On the population diversity control of evolutionary algorithms for production scheduling problems

In the literature, many evolutionary algorithms have been proposed for the production scheduling problems such as genetic algorithm, particle swarm optimization, differential evolution, and so on. However, these algorithms mainly focus on the efficiency of local search methods but seldom tackle the control of population diversity. Therefore, this paper aims to check whether the population diversity control strategy can significantly impact the performance of evolutionary algorithms for the production scheduling problems. With this aim in mind, a population diversity control strategy was proposed and applied in a basic evolutionary algorithm. The computational results on two representative production scheduling problems, i.e., the single machine scheduling and the permutation flowshop scheduling, show that the proposed population diversity control strategy can significantly improve the basic evolutionary algorithm and that this algorithm with the strategy (but without local search) can even be competitive with other powerful evolutionary algorithms with local search in the literature.