Multi-objective Evolutionary Algorithm with Variable Neighborhood Search for Optimizing Green Scheduling in a Re-Entrant Hybrid Flow Shop with Dynamic Events

Abstract In previous studies on re-entrant hybrid flow shops, the impact of dynamic events was often ignored despite being a common occurrence in practical production. To address this issue and simultaneously reduce energy consumption, a multi-objective evolutionary algorithm with variable neighborhood search (MOEA-VNS) has been proposed to optimize the green scheduling problem in a re-entrant hybrid flow shop with dynamic events (RHFS-GDS).The approach involves creating a green dynamic scheduling optimization model, which aims to minimize the makespan, total energy consumption, and stability of rescheduling solutions. A hybrid green scheduling decoding method is then employed to select machines for each operation and calculate fitness. Additionally, three neighborhood structures are designed to improve the population diversity and optimality of the MOEA-VNS algorithm. Two rescheduling strategies are also adopted to handle dynamic events that may occur during production. Experimental results demonstrate that these approaches are effective in solving the RHFS-GDS problem and can guide actual production. By incorporating dynamic events and rescheduling strategies into the optimization process, the proposed MOEA-VNS algorithm provides a comprehensive solution to the complex challenges faced by re-entrant hybrid flow shops in practical production environments.