0 1 0 0 3
0 4 5 -1 0 0 1 0 10
0 4 8 0 -1 0 0 1 5

𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠4 𝑅𝑅1 𝑅𝑅2 𝑅𝑅3 Solution

1 -4+6M -6+16M -M -M 0 0 0 18M
0 -2 3 0 0 1 0 0 3
0 4 5 -1 0 0 1 0 10
0 4 8 0 -1 0 0 1 5

𝑧𝑧 − (4 − 6𝑀𝑀)𝑥𝑥1 − (6 − 16𝑀𝑀)𝑥𝑥2 − 𝑀𝑀𝑠𝑠1 − 𝑀𝑀𝑠𝑠4 = 18𝑀𝑀

(e)
𝑧𝑧 − 3𝑥𝑥1 − 2𝑥𝑥2 − 𝑀𝑀𝑅𝑅1 − 𝑀𝑀𝑅𝑅3 = 0
𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠4 𝑅𝑅1 𝑅𝑅3 Solution
1 -3 -2 0 -M -M 0
0 -2 3 0 1 0 3
0 4 8 -1 0 1 5

𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠4 𝑅𝑅1 𝑅𝑅3 Solution

1 -3+2M -2+11M -M 0 0 8M
0 -2 3 0 1 0 3
0 4 8 -1 0 1 5

𝑧𝑧 − (3 − 2𝑀𝑀)𝑥𝑥1 − (2 − 11𝑀𝑀)𝑥𝑥2 − 𝑀𝑀𝑠𝑠4 = 8𝑀𝑀

Question 3
Consider the following Linear Programming:
𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 𝑧𝑧 = 3𝑥𝑥1 + 2𝑥𝑥2
Subject to
4𝑥𝑥1 − 𝑥𝑥2 ≤ 8
4𝑥𝑥1 + 3𝑥𝑥2 ≤ 12
4𝑥𝑥1 + 𝑥𝑥2 ≤ 8
𝑥𝑥1 ≥ 0, 𝑥𝑥2 ≥ 0
(a) Show that the associated simplex iterations are temporarily degenerate.
(b) Verify the results by solving the problem graphically.
(a)
4𝑥𝑥1 − 𝑥𝑥2 + 𝑠𝑠1 =8
4𝑥𝑥1 + 3𝑥𝑥2 + 𝑠𝑠2 = 12
4𝑥𝑥1 + 𝑥𝑥2 + 𝑠𝑠3 = 8
𝑥𝑥1 , 𝑥𝑥2 , 𝑠𝑠1 , 𝑠𝑠2 , 𝑠𝑠3 ≥ 0
Basic 𝑧𝑧 𝑥𝑥1 𝑥𝑥2 𝑠𝑠1 𝑠𝑠2 𝑠𝑠3 Solution
z 1 -3 -2 0 0 0 0
𝑠𝑠1 0 4 -1 1 0 0 8
𝑠𝑠2 0 4 3 0 1 0 12
𝑠𝑠3 0 4 1 0 0 1 8

Basic 𝑥𝑥1 Solution Ratio

𝑠𝑠1 4 8 8� = 2
4
𝑠𝑠2 4 12 12� = 3
4
𝑠𝑠3 4 8 8� = 2
4
since two variables have the same minimal ratio with the entering variable, then this
system is degenerate.
(b)
Question 4
A company manufactures two products, A and B. The unit revenues are $2 and $3,
respectively. Two raw materials, M1 and M2, used in the manufacture of the two
products have daily availabilities of 8 and 18 units, respectively. One unit of A uses 3
units of M1 and 2 units of M2, and 1 unit of B uses 2 units of M1 and 6 units of M2.
(a) Determine the dual prices of M1 and M2 and their feasibility ranges.
(b) Suppose that 4 additional units of M1 can be acquired at the cost of 30 cents
per unit. Would you recommend the additional purchase?
(c) What is the most the company should pay per unit of M2?
(d) If M2 availability is increased by 5 units, determine the associated optimum
revenue.
(a)
Decision Variables
𝑥𝑥1 → the number of A manufactured
𝑥𝑥2 → the number of B manufactured
Objective Function
𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 2𝑥𝑥1 + 3𝑥𝑥2
Constraints
𝑥𝑥1 ≥ 0, 𝑥𝑥2 ≥ 0, 3𝑥𝑥1 + 2𝑥𝑥2 ≤ 8, 2𝑥𝑥1 + 6𝑥𝑥2 ≤ 18
Optimum Solution
𝑥𝑥1 = 0.857143, 𝑥𝑥2 = 2.71429
2(0.857143) + 3(2.71429) = $9.85714
Dual prices apply to M1 and M2 feasibility ranges within which the optimum is the
intersection of M1 and M2 in the first quadrant. The main assumption is that M1 and
M2 are changed one at a time, while keeping the other constraint unchanged. For this
to occur, M1 allowable change must coincide with point (0, 3) as the minimum and
point (9, 0) as the maximum. Similarly, M2 range coincides with point (2.667, 0) as the
minimum and point (0, 4) as the maximum.
M1:
Corner Point M1 z Example Calculations
(0, 3) 6 9 M1 = 3 × 0 + 2 × 3 = 6
(9, 0) 27 18 z = 2 × 9 + 3 × 0 = 18
18−9
M1 dual price = = .42857/unit, Range 6 ≤ 𝑀𝑀1 ≤ 27
27−6
M2:
Corner Point M2 z Example Calculations
(2.667, 0) 5.334 5.334
(0, 4) 24 12
5.334−12
M2 dual price = 5.334−24 = .35712/unit, Range 5.334 ≤ 𝑀𝑀2 ≤ 24
(b)
Dual price M1 = $.43 in the range 6 ≤ 𝑀𝑀1 ≤ 12.
Revenue increase = . 43 × 4 = $1.72, which is larger than increase in
cost = . 3 × 4 = $1.2, so the purchase is recommended
(c)
Dual price M2 = $0.35712 in the range 5.334 ≤ 𝑀𝑀2 ≤ 24.
Any unit cost less than the dual price is acceptable.
(d)
M2 is increased from 18 to 23, which falls in the feasibility range. The increase in
revenue is 5 × .35712 = $1.7856.
The new optimal solution is 9.85714 + 1.7856 = $11.64274.