Originally posted by: cpgt

I've been trying to find out if CP or MP is faster and have come up empty. So I've been doing my own informal testing using facility.cs example using both MP and CP, and CP is faster, sometimes by a factor of 2X, 3X. Is this expected for the type of problem and is there more official in-depth benchmarks comparing the two.

Originally posted by: Petr Vilím

Hello,

to my knowledge, there is no generic benchmark for comparing MIP and CP. The reason that the performance of both tools depends on the problem, for some problems MIP is better, for some CP is better.

There is also a difference in modeling the problem. CP has richer language for describing the problem, especially for scheduling problems. Taking a MIP model and putting it to a CP solver without any change isn't a good idea.

If your problem is similar to the example facility then yes, CP is probably a better choice.

Best regards, Petr

Originally posted by: cpgt

I've found several IBM slides which show benchmarks for MIP vs CP, see slides 19& 24.