Tabu Search in Integer Programming

Abstract

We have already observed that tabu search can be used to modify decision rules for basis exchange (pivoting) methods, and therefore can be applied to solving mixed zero-one integer programming problems and various nonlinear programming problems by this means. Such applications exploit the fact that optimal solutions to these special zero-one discrete and nonlinear problems can be found at extreme points, and that pivoting methods provide a mechanism to move from one extreme point to another. In addition, however, tabu search can also be applied to solve integer and nonlinear programming problems in other ways. In this section we focus on integer programming (IP) problems, including mixed integer programming (MIP) problems where some variables can take on continuous values, as well as pure IP problems, where all variables must receive integer values.