Scribd
Upload
105 views

CS 525: Linear Programming 1993 Final Exam Solution

This document contains solutions to exam problems involving linear programming. It first finds the maximum value and optimal solution to a maximization problem. It then solves a parametric linear program for different values of the parameter t and provides a summary table of the results. Finally, it evaluates several claims about linear programming, determining whether they are true or false and providing brief justifications.

Uploaded by

Abdesselem Boulkroune
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PS, PDF, TXT or read online on Scribd
105 views

CS 525: Linear Programming 1993 Final Exam Solution

This document contains solutions to exam problems involving linear programming. It first finds the maximum value and optimal solution to a maximization problem. It then solves a parametric linear program for different values of the parameter t and provides a summary table of the results. Finally, it evaluates several claims about linear programming, determining whether they are true or false and providing brief justifications.

Uploaded by

Abdesselem Boulkroune
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PS, PDF, TXT or read online on Scribd
You are on page 1/ 4

CS 525 : Linear Programming

1993 Final Exam Solution

1.
max x21 + x1 x2 x22 4x1 + x2
2 2

s.t. x1 + x2  2
x1 ; x2  0
What is the maximum value and where is it attained ?
Solution: This is a maximization problem.
max x21 + x1 x2 x22 4x1 + x2  min x21 x1 x2 + x22 + 4x1 x2
2 2 2 2

The Lagrangian is given by

L(x; u) = x21 x1 x2 + x22 + 4x1 x2 u(x1 + x2 2)


2 2

so that  
rx L(x; u) = x1 x2 u + 4
x1 + x2 u 1
with corresponding LCP(M; q)
2 3 2 3
x1 x1 x2 u + 4
0  4 x2 5 ? 4 x1 + x2 u 1 50
u x1 + x2 2
Using this to de ne M and q we get the following steps.
>> M=[1 -1 -1;-1 1 -1;1 1 0];
>> q =[4 -1 -2]';

>> H = lemketbl(M,q);
z1 z2 z3 1
--------------------------------------------
w1= | 1.0000 -1.0000 -1.0000 4.0000
w2= | -1.0000 1.0000 -1.0000 -1.0000
w3= | 1.0000 1.0000 0.0000 -2.0000

1
>> H = addcol(H,[1 1 1]','z0',4);
z1 z2 z3 z0 1
-------------------------------------------------------
w1= | 1.0000 -1.0000 -1.0000 1.0000 4.0000
w2= | -1.0000 1.0000 -1.0000 1.0000 -1.0000
w3= | 1.0000 1.0000 0.0000 1.0000 -2.0000
>> H = ljx(H,3,4);
z1 z2 z3 w3 1
-------------------------------------------------------
w1= | 0.0000 -2.0000 -1.0000 1.0000 6.0000
w2= | -2.0000 0.0000 -1.0000 1.0000 1.0000
z0= | -1.0000 -1.0000 -0.0000 1.0000 2.0000
>> H = ljx(H,2,3);
z1 z2 w2 w3 1
-------------------------------------------------------
w1= | 2.0000 -2.0000 1.0000 0.0000 5.0000
z3= | -2.0000 0.0000 -1.0000 1.0000 1.0000
z0= | -1.0000 -1.0000 0.0000 1.0000 2.0000
>> H = ljx(H,3,2);
z1 z0 w2 w3 1
-------------------------------------------------------
w1= | 4.0000 2.0000 1.0000 -2.0000 1.0000
z3= | -2.0000 -0.0000 -1.0000 1.0000 1.0000
z2= | -1.0000 -1.0000 0.0000 1.0000 2.0000

Hence, x1 = 0; x2 = 2 and the optimal value z  = 0.


2. Solve the following parametric linear program for all values of t and give
a summary table of your results.
maximize (2 t)x1 4x2
x1 x2  1
s.t. x1  2
x1 ; x2  0
Solution : We convert the problem rst.
>> A =[-1 1;-1 0];
>> p=[-2 4]'; q =[1 0]';
>> b =[-1 -2]';
>> H = totbl(A,b,p);

2
x1 x2 1
---------------------------------
x3= | -1.0000 1.0000 1.0000
x4= | -1.0000 0.0000 2.0000
---------------------------------
z= | -2.0000 4.0000 0.0000
>> H = addrow(H,[q' 0],'z0',3);
x1 x2 1
---------------------------------
x3= | -1.0000 1.0000 1.0000
x4= | -1.0000 0.0000 2.0000
z0= | 1.0000 0.0000 0.0000
---------------------------------
z= | -2.0000 4.0000 0.0000
>> %
>> % If t >= 2 then this is an optimal tableau
>> % An optimal solution is (x1, x2) =(0, 0)
>> % Note: The optimal value is -(0t + 0) = 0
>> %
>> % If t < 2
>> H = ljx(H,1,1);
x3 x2 1
---------------------------------
x1= | -1.0000 1.0000 1.0000
x4= | 1.0000 -1.0000 1.0000
z0= | -1.0000 1.0000 1.0000
---------------------------------
z= | 2.0000 2.0000 -2.0000
>> % If -2<= t <= 2 then this is an optimal tableau
>> % An optimal solution is (x1, x2) =(1, 0)
>> % Note: The optimal value is -(t - 2) = -t + 2
>> % If t < -2
>> H = ljx(H,2,2);
x3 x4 1
---------------------------------
x1= | 0.0000 -1.0000 2.0000
x2= | 1.0000 -1.0000 1.0000
z0= | 0.0000 -1.0000 2.0000
---------------------------------
z= | 4.0000 -2.0000 0.0000
>> % If t <= -2 then this is an optimal tableau
>> % An optimal solution is (x1, x2) =(2, 1)
>> % Note: The optimal value is -(2t) = -2t

3
Summary
Parameter t Range Minimum Value z (t) Solution Point x(t)
t2 0 (0; 0)
2t<2 2 t (1; 0)
t 2 2t (2; 1)
3. For each claim below, state rst whether it is true or false. Then in no
more two sentences or a simple 0ne- or two-line counterexample ( for the
false case), justify your answer. An unjusti ed \True" or \False" answer
gets no credit.
(a) If a primal LP is infeasible then the dual LP is also infeasible.
Answer : False min x s.t. x  1; x  0, max u s.t. u  1; u  0
(b) A solvable quadratic program must have a vertex solution.
Answer : False min x2 s.t. 1  x  1
(c) The problem minx kAx bk1 s:t:x  0 always has a solution where
A is any m  n matrix.
Answer : True Feasible LP with objective bounded below by zero.
(d) Let A = A0; then the linear program :
max
x
b0 x s.t. Ax  b; x  0
is solvable if it is feasible.
Answer : True Because dual is feasible:
min
u
b0 u s.t. A0 u = Au  b; u  0

You might also like