华中科技大学学报（自然科学版）2025，Vol.53Issue(6)：46-53,8.DOI:10.13245/j.hust.250607
基于双种群协同进化算法的晶圆制造跨区多目标调度方法
Dual-population co-evolutionary algorithm based multi-objective cross-area scheduling method in wafer manufacturing
摘要
Abstract
To address the problem that the multi-objective cross-area scheduling in wafer manufacturing systems faces the challenges of generating feasible solutions and often requiring frequent manual interventions,due to multiple time windows constraints caused by residence time and cleaning process within and between different areas,a multi-objective mixed integer programming model with objectives of minimizing the total tardiness and total production cost was established,and a dual-population co-evolutionary scheduling optimization algorithm was proposed based on adaptive relaxation factors.First,to balance the multi-objectives and various constraints,a dual-population co-evolution mechanism was designed,including a normal population that prioritized feasibility and a convergent population that prioritized convergence and diversity.According to the state of the convergent population,the normal population was selectively provided with excellent individuals to jump out of the local optimum.Then,adaptive relaxation factors were implemented to retain high-quality infeasible solutions within the convergent population to enhance the exploration ability in solution space and improve solution diversity.For infeasible solutions generated by crossover and mutation operators,individual repair strategies were devised to enhance population feasibility.Finally,results of 27 sets of algorithmic experiments and 6 consecutive months of testing of the wafer manufacturing simulation system show that the proposed algorithm achieves a Pareto solution set with superior convergence and diversity,which effectively reduces both the total delay and total production cost in the wafer manufacturing system.关键词
晶圆制造系统/多时间窗约束/多目标优化/自适应松弛因子/双种群协同进化算法Key words
wafer manufacturing system/multiple time window constraints/multi-objective optimization/adaptive relaxation factor/dual-population co-evolutionary algorithm分类
机械制造引用本文复制引用
张朋,金孟宇,王明,吴立辉,张洁..基于双种群协同进化算法的晶圆制造跨区多目标调度方法[J].华中科技大学学报（自然科学版）,2025,53(6):46-53,8.基金项目
国家重点研发计划资助项目(2022YFB3305003) （2022YFB3305003）
国家自然科学基金资助项目(52005099). （52005099）
