关键词: 双资源约束柔性作业车间调度, 机械加工车间, 逆调度, 多目标优化, 差分进化算法

Abstract: In order to improve the efficiency and stability of machining workshops in dynamic production environments, an inverse scheduling problem model of flexible machining workshops was established considering machine and worker constraints. The model aimed to minimize makespan, machine energy consumption and inverse deviation index by adjusting workpiece scheduling, worker work, and machining parameters. Aiming at the problem characteristics, an improved differential evolution algorithm was proposed. In the algorithm, a hybrid double-layer encoding method was designed to reduce the search difficulty. Two initialization methods were proposed to improve the population quality based on dispatching rules. In order to strengthen and balance the global and local search, adaptive genetic operations and neighborhood search strategies were designed based on elite selection. The Hamming distance was improved, and a crowding operator was proposed to reflect the true diversity of the population. In the experiment, 33 test instances were constructed and the proposed algorithm was compared with the other 7 algorithms to verify the performance. Finally, the real inverse scheduling cases of a hydraulic cylinder production workshop in two different dynamic environments were analyzed. The results show that the proposed algorithm may effectively reduce the makespan by 4.2 % and the machine energy consumption by 20.2 % with a little change in the original schedule.

Key words: dual-resource constrained flexible job shop scheduling, machining workshop, inverse scheduling, multi-objective optimization, differential evolution algorithm