A dynamic co-evolution compact genetic algorithm for E/T problem

摘要

In this paper, a dynamic co-evolution compact genetic algorithm (DCCGA) is proposed for flexible flow shop scheduling problem (FFSP) to minimize the total earliness and tardiness (E/T) penalties. In this new algorithm, a dynamic co-evolution mechanism containing two probabilistic models and a best individual inheritance strategy are integrated into the compact genetic algorithm (CGA). For improving the stability of the evolutionary trend in the evolution processes, the diversity of evolution trend and the convergence speed. Lastly, the experimental results show that, DCCGA outperforms CGA by 11.74% on the problem we study.