Some Easy Postoptimality Analysis for Zero-One Programming

Abstract

Some easy postoptimality analysis for zero-one programming is developed. First, a procedure is presented for obtaining a set K of 0-1 solutions whose objective function values are within a specified tolerance of the optimum. Then sufficiency conditions for K to contain the optimum of a revised problem are derived. Computational experience with a number of test problems indicates that K can be economically obtained and that the sufficiency tests provide a significant postoptimization capability.