
From:  Andrew Makhorin 
Subject:  Re: [Helpglpk] the theoretic formula about the integrality gap for MILP and 01 knapsack integer programing model 
Date:  Thu, 03 Dec 2015 22:52:02 +0300 
> How to estimate the "bestpossible" and "epsilon" > without solving an integer programming model ? Find any integer feasible solution to MIP (not solving MIP exactly), e.g. with a primal heuristicit gives you "bestfound". > > > How to estimate the "bestpossible" and "epsilon" > without solving an linear programming model ? Find an optimal solution to any (not LP) relaxation of MIPit gives you "bestpossible". Take "epsilon" as a smallest floatingpoint number such that 1+eps > 1.
[Prev in Thread]  Current Thread  [Next in Thread] 