关键词: 设置与运输时间, 机器可变加工速度, 柔性作业车间绿色调度, 聚类交叉, 衍生策略, 自适应局部搜索

Abstract: For the Multi-factor Flexible Job shop Green Scheduling Problem with Setup and Transportation time constraints and Variable machine processing Speeds (MFJGSP-STVS), a mathematical model with completion time and energy consumption as optimization objectives was constructed, and an Enhanced Muti-objective Evolutionary Algorithm (EMoEA) was proposed to solve the problem. A three-layer integer encoding method was adopted in the algorithm. The Machine Idle time Preference rule (MIP) and Turning On/Off strategy (TOF) were applied in decoding to optimize objectives. Heuristic rules such as Global Search (GS) were employed to generate the initial population. A cluster crossover approach was designed based on a non-dominated hierarchy to accelerate the algorithm's convergence. To prevent the algorithm from converging prematurely and falling into local optimum, a derivative strategy was proposed to diffuse the non-dominated solution set, and an adaptive local search strategy based on the critical path was designed to further enhance the exploration capability of the algorithm. In simulation experiments, each design demonstrated better Hypervolume (HV) and Inverted Generational Distance (IGD) metrics compared to the original multi-objective evolutionary algorithm, and EMoEA exhibited advantages in both HV and IGD metrics, converged faster, and achieved the optimal objective value in most instances compared to the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Hybrid Jaya Algorithm (HJaya). The results show that the proposed designs are effective, and EMoEA can efficiently address the MFJGSP-STVS, providing high-quality scheduling solutions for businesses.

Key words: setup and transportation time, variable processing speed, flexible job shop green scheduling, cluster crossover, derivation strategy, adaptive local search