Hybrid genetic algorithms for multi-period part type selection and machine loading problems in flexible manufacturing system

Abstract

This paper addresses the multi-period part type selection and machine loading problems in flexible manufacturing system (FMS) with the objective of maximizing system throughput and maintaining balance of the system for the whole planning horizon. Various flexibilities including machine and tool flexibility, routing flexibility, and alternative production plans are considered. Hybridization of real coded genetic algorithms (RCGA) and variable neighborhood search (VNS) is proposed to simultaneously solve these NP-hard problems for the whole periods. The proposed hybrid genetic algorithms (HGA) are designed to balance the power of the algorithms to explore a huge search space and to exploit local search areas. The experiments show that addressing the problems for the whole periods simultaneously will produce better results comparable to those achieved by the sequential approach.