Multi-objective optimization with improved genetic algorithm

Abstract

We extend an improved GA (GA-SRM) to the multi-objective flowshop scheduling problem (FSP) in order to obtain better pareto-optimum solutions (POS). Two kinds of cooperative-competitive genetic operators in GA-SRM, CM and SRM, are extended to ones suitable for FSP in which solutions (individuals) are represented as permutations. Simulation results verify that GA-SRM shows better performance for the multi-objective optimization problem (MOP), and consequently better POS are obtained than conventional approaches with canonical GA.