[Coin-lpsolver] CLP: How do deal with a solution out of bounds?

John J Forrest jjforre at us.ibm.com
Mon Oct 9 09:11:08 EDT 2006


Thomas,

You should have had a message saying the problem was infeasible and you 
should be able to check the status i.e. model.status()!=0 or 
!model.isProvenOptimal().   The simplex algorithm always returns a 
solution but will say if it is infeasible.

John Forrest



"Thomas Witkowski" <s7118188 at inf.tu-dresden.de> 
Sent by: coin-lpsolver-bounces at list.coin-or.org
10/03/2006 02:38 PM
Please respond to
s7118188 at inf.tu-dresden.de


To
coin-lpsolver at list.coin-or.org
cc

Subject
[Coin-lpsolver] CLP: How do deal with a solution out of bounds?






Hello,

I'm new to programming with CLP and I have a problem where I get a
solution which is out of the bounds I have set before. I want to solve a
set of equations without a min/max function. The coefficient-matrix is as
follows:

1      1      0      0      0      0
0      0      1      1      0      0
0      0      51.32  100    -1     0
0      0      51.92  51.02  -1     0
48.68  48.08  0      0      0      -1
0      48.98  0      0      0      -1
0      0      51.92  51.32  -1     0
0      48.68  0      0      0      -1

The first two rows have to be equal to 1 the following four rows have to
be equal to 0 and the both last can range between -100 and 0. The bounds
for the first four columns are 0 - 1 and for the others 0 - 100. Solving
this system with the primal-algorithm I get a solution for the fourth row
to be -0.00365258. But how can this be? The row should be equl to 0! From
an other programm I know that this equations do not have a solution, so
why I get one? Maybe some of you can help me with this problem? I would be
thankful!

Regards,

Thomas

_______________________________________________
Coin-lpsolver mailing list
Coin-lpsolver at list.coin-or.org
http://list.coin-or.org/mailman/listinfo/coin-lpsolver

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/clp/attachments/20061009/bbde691d/attachment.html>


More information about the Clp mailing list