I have a binary LP problem, about 9100 binary variables and 8500 constraints after presolving (set to "Aggressive"). I have no objective function, and there is a decent chance that the model is infeasible (although I hope to find a solution). I have the latest instance running for a 11 days without finding a feasible solution. Currently 3415602 explored and 454289 unexplored nodes.
I was wondering if there are any settings that could help speed up this computation. Tuning is not able to give any improvements, since no feasible solutions are found in any trial. Currently I have:
I turned up the Heuristics after running a few weeks at the default without finding anything. I have also tried changing some of the priority levels for the variables, which fall into 4 main types (I can give more details on this if it's relevant).
I have seen some general information saying it may help to use variable probing and node presolve, but I cannot find information on these settings in Gurobi.
Please sign in to leave a comment.