Scribd
Upload
48 views6 pages

IE 471 - Operations Research I Fall 2012 Homework 3 Solution Due Friday September 21, 2012

The document provides the solution to homework problems involving linear programming and the simplex method. It finds the optimal solution for two linear programming problems by initializing a starting basic feasible solution and iteratively increasing non-basic variables until an optimal solution is reached where no further improvements can be made. Corner point solutions and augmented forms are provided along with step-by-step workings to arrive at the optimal objective function values through multiple simplex iterations.

Uploaded by

smerino92
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
48 views6 pages

IE 471 - Operations Research I Fall 2012 Homework 3 Solution Due Friday September 21, 2012

The document provides the solution to homework problems involving linear programming and the simplex method. It finds the optimal solution for two linear programming problems by initializing a starting basic feasible solution and iteratively increasing non-basic variables until an optimal solution is reached where no further improvements can be made. Corner point solutions and augmented forms are provided along with step-by-step workings to arrive at the optimal objective function values through multiple simplex iterations.

Uploaded by

smerino92
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

IE 471 Operations Research I Fall 2012 Homework 3 Solution Due Friday September 21, 2012 Problem 1: a) We recall that

t the points where Z = constant correspond to a line, and when we vary the value of the constant, we obtain lines that are parallel to each other. With trial and error, one can pick an objective function value that is achievable within the feasible region e.g. Z = 50, corresponding to a line that goes through the points (10,0) and (0,5), can be achieved by several feasible solutions. As our goal is to maximize Z, we draw parallel lines above the line Z = 50 (that yield higher values of Z) until we are about to leave the feasible region. The last such line that can be drawn is the one that goes through the point (2,10) (intersection of the lines x1+ 2x2 = 18 and x1 + x2 = 12). Any line above it would not contain any feasible solution. This line corresponds to Z = 5*2 + 10*10 = 110. Optimal solution: (x1*, x2*) = (2,10) and Z = 110.

10 corner point solutions are circled (5 feasible, 5 infeasible)

b-d) CPF (0,0) (0,9) (2,10) (4,8) (7.2,0) Constraint boundary equations x1 = 0 x2 = 0 x1 = 0 -x1+2x2 = 18 -x1+2x2 = 18 x1+x2 = 12 x1+x2 = 12 5x1+2x2=36 5x1+2x2=36 x2 = 0 Adjacent CPF solutions (0, 9) (7.2, 0) (2, 10) (0, 0) (4,8) (0, 9) (2, 10) (7.2, 0) (0, 0) (2, 10) Z 0 90 110 100 36

CPF with highest Z is (2,10) so (2,10) is an optimal solution. e) Start at (0,0). Check Z at the adjacent CPFs (0, 9) and (7.2, 0): Z = 90 and 36 respectively. Both lead to a higher value of Z so (0, 0) is not optimal. Increasing x2 gives the best rate of improvement (10 vs. 5 if x1increases), so the simplex method goes to (0, 9) where Z = 90. Check Z at the adjacent CPFs (0, 0) and (2, 10): Z = 0 and 110 respectively. Only (2, 10) leads to a higher value of Z so go to (2,10) where Z = 110. Check Z at the adjacent CPFs (0, 9) and (4,8): Z = 90 and 100 respectively. Both lead to a lower value of Z so (2,10) is optimal. The simplex method stops.

Problem 2:

a) Augmented form: maximize subject to 30x1 + 30x2 x1 + x3 x2 + x4 40x1 + 50x2 + x5 20x1 + 16x2 + x6 x1, x2, x3, x4, x5, x6 =1 =1 = 60 = 24 "0

b)
! Nonbasic Variables Basic Variables x1, x 4 x 2, x 3 , x 5, x 6 A x4 , x5 x1, x 2 , x 3 , x 6 B x 5, x 6 x1, x 2 , x 3 , x 4 C x 3, x 6 ! x1, x 2 , x 4 , x 5 D ! x 2, x 3 ! x1, x 4 , x 5 , x 6 E ! ! x , x x F 1 2 3, x 4 , x 5, x 6 ! ! ! ! ! ! ! The coordinates of the CPFs were found by solving for the intersection of the two corresponding ! !and 40x1 + 50x2 = 60). constraint boundaries (e.g., point B is at the intersection of x2 = 1 ! !

CPF Solution (0,1) (1 4 ,1) 2 2 ( 3 , 3) (1, 1 4) (1,0) ! (0,0) !

BF Solution (0,1,1,0,10,8) 3 (1 4 ,1, 4 ,0,0,3) 2 1 1 (2 3 , 3 , 3 , 3 ,0,0) 3 (1, 1 4 ,0, 4 ,7.5,0) (1,0,0,1,20,4) (0,0,1,1,60,24)

c)

BF Solution A: Set x1 = x 4 = 0 and solve

x3 = 1 x2 = 1 ! 50 x 2 + x 5 = 60 " x 5 = 10 16 x 2 + x 6 = 24 " x 6 = 8
BF Solution B: Set x 4 = x 5 = 0 and solve

x2 = 1

! 40 x1 + 50 x 2 = 60 "
x1 + x 3 = 1 " x3 =

x1 =

1 4

3 4 20 x1 + 16 x 2 + x 6 = 24 "

x6 = 3

BF Solution C: Set x 5 = x 6 = 0 and solve

40 x1 + 50 x 2 = 60 20 x1 + 16 x 2 = 24 ! x1 + x 3 = 1 x2 + x4 = 1
From the last two equations, x 2 = x 3 =
2 1 and from the first two equations x 3 = x 4 = . 3 3

!
!

BF Solution D: Set x 3 = x 6 = 0 and solve

x1 = 1

! 20 x1 + 16 x 2 = 24 "

x2 =

1 4

3 4 40 x1 + 50 x 2 + x 5 = 60 " x2 + x4 = 1 " x4 =

x 5 = 7.5

BF Solution E: Set x 2 = x 3 = 0 and solve

x1 = 1 x4 = 1 ! 40 x1 + x 5 = 60 " 20 x1 + x 6 = 24 " x 5 = 20 x6 = 4

BF Solution F: Set x1 = x 2 = 0 and solve

x3 = 1 x4 = 1 ! x 5 = 60 x 6 = 24

! 3: Problem
We first write the problem in augmented form. maximize subject to

Z " 2 x1 " x 2
!

=0

x1 + x 2 + x 3 =4 4 x1 + x 2 + x 4 = 10 x1, x 2 , x 3 , x 4 #0

Initialization: x1 = x 2 = 0 " x 3 = 4, x 4 = 10, Z = 2 x1 + x 2 = 0 , is not optimal since the improvement rates are positive. Since it offers a rate of improvement of 2, choose to increase x1, ! which becomes the entering basic variable for Iteration 1. Given x 2 = 0 , the highest possible increase in x1 is found by observing that:

x 3 = 4 " x1 # 0 $ x1 % 4
!

! !

5 x 4 = 10 " 4 x1 # 0 $ x1 % 2

The minimum of these two bounds is 5/2, so x1 can be raised to 5/2 and x 4 = 0 leaves the basis. Using Gaussian elimination, we obtain:

"1 2 x2
3 4

x2 + x3 " 1 4 x4 = +1 4 x4 =

+1 2 x4 = 5
3 2 5 2

x1 + 1 4 x2 x1, x 2 , x 3 , x 4

#0

,0, 3 ,0) is not optimal since Z = 5+ 0.5x2 0.5x4 and the rate of improvement for Again ( 5 2 2 1/2 > 0. The highest possible increase in x 2 is found by observing that:

x 2 is

!
!

and

x 2 can be increased to 2. Consequently, x 3 becomes 0. By Gaussian elimination:


!

3 3 " x # 0 $ x2 % 2 2 4 2 5 1 x1 = " x 2 # 0 $ x 2 % 10 2 4 x3 =
2 3

x3 + 1 3 x4 = 6

1 x2 + 4 3 x! 3 " 3 x4 = 2 1 x1 "1 3 x3 + 3 x4 = 2 x1, x 2 , x 3 , x 4 #0

The current solution is optimal, since increasing x 3 or x 4 would decrease the objective value. Hence ! x * = (2,2,0,0) and Z * = 6 .

You might also like