Originally posted by: Mintch Zulitch

Hi,

How can someone show that CP actually yields the optimal solution?

Is there any optimality proof for the CP used in IBM ILOG?

And Is there any supplementary materiel describing how the IBM Ilog embedded the Optimization Concept into CP?

Hi,

you will see a proof if you run floatexpr example.

In the log you will see:

 ! ----------------------------------------------------------------------------
 ! Search terminated normally, 1 solution found.
 ! Best objective         : 5,87794 (optimal - effective tol. is 0,000587794)

If CPLEX CPO had stopped because of a time or a branch limit, that would have been said.

regards