BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, PILANI, HYDERABAD CAMPUS

ME F320 – ENGINEERING OPTIMIZATION

SEMESTER – II, 2021-22 Comprehensive Examination PART-B (Closed Book)
12/05/2022 (Thu) 2:00–5:00 p.m. Max. Duration: 120 min Max. Marks = 100 M

Note: There are total 5 questions printed on one side of this sheet. Start answering each question from a new page.
Clarity of presentation will carry its own marks. Please find Normal distribution table on the back side.

1. Siyaram’s store sells two types of shirts, Linen and Denim, at a profit of Rs. 300 and Rs. 800 per shirt
respectively. The store maintains two tailors, P and Q, for stitching the shirts. Tailor P can devote at least
7 hours per day and Tailor Q can devote at the most 14 hours per day. Tailors P and Q spend 2 hours and
5 hours respectively stitching a linen shirt, and 4 hours and 3 hours respectively stitching a denim shirt.
a. Formulate the LPP and solve it by using Two-phase method to find the optimum profit value.
b. Now, considering the realistic scenario of stitching the number of shirts per day as exact integers,
find the optimal integer solution using Gomory’s cutting plane method. [15 + 10 = 25 M]

2. There are three factories A, B and C, which supply goods to four dealers W X Y Z
W, X, Y and Z. The production capacities of these factories are 1000, 600 A 2 2 2 4
and 900 units per month, respectively. The requirements from the dealers B 3 5 3 2
are 800, 800, 500 and 400 units per month, respectively. The per unit C 4 3 2 1
returns (excluding transportation cost) are Rs. 8, 7, 9 at the three factories. The table gives the unit
transportation costs from the factories to the dealers. Using the least cost method for IBFS, determine the
optimal solution with the maximum profit value. [15 M]
Immediate Optimistic Most likely Pessimistic
3. A project network has the S.No. Activity
Predecessor time (a) time (m) time (b)
following activities:
1 A - 2 4 12
a. Determine the critical
2 B - 10 12 26
path for the project
network. 3 C A 8 9 10
b. Find the probability of 4 D A 10 15 20
completing the project 5 E A 7 7.5 11
within 25 to 35 days? 6 F B, C 9 9 9
[10 + 5 = 15 M] 7 G D 3 3.5 7
8 H E, F, G 5 5 5

4. A car wash facility has two bays for service and a total of four car spaces including the ones being serviced.
Following Markovian distribution, cars arrive at the facility at a rate of 3 per hour, while the service time
for each car has a mean of 30 minutes. (Write your answers up to 4 decimal places)
a. Draw the Markov-chain and find the steady state probabilities.
Based on this, determine the following:
b. Amount of business lost to competition per 8-hour day due to the limited system capacity.
c. Expected number of cars in the facility.
d. Expected waiting time of a car before receiving service.
e. Probability that the arriving car has to wait before receiving service.
f. Capacity of the facility that will ensure that, on the average, no more than 5% of the arriving cars are
turned away. [12 + 4*2 + 10 = 30 M]

5. Consider the following non-linear function f(X) = x13 + x23 − 3x1x2

a. Minimize f(X), subject to g(X): x1 + x2 = 3 & x1, x2 ≥ 0, using Lagrangean method.
b. Minimize f(X), subject to g(X): x1 + x2 ≤ 3 & x1, x2 ≥ 0, using KKT conditions. [6 + 9 = 15 M]
Table for Normal Distribution Function
 ME F320 – Engineering Optimization SOLUTION (Compre. Exam. Part-B)

Solution 1: [25 M] Adding this additional constraint at the bottom of optimal

1(A) Max. Z = 300x + 800x simplex table, the new table so obtained is:
1 2 Basic x1 x2 S1 s1 y RHS Ratio
s.t. 2x + 4x ≥ 7 Iteration Z 3100/3 0 0 800/3 0 11200/3
1 2
0 x2 5/3 1 0 1/3 0 14/3
5x + 3x ≤ 14 s1 enters S1 14/3 0 1 4/3 0 35/3
1 2 y leaves
y -2/3 0 0 -1/3 1 -2/3
x ,x ≥0
1 2
Max. Z = 300x + 800x
1 2 By using dual simplex
s.t. 2x + 4x - S + A = 7 R ⟶ R + 800R ; R ⟶ R + R ; R ⟶ R + 4 R ; R ⟶ (-3) R ;
1 2 1 1
Iteration Z 500 0 0 0 800 3200
5x + 3x + s = 14 1 x2 1 1 0 0 1 4
1 2 2 S1 2 0 1 0 4 9
x ,x , S , s , A ≥ 0 s1 2 0 0 1 -3 2
1 2 1 2 1
Hence, integer optimal solution is arrived with value of
Phase-1
Max. R = -A variables as: x = 0, x = 4; Max Z = 3200
1 2
1
s.t. 2x + 4x - S + A = 7
1 2 1 1 Solution 2: [15 M]
5x + 3x + s = 14
1 2 2
Profit = return – transport cost 
x ,x ,S ,s ,A ≥0
1 2 1 2 1 Profit matrix =
Basis x1 x2 S1 s2 A1 RHS Ratio
Iteration 0 R 0 0 0 0 1 0
W X Y Z
A1 2 4 -1 0 1 7 A 6 6 6 4
s2 5 3 0 1 0 14
R ⟶R - R ; B 4 2 4 5
Iteration 1
x 2 enters
R
A1
-2
2
-4
4
1
-1
0
0
0
1
-7
7 7/4
C 5 6 7 8
A 1 leaves s2 5 3 0 1 0 14 14/3
R ⟶ R + R ; R ⟶ R /4 ; R ⟶ R - 3R /4 ; The problem is of maximization type. Convert this
Iteration 2 R
x2
0
1/2
0
1
0
-1/4
0
0
0
7/4
into minimization by subtracting all the entries by the
s2 7/2 0 3/4 1 35/4 highest entry 8.
Phase 2
W X Y Z
Eliminating the artificial variables and its column A 2 2 2 4
and bringing in the original objective function, B 4 6 4 3
Max Z = 300x + 800x2+ 0 s1 + 0 s2
1 C 3 2 1 0
Iteration 0
Basis
Z
x1
-300
x2
-800
S1
0
s2
0
RHS
0
Ratio Using Least cost method for the IBFS and applying
x2 1/2 1 -1/4 0 7/4 MODI’s optimality algorithm:
s2 7/2 0 3/4 1 35/4
R ⟶ R + 800R ;
Iteration 1 Z 100 0 -200 0 1400
S1 enters x2 1/2 1 -1/4 0 7/4 1.75/-
s2 leaves 0.25
s2 7/2 0 3/4 1 35/4 8.75/
0.75
R ⟶ R + 800R /3 ; R ⟶ R + R /3 ; R ⟶ 4R /3;
Iteration 2 Z 3100/3 0 0 800/3 11200/3
x2 5/3 1 0 1/3 14/3
S1 14/3 0 1 4/3 35/3
Hence, the optimal solution is:
x = 0, x = 4.6; Max Z = 3733.3
1 2
1(B) To obtain the integer valued solution, construct a
Gomory's fractional cut, with the help of x -row:
2
5/3 x + x + 1/3 s = 14/3
1 2 1
=> x + x - 4 = 2/3 - 2/3 x - 1/3 s
1 2 1 1
LHS is integer, while RHS is fractional part, so the
fractional cut will become 2/3 - 2/3 x - 1/3 s ≤ 0 Max Profit
1 1
= (200*6) + (800*6) +(600*4) +(500*7) +(400*8)
- 2/3 x - 1/3 s ≤ -2/3
1 1 = Rs. 15,100/-
Solution 3: [15 M] Solution 4: [30 M]
Queuing system: M/M/2: FCFS/4/∞ with
λ = 3 /hr, µ = 2 /hr, c = 2, N=4
State transition diagram

[2]
(a) Steady state probabilities:
𝑝 = 𝑝 ,𝑝 = 𝑝 ,𝑝 = 𝑝 ,𝑝 = 𝑝
=> 𝑝 = 1.5𝑝 , 𝑝 = 1.125𝑝 , 𝑝 = 0.843𝑝 , 𝑝 = 0.632𝑝

Since ∑ 𝑝 = 1 => 𝑝 (5.1) = 1

Hence,
state (n) Probability (pn)
0 0.1960
1 0.2941
Mean (PCT), 𝜇 = 29, Variance (PCT) = 6, 𝜎 = 2.445
2 0.2205
𝑥−𝜇 3 0.1652
𝑧= 4 0.1239 [10]
𝜎
Probability of project completion between 25 to 35 (b) Amount of business lost to competition per 8-hour
days= day due to limited system capacity
= 8*λ*p4 = 8*3*0.1239
P (25 < PCT < 35) = P ( <Z< ) = 2.973 car per day [2]
. .

= P (-1.6359 < Z < 2.4539) (c) Expected number of cars in the facility:
= P (Z < 2.4539) – P (Z < -1.6359) Ls = 1* 𝑝 + 2* 𝑝 + 3*𝑝 + 4*𝑝
= (1*0.294) + (2*0.2205) + (3*0.1652) + (4*0.1239)
= 0.9929 – (1- 0.9495) = 1.7263 cars [2]

= 0.9424 (d) Expected waiting time of a car before receiving

= 94.24 % service: Wq = ;
where Lq = 1*𝑝 + 2*𝑝 = 0.413
λ = λ (1 − 𝑝 ) = 2.6283
.
So, Wq = = = 0.1571 hr [2]
.

(e) Probability that the car has to wait before receiving

service = 𝑝 + 𝑝 = 0.3857 [2]

(f) Capacity of the facility that will ensure that, on the

average, no more than 5% of the arriving cars are
turned away: [10]
Solution 5: [15 M]
a)
L(X) = f(X) – λ g(X)
= x13 + x23 − 3x1x2 – λ (x1 + x2 – 3)
∇ L (X) = 0
3x12 − 3x2 – λ = 0---------------1
3x22 − 3x1 – λ = 0----------------2
x1 + x2 – 3 = 0  x2 = 3 - x1,
using above equation x1 = 3/2, x2 = 3/2
6𝑥 −3
Now Hessian matrix =
−3 6𝑥
9 −3
@ (3/2, 3/2), 𝐻 , =
−3 9
Since, M1 = 9, M2 = 27
H is positive definite.
Therefore, @ (3/2,3/2) f(X) is min=0 @ (3/2,3/2)

b)
L(X) = f(X) – λ [g(X) + S2]
= x13 + x23 − 3x1x2 – λ [(x1 + x2 – 3) + S2]
∇ L (x) = 0
 3x12 – 3x2 – λ = 0
3x22 – 3x1 – λ = 0
x1 + x2 – 3 + S2 = 0
λ*S =0
λ = 0, S ≠ 0 λ ≠ 0, S = 0
 Case I: (x1, x2) = (0, 0)  Case III:
Case II: (x1, x2) = (1, 1) (x1, x2) = (3/2, 3/2)
6𝑥 −3
Now Hessian matrix =
−3 6𝑥
Case-I
0 −3
λ = 0, (x1, x2) = (0, 0), S = 3 H(0, 0) =
−3 0
∵ M1 = 0, M2 = -9
Positive semi-definite. So, not exactly minima.
Case-II
6 −3
λ = 0, (x1, x2) = (1, 1), S =1 H(1, 1) =
−3 6
M1 = 6, M2 = 27
Thus, positive definite
 Min f(X) @ (1, 1) = -1
Case-III
S = 0, (x1, x2) = (3/2, 3/2), λ = 9/4
∵ In KKT condition
λ ≥ 0 (condition for maxima), Not a feasible solution

∴ Finally, Min f(X) @ (1, 1) = -1