Originally posted by: cplexforMatlabuser

Hello,

I use the CPLEX Optimization toolbox (Academic Initiative Version 12.2) for MATLAB to solve a mixed integer linear optimzation problem (with the help of the toolbox function cplexmilp). I modified my model that previously got solved and now I get an error message "Row 'c130' infeasible, all entries at implied bounds". What exactlly does this mean? Does this mean that row 130 in the inequality constraint matrix Aineq that I pass to cplexmilp is infeasible? How could I trace down the error?
I would be grateful for any kind of hint!
Furthermore I would like to know, whether it is good to specify upper bounds, even if they are already implied by the inequality constraints of Aineq? I would suppose that this makes the model solving faster, is that right?
Thanks a lot,

Amy
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

The error message means that CPLEX proved your problem to be infeasible in presolve.
Presolving was able to fix a bunch of variables and these fixings rendered the row named c130 (which is the 130th constraint if you don't use explicit row names) infeasible.
To investigate this issue you can use the conflict refiner:

• Export your problem into a .sav file before solving it.
• Use the interactive optimizer to analyze the conflict:

read <your problem>.sav
conflict
display conflict all

This will list a minimal set of conflicting constraints/bounds. From this set you should be able to tell what is wrong with your model.

Yes, in general it is a good idea to specify bounds for variables, even if they are implied by constraints.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: John Cui

Daniel said is correct, and also, you can use Cplex class in matlab to call conflict feature of cplex.

cplex=Cplex();
cplex.Model.obj = ...;
cplex.Model.A = ...;
cplex.Model.lb = ...;
cplex.Model.ub = ...;
cplex.Model.lhs = ...;
cplex.Model.rhs = ...;
cplex.refineConflict();
cplex.writeConflict('model.clp')

John Cui
#CPLEXOptimizers
#DecisionOptimization