Mixed 0-1 Programming through Benders' Partitioning Procedures and Genetic Algorithms

Abstract

In this paper, we focus on mixed 0-1 programming problems and propose an approximate solution method by combined use of the Benders' patitioning procedures and genetic algorithms. After overviewing the Benders' patitioning procedures, we examine how to apply genetic algorithms to the subproblems with 0-1 variables. Then we propose a coding method for applying genetic algorithms to the subproblems and a new decoding method for satisfying the constraints. Through computational experiments, the feasibility and efficiency of the proposed method are examined and compared with the branch and bound method for mixed 0-1 programming problems with respect to CPU time and perfomance.