sci.op-research

Amy wrote:
> Hi, Dear All,
>
> I'm generating cuts for my MIP problem by applying the separation
> procedure for multiple iterations (say 10 iterations, and get the
> lower bound Obj1), after that I solve the problem as IP. Now I'm
> thinking to remove the loose cuts before brand and bound and doing
> some test. After solving LP and adding cuts for 10 iterations, I
> printed out all the cuts and identify the tight cuts. If I only put
> those tight cuts to the original model and solve it as LP, shall I get
> the same objective value as Obj1?
>
> Thanks,
>
> Amy

Yes (provable using duality theory).

/Paul