An adaptive genetic algorithm with multiple operators for flowshop scheduling

Abstract

Genetic algorithms (GAs) are a class of effective parallel searching algorithms inspired by the idea of “survival of the fittest”, which has been successfully applied to a variety of problems, especially in the fields of manufacturing and scheduling. However, it is reported that traditional GAs often suffer from the weaknesses of premature convergence as well as parameter and operator dependence. So far, many improved methods with adaptive parameters or hybrid structures have been proposed, but there is little literature considering the adaptive control of genetic operators. In this paper, an adaptive GA (AGA) with multiple operators is proposed for flowshop scheduling, which is a typical NP-hard optimisation problem with many industrial applications and has been widely studied in both academic and engineering fields. In AGA, multiple different genetic operators are employed in an adaptive hybrid way to enhance the exploration and exploitation abilities so as to prevent premature convergence and achieve superior performance. It especially important to stress that the utilising ratio of each operator for hybridisation is adaptively and dynamically controlled during the evolutionary searching process. Simulation results based on benchmarks demonstrate the effectiveness of AGA by contrast with traditional GAs. And the effect of the adaptive control of the operator and the effects of some parameters on the optimisation performance are discussed as well.