Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

Indian Institute of Management @ Bangalore

PGP/PGPBA Term 4: Optimization


Homework - II

Homework Deadline: August 14th, 2022 (Sunday) @ 5PM.


Groups: Group number and team members as decided in class.
Peer Evaluation: Fill out the enclosed peer evaluation form, and submit
separately, at the end of the assignment.
Assignment Deliverables:

One professionally written homework solutions per group, inclusive of the following:

1. A cover page, with assignment title, names of team members, etc.


2. Problems to be presented in the order given in the homework.
3. All solutions must be typed and all math notation must be presented using “Equation
Editor” (or equivalent).
4. All additional information such as the GAMS files and graphs (with descriptions) must
be submitted.
5. All homework to be submitted on A4 pages, 12-point font, and 1.5 line spacing.

Instructions:

You can work in groups for this assignment, and only one final homework solutions document
must be submitted per group. Submit your assignment (including any codes) in a “.zip folder”,
with the name of the folder “Group_Number”, on the course website.
1. Use the simplex method for lower-upper bounded variables to solve the following linear
program, beginning with x1 at its upper bound, and x 2 , x3 at their lower bounds:

Maximize 2 x1 + x2 + 3x3
subject to: 3x1 + x2 + x3  12
− x1 + x 2  4
x2 + 2 x3  8
0  x1  3
0  x2  5
0  x3  4 .

2. Consider the following problem:

Maximize 3 x1 + 2 x 2
subject to: x1 + 2 x 2  8
x1 + x2  5
0  x1  3
0  x2  4 .

(a) Draw the feasible region, indicating slack variables associated with each constraint,
gradient of the objective function, and graphically identify the optimum.
(b) Beginning with x1 and x 2 respectively at their upper bounds, add artificial variables
and find an initial basic feasible solution to the problem.
(c) Starting with this basic feasible solution, determine the optimum using the lower-upper
bounded simplex method.

3. Show that the following two problems are equivalent:

P1: Maximize ct x P2: Maximize ct x


subject to: b1  Ax  b2 subject to: Ax + s = b2
x  0. x  0
0  s  b2 − b1 .

Use the simplex method for lower-upper bounded variables to solve the following problem
after reformulating it as shown above:

Maximize − 3 x1 + 4 x 2
subject to: 3  x1 + x 2  4
− 15  3 x1 − 5 x 2  2
(x1 , x2 )  0 .

4. Categorize each of the functions below as convex, concave, or nonconvex. Be sure to


provide adequate reasoning. Also, consider if a function is in more than one category.

(a) f ( x) = x 2 , x  .
(b) f ( x1 , x2 ) = x12 − x22 , ( x1 , x2 )  2

(c) f ( x1 , x2 ) = x12 + x22 , ( x1 , x2 )  2

(d) f ( x1 , x2 , x3 ) = 2 x1 + 3x2 − 4 x3 , ( x1 , x2 , x3 )  3

(e) f ( x) = c T x , x  n
, c n
, where c is a constant vector
(f) f ( x) = c x , x 
T n
, c n
, where c is not a constant vector
(g) f ( x1 , x2 , x3 ) = − x − 3x − 2 x23 − 4 x1 x2 + 2 x1 x3 + 4 x2 x3 , ( x1 , x2 , x3 ) 
2
1
2
2
3

5. Consider the following nonlinear program:

2
 3
 x1 −  + ( x2 − 5)
2
Minimize
 2
subject to: − x1 + x2  5
2 x1 + 3x2  11
( x1 , x2 )  0 .

(a) Draw the feasible region and contours of the objective function. Graphically identify
the optimum solution.
(b) How would the solution change if the objective function changes to a maximization
problem?

6. Consider the following nonlinear programming problem:

( x1 − 4) + ( x2 − 2 )
2 2
Minimize
subject to: 4 x12 + 9 x22  36
x12 + 4 x2 = 4

( x1, x2 )  X  ( x1, x2 ) 
2 x1  −3 .
(a) Draw the feasible region and contours of the objective function. Graphically identify
the optimum solution.
(b) How would the solution change if the objective function changes to a maximization
problem?

7. Let f ( x1 , x2 ) = 2 x1 + 6 x2 − 2 x12 − 3x22 + 4 x1 x2 , where ( x1 , x2 )  2


. Maximize f ( x1 , x2 )
using the line search method, beginning with the initial solution ( x1 , x2 ) = ( −1, − 1) . You
will need to write a small script (in any language of your choice) to generate the optimum
solution.

8. Consider Problem 5. Is this a convex programming problem for parts (a) and (b)? Solve
Problem 5(a) and 5(b) using the KKT conditions. Do the KKT conditions yield a globally
optimal solution in both cases? Why?

9. Consider the following nonlinear programming problem:

( x1 − 3) + ( x2 − 2 )
2 2
Minimize
subject to: x12 + x22  5
x1 + 2 x2  4
( x1 , x2 )  0 .

(a) Draw the feasible region and contours of the objective function. Graphically identify
the optimum solution.
(b) Solve the problem using the KKT conditions.
10. Find the shortest path from node 1 to node 9 using dynamic programming.

4 1
3 4
2 2 7 3
4 6
4
5
1 9
4 3
3
1 4
3 8
3
5
3
6

11. The sales manager for a publisher of college textbooks has six traveling salespeople to
assign to three different regions of the country. She has decided that each region should be
assigned at least one salesperson and that each individual salesperson should be restricted
to one region only. The following table gives the estimated increase in sales (in appropriate
units) in each region if it were allocated various numbers of salespeople. Use dynamic
programming to find the number of salespeople to be assigned to respective regions so as
to maximize sales.

Region
Salespeople
1 2 3
1 35 21 28
2 48 42 41
3 70 56 63
4 89 70 75

12. Solve the following problem using dynamic programming.

Maximize 2 x12 + 2 x2 + 4 x3 − x32


subject to: 2 x1 + x2 + x3 = 4
( x1 , x2 , x3 )  0.

You might also like