Wednesday, August 1, 2012

Poor feedback from solver

Here is an example where a global solver fails.  This is actually a run on NEOS. That something goes wrong we have to accept: that can happen. However the feedback is rather poor. As a modeler I have no clue what to do now.

I think that software should be be built in a much more defensive style. Good feedback especially when something is going wrong is underappreciated by many developers.

Note: this may be a NEOS related problem, I don’t know at this stage.





S O L V E S U M M A R Y

MODEL ds OBJECTIVE w
TYPE QCP DIRECTION MAXIMIZE
SOLVER LINDOGLOBAL FROM LINE 35

**** SOLVER STATUS 13 System Failure
**** MODEL STATUS 13 Error No Solution
**** OBJECTIVE VALUE NA

RESOURCE USAGE, LIMIT NA 1000.000
ITERATION COUNT, LIMIT NA 2000000000
EVALUATION ERRORS NA 0

LINDOGLOBAL Jul 14, 2011 23.7.3 LEX 27723.27726 LEG x86_64/Linux

LINDOGLOBAL Driver
Lindo Systems Inc, www.lindo.com

Lindo API version 6.1.1.588 built on Jun 29 2011 22:29:58

No solution returned
GAMS Rev 237 LEX-LEG 23.7.3 x86_64/Linux 08/01/12 06:01:26 Page 4
G e n e r a l A l g e b r a i c M o d e l i n g S y s t e m
Solution Report SOLVE ds Using QCP From line 35


**** SOLVER STATUS FILE LISTED BELOW
=C
=CLINDOGLOBAL Jul 14, 2011 23.7.3 LEX 27723.27726 LEG x86_64/Linux

=0LINDOGLOBAL Jul 14, 2011 23.7.3 LEX 27723.27726 LEG x86_64/Linux

=C
=C LINDOGLOBAL Driver
=C Lindo Systems Inc, www.lindo.com
=C
=CLindo API version 6.1.1.588 built on Jun 29 2011 22:29:58
Barrier Solver Version 6.0.0.114, Nonlinear Solver Version 3.15A
Platform Linux x64

Number of constraints: 102 le: 0, ge: 0, eq: 102, rn:
0 (ne:0)
Number of variables : 42 lb: 0, ub: 0, fr: 42, bx:
0 (fx:0)
Number of nonzeroes : 1083 density=0.0025(%)

Nonlinear variables : 42
Nonlinear constraints: 61
Nonlinear nonzeroes : 982+0

Starting global optimization ...

Number of nonlinear functions/operators: 1
EP_MULTIPLY

Starting GOP presolve ...

Pre-check unboundedness
**** SOLVER STATUS FILE LISTED ABOVE



Update: there seems to be a bug involved. Just for a fraction of second or so I saw a message:



Error: *** glibc detected *** /opt/gams/latest/gmsgenux.out: double free or corruption (out): 0x0000000009286c20 ***



It is noted that this is not the latest GAMS version, which is 23.9.

No comments:

Post a Comment