Originally posted by: cjg

I'm running into a situation with CPLEX/OPL where some of the constraints in a model are being violated, but the solution is being reported as feasible and optimal. I'm working with a linear binary MIP. When I turn all float variables into ints and switch to CP solver I get correct solutions. I'm using OPL 6.3. I'm an OPL newbie and would greatly appreciate any help.

Thanks,
Chris
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

There is an explanation of this sort of problem and tips for dealing with it in one of the CPLEX FAQs at http://www-01.ibm.com/support/docview.wss?uid=swg21399984. I found this particular FAQ by accessing the main CPLEX support page at http://www-947.ibm.com/support/entry/portal/Overview/Software/WebSphere/IBM_ILOG_CPLEX and searching on "big M" since that construct is the usual source of this sort of problem.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: cjg

Hi Mary,

Thanks so much! This was exactly the problem, and I was using an M value that was huge. I made it smaller and this fixed the problem. I really appreciate your reply.

Chris
#DecisionOptimization
#OPLusingCPLEXOptimizer