Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 9

Indian Institute of Management Bangalore

QUANTITATIVE METHODS II
Quiz 3
Saturday, December 24th, 2005

Time : 105 minutes


Total No. of Pages : 9 Name ________________________
Total No. of Questions : 3 Roll No. ________________________
Section: _______________________

1. Answer all questions only in the space provided following the question.
2. Show all work and give adequate explanations to get credit.
3. You may use the backside of the pages for rough work only if needed. Do NOT attach
any rough work/sheets.
4. Clearly define all variables used in your formulations. Label all constraints.
Question I (25 Points):

Consider the following Binary Linear Integer Programming (BIP) problem:

(P) Minimize 7x1 + 3x2 + 2x3 + 4x4


s.t. 4x1 + 2x2 - x3 + x4 >= 3
4x1 + 3x2 + 4x3 - 2x4 >= 7
x1, x2, x3, x4 {0, 1}.

The above problem needs to be solved using branch-and-bound.

(i) State the LP relaxation to problem (P). [5 points]

Minimize 7x1 + 3x2 + 2x3 + 4x4


s.t. 4x1 + 2x2 - x3 + x4 >= 3
4x1 + 3x2 + 4x3 - 2x4 >= 7
0 x1 1, 0 x2 1, 0 x3 1, 0 x4 1

(ii) Solve (P) using branch-and-bound (b-and-b). Use the linear programming
relaxation. On the last page of this exam (appendix), partial information on some
linear programs along with their corresponding optimal solution is provided for
your benefit. Use a depth-first-search strategy to guide your search for the optimal
solution. While branching on a variable xi, set the branch xi = 0 towards the left
and xi = 1 towards the right. Also, while selecting a b-and-b node, choose the left
node before the right node. As part of your answer, draw clearly the branch-and-
bound tree and number the nodes in the order that you visit them. At each node i,
state the optimal solution, objective function value (LPi) and the current best-
known feasible value, f. State clearly the optimal solution found. Give adequate
explanation for each step. {Note that the information given in the appendix is only
partial. The rest you are required to figure out yourself}. [20 points]
Branch-and-Bound Tree

0 (LP0) = 7.0, x1=0.4, x2=1.0


x3 = 0.6, f =

x1=0 x1=1

(LP1) = , x1=1, x3=0.75


1 2
Fathomed Infeasible, f = (LP2) = 8.5,f =

x3 = 0 x3 = 1

3 x1=1, x2=1 x1=1,


Fathomed (LP3) = 10,f =10 4 x3=1

Fathomed (LP4) = 9,
f = 9.

The optimal solution is: x1 = 1, x3 = 1, x2 = 0, x4 = 0, Optimal objective function


value is 9.

Note that the optimal solution to the LP relaxation of (P3) = {P: x1=1, x3=0} is
obtained directly from the 4th LP problem given in appendix.

Note that for (P4) = {P:x1=1, x3=1}: its LP relaxation becomes

Minimize 3x2 + 4x4


s.t.
2x2 + x4 >= 0
3x2 2x4 >= -1
0 <= x2 <= 1, 0 <= x4 <= 1

Since x2 = 0, x4 = 0 satisfies the constraints for the above problem it is also optimal.
Question II (25 Points):

At a machine tool plant five jobs must be completed each day. The time it takes to do
each job depends on the machine used to do the job. If a machine is used at all, there is a
setup time required. Once a machine is setup, then it can be used to process one or more
jobs. The relevant time (in minutes) are given in Table 1 below. For those machine-job
combinations for which the time is not provided, it is understood that the machine cannot
process that job. Finally, machine 1 can be used only if both machine 3 and machine 4 are
used.

Table 1.
Machine
Setup
Time
Job 1 Job 2 Job 3 Job 4 Job 5
Machine 1 42 70 93 - - 30
Machine 2 - 85 45 - - 40
Machine 3 58 - - 37 - 50
Machine 4 58 - 55 - 38 60
Machine 5 - 60 - 54 - 20

a. Suppose that the companys goal is to minimize the sum of the setup and
machine operation times needed to complete all jobs. Formulate an appropriate
Linear Integer Programming model for this problem. Define the decision
variables carefully, and describe the purpose of each constraint carefully.
[20 points]
Decision Variables:

Yj = 1, implies that Machine i is setup,


= 0 otherwise, for i = 1,,5

Xij = 1 if job j is assigned to machine i,


= 0 otherwise.

Minimize 30Y1 + 40Y2 + 50Y3 + 60Y4 + 20Y5 + 42X11 + 70X12 + 93X13 + 10000X14
+ 10000X15 + 10000X21 + 85X22 + 45X23 + 10000X24 + 10000X25 + 58X31
+ 10000X32 + 10000X33 + 37X34 + 10000X35 + 58X41 + 10000X42 + 55X43
+ 10000X44 + 38X45 + 10000X51 + 60X52 + 10000X53 + 54X54 + 10000X55

Subject to:
X1j + X2j + X3j + X4j + X5j = 1 for j = 1,,5 {these constraints say that
each job must be completed, hence assigned to exactly 1 machine}.

Xij Yi for all i = 1,,5, j = 1,5. {these constraints say that job j
cannot be assigned to machine i, unless the machine is setup}.

Y1 Y3, Y1 Y4 {these 2 constraints enforce the condition that


machine 1 can be used only if both machine 3 and machine 4 are
used}.

Yi = {0, 1} for all i = 1,..5. Xij = {0, 1} for all i = 1,..5 and j = 1,..,5.

Note: for those jobs j which cant be processed by machine I, a high


cost of 10000 is assigned to ensure such an assignment does not take
place.

b. Consider the situation where, if Machine 4 is used to process Job 1, Job 3 and Job 5 then, due
to processing efficiency, 10 minutes can be saved. That is, the time required to process these
three jobs will be (60+58+55+38-10) minutes. However, no such savings is accrued if
Machine 4 is used to process two or less jobs. Show clearly how the formulation in a. can be
modified to accommodate such a condition. [5 points]

Let Z = 1 if Machine 4 processes all three jobs (Job 1, Job 2 and Job 5),
= 0 otherwise.

The New Objective Function = Old Objective Function 10Z.

In the constraint set, we add the constraints

Z <= X41
Z <= X42
Z <= X45
Question III (20 Points):

ABC Computers produces three different types of computers: Epic, Galaxy and Utopia. The
production of all computers is conducted on a complex and modern assembly line. The
production of an Epic requires 5 hours on the assembly line, a Galaxy requires 8 hours and a
Utopia requires 12 hours. The normal operation hours of assembly line are 170 per month.
The marketing and accounting departments have estimated that profits per unit for the three
types of computers are Rs.10000 for Epic, Rs.14400 for Galaxy and Rs.25200 for the Utopia.
The marketing department further reports that the demand is such that the firm can expect to
sell all the computers it can produce in the month. The company must meet the demand of
the North Eastern sales district for 5 Epic, 5 Galaxy and 8 Utopia computers. In addition the
Chairman of the company would also like the following targets to be met:

a) Limit overtime operations of the assembly line to 20 hours.


b) Make a profit of at least Rs. 400000
c) Meet the sales goal for each type of computers: Epic 10, Galaxy 12 and Utopia
10.

The management has assessed its priorities for the three targets. They feel that the priority
level for not exceeding overtime is the highest, that of achieving the desired profit comes
second and that of sales goals for each type of computer comes last. Given these
priorities, we wish to formulate and solve the problem of determining the optimal product
mix. Use the variables defined below.

xi : units of computer i produced, i = 1,2,3.

a. Formulate the Priority Level 1 problem (P1). State the objective function and
constraints clearly. Define any other variables required. [3 points]

Min d+1

s.t.
x1 >= 5 {Min production of Epic hard constraint}
x2 >= 5 {Min production of Galaxy hard constraint}
x3 >= 8 {Min production of Utopia hard constraint}
5x1 + 8x2 + 12x3 + d-1 d+1 = 190 {Goal constraint 1}
10000x1 + 14400x2 + 25200x3 + d-2 d+2 = 400000 {Goal constraint 2}
x1 + d-3 d+3 = 10 {Goal constraint 3}
x2 + d-4 d+4 = 12 {Goal constraint 3}
- +
x3 + d 5 d 5 = 10 {Goal constraint 3}

d+i, d-i>= 0 for i= 1,,5


Using SOLVER the following optimal solution was obtained for the P1 problem.
x1 = 5.0, x2 = 5.0, x3= 8.0 and all the other variables have zero value at optimality.
The optimal objective function value is zero. Answer the following questions:

b. Is the priority level 1 goal met? Explain. If not, how much is it missed by? [2 points]

Using solution, x1 = 5.0, x2 = 5.0, x3= 8.0 we find that the number of hours
utilized is 161. Hence, d+1 = 0 and priority goal 1 is met.

c. Is the priority level 2 goal met? Explain. If not, how much is it missed by? [2 points]

No. Using solution, x1 = 5.0, x2 = 5.0, x3= 8.0 we find that the profit made is
Rs 323600. Hence priority level 2 goal is missed by Rs 76,400.

d. Formulate the Priority Level 2 problem (P2) which ensures that subsequent solutions
will not degrade the solution already obtained in terms of the priority level 1 targets.
State the objective function and the constraints clearly. Define all variables used if not
already defined above. [4 points]

Min d-2

s.t.
x1 >= 5 {Min production of Epic hard constraint}
x2 >= 5 {Min production of Galaxy hard constraint}
x3 >= 8 {Min production of Utopia hard constraint}
5x1 + 8x2 + 12x3 + d-1 d+1 = 190 {Goal constraint 1}
10000x1 + 14400x2 + 25200x3 + d-2 d+2 = 400000 {Goal constraint 2}
x1 + d-3 d+3 = 10 {Goal constraint 3}
- +
x2 + d 4 d 4 = 12 {Goal constraint 3}
x3 + d-5 d+5 = 10 {Goal constraint 3}
+
d 1= 0 {Attainment of Goal constraint 1}
d+i, d-i>= 0 for i= 1,,5

e. Given that the optimal solution obtained for the P2 problem is:
x1 = 5.0, x2 = 5.0, x3 = 11.03, find the values of all other variables and find the
optimal objective function value for the P2 problem. [3 points]

Given the optimal solution, x1 = 5.0, x2 = 5.0, x3 = 11.03, we find that the hours
utilized is 197.36 and the profit obtained is Rs 399956. Hence, d+1 = 7.36 and d-2 = 44.
f. Formulate the Priority Level 3 problem (P3) which ensures that subsequent solutions
will not degrade the solution already obtained in terms of the priority level 1 and 2
targets. Assign weights to the penalties on meeting the targets according to the relative
profit function for each computer. State the objective function and the constraints clearly.
[4 points]
- - -
Min 10000d 3 + 14400d 4 + 25200d 5

s.t.
x1 >= 5 {Min production of Epic hard constraint}
x2 >= 5 {Min production of Galaxy hard constraint}
x3 >= 8 {Min production of Utopia hard constraint}
5x1 + 8x2 + 12x3 + d-1 d+1 = 190 {Goal constraint 1}
10000x1 + 14400x2 + 25200x3 + d-2 d+2 = 400000 {Goal constraint 2}
x1 + d-3 d+3 = 10 {Goal constraint 3}
x2 + d-4 d+4 = 12 {Goal constraint 3}
x3 + d-5 d+5 = 10 {Goal constraint 3}
+
d 1= 0 {Attainment of Goal constraint 1 from b.}
d-2 = 44 {Attainment of Goal constraint 2 from e.}
+ -
d i, d i>= 0 for i= 1,,5

g. Do you expect to be able to meet the priority level 3 goal? Explain. [2 points]

No. To meet priority level 3 goal, at least 10 Epic computers, 12 Galaxy


computers and 10 Utopia computers have to be produced. This will degrade
attainment of priority goal 1 constraint by 96, hence violating the constraint d+1 = 0.
Appendix:

LP Problems: Optimal Solution

1. Minimize 7x1 + 3x2 + 2x3 + 4x4 x1=0.4, x2=1, x3=0.6


x4 = 0
s.t. 4x1 + 2x2 - x3 + x4 >= 3 (LP) = 7.0
4x1 + 3x2 + 4x3 - 2x4 >= 7
0<= xi <= 1, i = 1,..,4

2. Minimize 7x1 + 3x2 + 2x3 + 4x4 Infeasible

s.t. 4x1 + 2x2 - x3 + x4 >= 3 (LP) =


4x1 + 3x2 + 4x3 - 2x4 >= 7
0<= xi <= 1, i = 1,..,4
x1 = 0

3. Minimize 7x1 + 3x2 + 2x3 + 4x4 x1=1, x3=0.75,


x2 = 0, x4 = 0
s.t. 4x1 + 2x2 - x3 + x4 >= 3 (LP) = 8.5
4x1 + 3x2 + 4x3 - 2x4 >= 7
0<= xi <= 1, i = 1,..,4
x1 = 1

4. Minimize 7x1 + 3x2 + 2x3 + 4x4 x1=1, x2=1.0,


x3 = 0, x4 = 0
s.t. 4x1 + 2x2 - x3 + x4 >= 3 (LP) = 10.0
4x1 + 3x2 + 4x3 - 2x4 >= 7
0<= xi <= 1, i = 1,..,4
x3 = 0

5. Minimize 7x1 + 3x2 + 2x3 + 4x4 x1=0.5, x2=1.0,


x3 = 1, x4 = 0
s.t. 4x1 + 2x2 - x3 + x4 >= 3 (LP) = 8.5
4x1 + 3x2 + 4x3 - 2x4 >= 7
0<= xi <= 1, i = 1,..,4
x3 = 1

You might also like