Birla Institute of Technology and Science, Pilani (Raj.

)
First Semester, 2023-24
MATH F212 (Optimization)
Mid Semester Examination (Closed Book)
Max. Marks: 70 Max. Time: 90 Minutes Date: OCT. 11, 2023, 11.00 am -12.30 pm
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Q.1 The manufacturer utilizes three resources: person-hours, machine hours, and cloth
material to produce two dress types, Type A and Type B. Type A dresses generate a profit
of ₹160 each, while Type B dresses yield a profit of ₹|180 each. The manufacturer’s daily
capacity allows for the production of either 50 Type A dresses or 20 Type B dresses in
terms of person- hours. When it comes to machine hours, the capacity permits the
production of 36 Type A dresses or 24 Type B dresses each day. The daily availability
of cloth material is limited but sufficient for 30 dresses of either type. Formulate this as
a linear programming problem to maximize profit and convert it into standard form. [6]

Q.2 Suppose that the feasible region of the LPP

Max 𝑐 𝑇 𝑥 subject to 𝐴𝑥 = 𝑏, is bounded,

𝑐1 𝑥1 𝑏1
where 𝐴 = [𝑎𝑖𝑗 ], 1 ≤ 𝑖 ≤ 𝑚, 1 ≤ 𝑗 ≤ 𝑛, 𝑐 = ( ⋮ ) , 𝑥 = ( ⋮ ) and 𝑏 = ( ⋮ ).
𝑐𝑛 𝑥𝑛 𝑏𝑚

Show that the optimal solution of this LPP always occur at one of the feasible region’s
extreme point. [10]

Q.3 Solve the following LPP by using the two-phase method:

Maximize z = 𝑥1 + 2𝑥2 + 3𝑥3 − 𝑥4

subject to 𝑥1 + 2𝑥2 + 3𝑥3 ≥ 15
2𝑥1 + 𝑥2 + 5𝑥3 = 20
𝑥1 + 2𝑥2 + 𝑥3 + 𝑥4 = 10

𝑥1 , 𝑥2, 𝑥3 , 𝑥4 ≥ 0 [14]

Q.4 Write the dual of the following primal problem

Min z = 5𝑥1 + 6𝑥2

subject to 𝑥1 + 𝑥2 ≥ 2
4𝑥1 + 𝑥2 ≥ 4
𝑥1 , 𝑥2 ≥ 0

Solve primal problem by dual simplex method. [12]

Q.5 Solve the LPP by Revised Simplex method

Max z = 𝑥1 + 2𝑥2
Subject to 𝑥1 + 𝑥2 ≤ 3
𝑥1 + 2𝑥2 ≤ 5
3𝑥1 + 𝑥2 ≤ 6
𝑥1 , 𝑥2 ≥ 0 [10]

P.T.O.
Q.6 The optimal table of the following LP

Minimize z = 2x1 + x2
Subject to 3x1 + x2 = 3
4x1 + 3x2 ≥ 6
x1 + 2 x2 ≤ 3
x1, x2 ≥ 0
by solving M-method is

Basis Z x1 x2 s1 R1 R2 s3 Solution

1 1 2 1 12
z 0 0  M M 0
5 5 5 5

0 1 3 1 3
x1 1 0  0
5 5 5 5

0 3 4 3 6
x2 0 1   0
5 5 5 5

s3 0 0 0 1 1 -1 1 0

Suppose one activity (𝑥3 ) is added to the given LPP with coefficients (2, 1, 1)T for constraints
and its coefficient in the objective function is 3. Use post optimal analysis to find the solution
of new LPP. [6]

Q.7 Find the initial basic feasible solution for the following transportation problem by VAM.

D1 D2 D3 D4
S1 1 5 3 3 32
S2 3 7 1 2 15
S3 0 2 7 3 12
21 12 11 17

This solution is degenerate or not. Justify your answer. [12]

****END****
Q 1. The manufacturer utilizes three resources: person-hours, machine hours, and cloth material to
produce two dress types, Type A and Type B. Type A dresses generate a profit of |160 each,
while Type B dresses yield a profit of |180 each. The manufacturer’s daily capacity allows
for the production of either 50 Type A dresses or 20 Type B dresses in terms of person-hours.
When it comes to machine hours, the capacity permits the production of 36 Type A dresses or
24 Type B dresses each day. The daily availability of cloth material is limited but sufficient for
30 dresses of either type. Formulate this as a linear programming problem to maximize profit
and convert it into standard form.
Solution. Let x1 type A and x2 type B dresses are produced. Then, our objective is

Maximize z = 160x1 + 180x2 . [1 Mark]

Since 50 type A dresses are produced daily, the person-hours required to produce 1 type A dress
is 1/50. Thus person-hours taken to produce x1 type dresses is x1/50.
Similarly, since 20 type B dresses are produced daily, the person-hours required to produce 1
type B dress is 1/20. Thus person hours taken to produce x2 type dresses is x2/20. Hence,
x 1 x2
+ ≤ 1 ⇒ 2x1 + 5x2 ≤ 100. [1 Mark]
50 20

Regarding machine hours, the capacity permits the production of 36 Type A dresses or 24 Type
B dresses each day. Thus
x1 x2
+ ≤ 1 ⇒ 2x1 + 3x2 ≤ 72. [1 Mark]
36 24
Since the daily availability of cloth material is limited but sufficient for 30 dresses of either type,
x1 x2
+ ≤ 1 ⇒ x1 + x2 ≤ 30. [1 Mark]
30 30
Further, we have the non-negativity constraints:

x1 , x2 ≥ 0. [1 Mark]

The standard form of the LPP is as follows:

Maximize z = 160x1 + 180x2

subject to 2x1 + 5x2 + x3 = 100
2x1 + 3x2 + x4 = 7
x1 + x2 + x5 = 30
x1 , x2 , x3 , x4 , x5 ≥ 0. [1 Mark]

Q 2. Suppose that the feasible region of the LPP

max cT x subject to Ax = b,
     
c1 x1 b1
 ..   ..   .. 
is bounded, where A = [aij ], 1 ≤ i ≤ m, 1 ≤ j ≤ n, c =  .  , x =  .  and b =  .  .
cn xn bm
Show that the optimal solution of this LPP always occurs at one of the feasible region’s extreme
points.
 Solution: For the sake of contradiction, assume that the optimal solution x∗ does not occur
at a corner point. [1 Mark]
Let x(1) , x(2) , . . . , x(k) be the corner points of the feasible region. Then, we can write x∗ as
(k)
x∗ = λ1 x(1) + λ2 x(2) + . . . + λk

Here, λ1 , λ2 , . . . , λk are positive constants that sum to 1 (i.e., λ1 + λ2 + . . . + λk = 1). [2 Marks]

Now, let’s compute the objective function value at x∗ :
k
X
T ∗
c x = λi cT x(i)
i=1

≤ max{cT x(i) : 1 ≤ i ≤ k}
= cT x(r) [3 Mark]

for some 1 ≤ r ≤ k. Thus cT x∗ ≤ cT x(r) . [1 Mark]

However, since x∗ is optimal, cT x∗ ≥ cT x(r) . [1 Mark]
Thus cT x∗ = cT x(r) . [1 Mark]
Thus, the optimal value is attained at the corner point x(r) . This completes the proof. [1 Mark]

Q 3. Solve the following LPP by using the two-phase method:

Maximize x1 + 2x2 + 3x3 − x4

subject to x1 + 2x2 + 3x3 ≥ 15
2x1 + x2 + 5x3 = 20
x1 + 2x2 + x3 + x4 = 10
x1 , x2 , x3 , x4 ≥ 0

Solution:
On converting the constraints into equalities, we get

1x1 + 2x2 + 3x3 + 0x4 − 1s1 + 1R1 + 0R2 + 0R3 = 15

2x1 + 1x2 + 5x3 + 0x4 + 0S1 + 0R1 + 1R2 + 0R3 = 20
1x1 + 2x2 + 1x3 + 1x4 + 0S1 + 0R1 + 0R2 + 1R3 = 10
x1 , x2 , x3 , x4 , S1 , R1 , R2 , R3 ≥0

where s1 is a surplus variable and R1 , R2 , R3 are artificial variables. [2 Marks]

(Note: 1 Mark will be deducted if the non-negativity constraint is missed.)
Phase 1: The objective function is minimize z = R1 + R2 + R3 and the initial simplex tableau
is
Basic x1 x2 x3 x4 s1 R1 R2 R3 Sol. Ratio
z 4 5 9 1 −1 0 0 0 45
R1 1 2 3 0 −1 1 0 0 15 5
R2 2 1 5 0 0 0 1 0 20 4
R3 1 2 1 1 0 0 0 1 10 10
Since z3 − c3 is the positive most, x3 enters the basis, and the ratio is minimum for R2 , so it
leaves the basis. [2 Marks]
The next tableau is as follows:
Basic x1 x2 x3 x4 s1 R1 R2 R3 Sol. Ratio
z 2/5 16/5 0 1 −1 0 −9/5 0 9
R1 −1/5 7/5 0 0 −1 1 −3/5 0 3 15/7

x3 2/5 1/5 1 0 0 0 1/5 0 4 20

R3 3/5 9/5 0 1 0 0 −1/5 1 6 30/9

Since z2 − c2 is the positive most, x2 enters the basis, and the ratio is minimum for R1 , so it
leaves the basis. [2 Marks]
The next tableau is as follows:
Basic x1 x2 x3 x4 s1 R1 R2 R3 Sol. Ratio
z 6/7 0 0 1 9/7 −16/7 −3/7 0 15/7
x2 −1/7 1 0 0 −5/7 5/7 −3/7 0 15/7 −
x3 3/7 0 1 0 1/7 −1/7 2/7 0 25/7 25
R3 6/7 0 0 1 9/7 −16/7 −3/7 0 15/7 15/9

ince z5 − c5 is the positive most, s1 enters the basis, and the ratio is minimum for R3 , so it
leaves the basis. [2 Marks]
The next tableau is as follows:
Basic x1 x2 x3 x4 s1 R1 R2 R3 Sol. Ratio
z 0 0 0 0 0 −1 −1 −1 0
x2 1/3 1 0 5/9 0 0 −1/9 5/9 10/3
x3 1/3 0 1 −1/9 0 0 2/9 −1/9 10/3
s1 2/3 0 0 7/9 1 −1 4/9 7/9 5/3

Since the correct solution is optimal and all artificial variables are zero, the given LPP is feasible,
and we proceed to phase 2. [2 Marks]
Phase 2: The objective function for Phase 2 is maximize x1 + 2x2 + 3x3 − x4 and the initial
table is as follows:
Basic x1 x2 x3 x4 s1 Sol.
z 2/3 0 0 16/9 0 50/3
x2 1/3 1 0 5/9 0 10/3 [2 Marks]
x3 1/3 0 1 −1 /9 0 10 /3
s1 2/3 0 0 7/9 1 5/3
Since zj − cj ≥ 0 for all j, the current BFS, x1 = 0, x2 = 10/3, x
3 = 10/3 is optimal and the
optimal value is z = 50/3. [1+1 Marks]
Alternate Solution:
On converting the constraints into equalities, we get

1x1 + 2x2 + 3x3 + 0x4 − 1s1 + 1R1 + 0R2 = 15

2x1 + 1x2 + 5x3 + 0x4 + 0S1 + 0R1 + 1R2 = 20
1x1 + 2x2 + 1x3 + 1x4 + 0S1 + 0R1 + 0R2 = 10
x1 , x2 , x3 , x4 , S1 , R1 , R2 ≥0
where s1 is a surplus variable and R1 , R2 are artificial variables. [2 Marks]
(Note: 1 Mark will be deducted if the non-negativity constraint is missed).
Phase 1: The objective function is minimize z = R1 + R2 and the initial simplex tableau is
Basic x1 x2 x3 x4 s1 R1 R2 Sol. Ratio
z 3 3 8 0 −1 0 0 35
R1 1 2 3 0 −1 1 0 15 5
R2 2 1 5 0 0 0 1 20 4
x4 1 2 1 1 0 0 0 10 10
Since z3 − c3 is the positive most, x3 enters the basis, and the ratio is minimum for R2 , so it
leaves the basis. [2 Marks]
The next tableau is as follows:
Basic x1 x2 x3 x4 s 1 R 1 R2 Sol. Ratio
z −1/5 7/5 0 0 −1 0 −8/5 3
R1 −1/5 7/5 0 0 −1 1 −3/5 3 15/7

x3 2/5 1/5 1 0 0 0 1/5 4 20

x4 3/5 9/5 0 1 0 0 −1/5 6 30/9

Since z2 − c2 is the positive most, x2 enters the basis, and the ratio is minimum for R1 , so it
leaves the basis. [2 Marks]
The next tableau is as follows:
Basic x1 x2 x3 x4 s1 R1 R2 Sol.
z 0 0 0 0 0 −1 −1 0
x2 −1/7 1 0 0 −5/7 5/7 −3/7 15/7

x3 3/7 0 1 0 1/7 −1/7 2/7 25/7

x4 6/7 0 0 1 9/7 −9/7 4/7 15/7

Since the correct solution is optimal and all artificial variables are zero, the given LPP is feasible,
and we proceed to phase 2. [2 Marks]
Phase 2: The objective function for Phase 2 is maximize x1 + 2x2 + 3x3 − x4 and the initial
table is as follows:
Basic x1 x2 x3 x4 s1 Sol. Ratio
z −6/7 0 0 0 −16/7 90/7
x2 −1/7 1 0 0 −5/7 15/7 − [2 Marks]
x3 3/7 0 1 0 1/7 25/7 25
x4 6/7 0 0 1 9/7 15/7 15/9

Since z5 − c5 is the negative most value of zj − cj , s1 enters the basis, and the ratio is minimum
for x4 , it leaves the basis. The next tableau is as follows:
Basic x1 x2 x3 x4 s1 Sol.
z 2/3 0 0 16/9 0 50/3
x2 1/3 1 0 5/9 0 10/3 [2 Marks]
x3 1/3 0 1 −1/9 0 10/3
s1 2/3 0 0 7/9 1 5/3

Since zj − cj ≥ 0 for all j, the current BFS, x1 = 0, x2 = 10/3, x

3 = 10/3 is optimal and the
optimal value is z = 50/3. [1+1 Marks]