Originally posted by: Eumpfenbach

I stop my problem when I find the first feasible solution. I want reference the objective gap in the python API (something like "while gap > # do ...").

I've messed around with it for over a half hour so I just figured I'd ask. solution.quality_metric.objective_gap seems like where the information would be but it doesn't have a get() method.

If I use solution.get_quality_metrics() it gives me the gap but I am not sure how to reference just this one number, since this function seems to print a list.

So yeah, I just want the gap at the time the solver stops. Absolute or relative will both work fine.

Thanks,
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

The relative gap is available via

c.solution.MIP.get_mip_relative_gap()

The absolute gap is (for minimization)

c.solution.get_objective_value() - c.solution.MIP.get_best_objective()

#CPLEXOptimizers
#DecisionOptimization

Originally posted by: Eumpfenbach

Thank you.

I was using dir(cpx.solution) and missed MIP because it is listed before all the __ ... __, whereas all the other attributes are listed after.
#CPLEXOptimizers
#DecisionOptimization