Hi,
I am solving the MIQP model with CPLEX IDE 12.5. The converging speed is very slow.
Moreover, as I increase the size, the CPLEX have some difficulty to find a feasible solution .
Any setting will help?
Thanks.
Hi,
I am solving the MIQP model with CPLEX IDE 12.5. The converging speed is very slow.
Moreover, as I increase the size, the CPLEX have some difficulty to find a feasible solution .
Any setting will help?
Thanks.
Regarding speeding up the search, please include a log file of the current run; examining the node log may
shed light on the source of the slow performance.
Regarding finding a feasible solution, since an MIQP has a quadratic objective but no quadratic constraints, you
can try running CPLEX on the associated MILP until at finds at least one solution (or running to optimality if it
doesn't take long), then using that solution as an advanced start for your MIQP.
Hi,
It will be helpful if you could provide a model formulation or output an LP formatted file.
Further, are the integer variables Binary? are these constraints all linear? Is the feasible region convex? It shouldn't be hard to find a feasible solution for MIQP problem, I think.
Bests, =)