A self-learning multi-population evolutionary algorithm for flexible job shop scheduling under time-of-use pricing

Due to the energy crisis and environmental downgrade, manufacturing companies face rising costs. The pressure on manufacturers to reduce costs and increase efficiency is increasing. The flexible job shop scheduling problem considering time-of-use pricing (FJSPTOUP) is studied to minimize the maximum completion time and total electricity cost. A new self-learning multi-population evolutionary algorithm (SLMPEA) is proposed to address the static scheduling problem. A self-learning module is designed to acquire knowledge from elite individuals in subpopulations. The knowledge discovered can guide the evolution of subpopulations. To improve the convergence of the algorithm and maintain the diversity of the population, we present an elite retention and enhancement strategy. Two enhancement operators are developed to elevate the quality of elite individuals, and a selection strategy is designed for the enhancement operators. A time-of-use-based adjustment strategy is proposed to reduce the total electricity costs without increasing the maximum completion time. Experimental results show that the proposed SLMPEA outperforms the comparative algorithms in solving FJSPTOUP.