Forum Discussion
Infeasible model after consecutive invocation
We have the following workflow implemented in Java with Xpress 9.2.2 8.14.5: 1. Build BCL MIP model 2. Fix subset of variables (xprbVar.fix(.)). This simplified model is referred to as model A 3...
First of all, I suggest to export the models to LP files and compare them, just to be sure that the changes are indeed what you suspect them to do. I.e., do XPRBprob.getXPRSprob().writeProb("stage.lp", "l").
I suggest to do that in three different points:
- For the original model (model1.lp)
- After fixing the variables and calling loadmat (model2.lp)
- After unfixing the variables and calling loadmat (model3.lp)
model1.lp and model3.lp should be the same.
Then, when analyzing model B (model3.lp) instead of setting PRESOLVE=0 you could play with the PRESOLVEOPS control to figure out whether this can be blamed easily on a particular reduction.
I am not sure I understood correctly: are you saying that model B is found infeasible only if it is solved right after model A in your application but when solved standalone then it is found feasible? In particular, if you solve the original model before even creating model B, is that found feasible?
Could you share the offending model? If not here, then maybe by email? Preferably as a file created by XPRBprob.getXPRSprob().saveAs(...) right before solving model B. That might tell whether there is something bad "left" from solving A in the model state.