GAMS World Forum

I see. Debugging models is hard. What you need to do is to take be "better" solution you get by just providing the .l (or leave this out because it has no effect) and look at this solution. Explain the solution to yourself and detect the flaws in this solution. Next look at the equation th...

I think you need to upload something that lets us reproduce the error. I can't do this from the code snippets you have pasted.

-Michael

Sure, if you provide the Excel files to see what's going on.

-Michael

I have no idea what you are asking. The attached GAMS model runs though nicely. I don't find any .fx, I only find some .l in your code. They have no effect (since a MIP solver usually ignores the starting point if not instructed otherwise). If I turn the .l into .fx I still get a good solution but w...

You have a Mac with an M1 or M2 chip but the GAMS 34 distribution is Intel only. So you would need an Intel Python running under Rosetta 2 on you machine. The native M1/M2 port for GAMS was introduced with GAMS 43 (see https://www.gams.com/latest/docs/RN_43.html#g4310_PLATFORMS). So many good reason...

You don't have a bi-level optimization problem. You have a two-stage stochastic program! The GAMS compiler gives you good help: 171 solve project max z use EMP scenario dict; **** $790,257 **** 257 Solve statement not checked because of previous errors **** 790 Symbol dimension mismatch **** The fol...

and it starts reporting solutions, which I've interpreted, rightly or wrongly, as the model finding solutions with the same objective value. This is wrongly assumed. If you would have send the entire width of the log I could have shown you the difference in incumbent values. With what you sent I ca...

Not clear what problem you describe. When I run the model you uploaded I get: --- Scenario s0: Records read: 4650 --- Solution statistics for scenario s0 --- Unmatched Records 0 --- Solver Status 1 Normal Completion --- Model Status 1 Optimal --- Objective Value 2.80346e+09 --- Resource Usage 281.79...

Hi ,not sure I understand everything what you write here. Degeneracy usually means (see https://glossary.informs.org/second.php?page=D.html#Degeneracy) that we have multiple solution of the same value. You describe that Cplex "always identifies several hundred solutions prior to solving". ...

UCM_h.log This all looks very much like the cluster is killing your jobs. GAMS runs the solver as an executable (communicating the model instance via scratch files) so even if the solver dies, the GAMS job should nicely continue and write proper lst files. Since this does not happen (and since the ...