Abstract: In the actual production of the green workshop, the producers are required to obtain the scheduling scheme of the multi-objective flexible job shop to meet the requirements in the shortest possible time. This paper proposes a method to solve scheduling problems by using individual historical information and limitation operators. This method decomposes multi-objective into a set of scalar sub-problem, and uses multi-objective evolutionary algorithm to optimize the sub-problem. The generation strategy with historical information is used to accelerate the speed of convergence. Then, a stable matching strategy, with limitation information, is used as selection operation and population with better diversity to be selected as the next parent. The result shows that compared with the original algorithm, the efficiency, cost and energy efficiency of this new algorithm are increased by 0.8%, 0.8% and 2.5%, respectively. And it is better than NSGA-II solution by 1.4%, 1.8% and 4.8%.

Key words: multi-objective evolutionary algorithm, historical information, limitation information, speed of convergence

摘要： 在绿色车间实际生产中，生产者要求在尽量短的时间内获得符合要求的多目标柔性作业车间的调度方案。提出一种使用个体历史信息和限制算子求解柔性作业车间优化调度问题的方法。该方法将多个优化目标分解为一组标量子问题，利用多目标进化算法优化子进行目标优化；在进化过程中，子代生成阶段使用历史信息，提高个体的改变量，加快收敛；在选择阶段，利用带有限制信息的稳定匹配选择策略选择多样性好的染色体种群作为下一次进化的父代种群，保证种群的多样性。实例仿真表明：相比已有算法，所提算法在效率、成本以及能效三个目标上分别提升0.8%、0.8%、2.5%，同时优于NSGA-II求解方案的1.4%、1.8%、4.8%。

关键词: 多目标进化算法, 历史信息, 限制信息, 收敛速度