Originated from
QueryCAT

Pb using the MATLAB's linprog solver

Hello, I have a question about the problem encountered with the MATLAB's linprog solver : is it possible to give me an example that illustrates the fact that linprog provides an infeasible solution while other solvers return a correct one ? Thank you for your reply. Regards.


Would you like to answer or comment?

Sign up for a free account, or sign in (if you're already a member).
  • 174 views
Share Send to a friend Watch Report
 
 

Posted Answers

Order by
 
165 thumbs up

I think, but not sure, that if you take A to be a matrix of very small numbers and the vector to b to be a vector of  big numbers, such that the different in scale will be something as 10^9, you'll get a wrong solution.


Posted 5 months ago ( permalink )
In reply to olivier's question
Rated as
#1 out of 2
0
0

Helpful?

line
line
line



 

Thank you for your help, but I did not manage to reproduce the problem. I have solved a large number of problems of this type :

f = rand(10, 1);
A = 1.0e-6 * rand(10, 4);
b = 1.06 * rand(4, 1);

and linprog always gives me a correct solution. I am working on MATLAB 6.5.1. Thank you for any other information.

Regards,

Olivier


Posted 5 months ago ( permalink )
In reply to eniv4's answer
Rated as
#2 out of 2
0
0

Helpful?

line
line
line



Sign in to participate

Got an answer for olivier? Would you like to comment on the posted answers, or vote for the one which you think is the best?

Sign up for a free account, or sign in (if you're already a member).

Feed - Subscribe to changes to this Q&A Blog
Copyright © 2006-2008, Yedda Inc. and respective copyright owners · CC License