Abstract: To improve equipment reliability and production efficiency,the integrated optimization of production scheduling and machine maintenance was studied within a semiconductor assembly and test flexible flow-shop,characterized by multi-variety mixed-line production and limited machine capacity.Considering the rigid-flexible hybrid constraints of buffer capacity and due window,an integrated optimization mathematical model was developed to minimize the makespan,average flow time,earliness/tardiness penalty and machine workload extreme deviation.To solve the problem efficiently,an Improved Bi-population Cooperative Evolutionary Algorithm(IBCEA)was proposed.In IBCEA,a hybrid initialization method was introduced to generate a high-quality initial population,and a hybrid local search strategy integrating Variable Neighborhood Search(VNS)and Orthogonal Learning(OL)was designed to enhance algorithm performance.In the numerical simulations based on the assembly and test production line of a semiconductor enterprise in Chengdu,the effectiveness and superiority of improved operators and proposed IBCEA were verified through algorithm comparison.

Key words: flexible flow-shop, integrated optimization of production scheduling and machine maintenance, limited buffer, due window, improved bi-population collaborative evolutionary algorithm

摘要： 为提高设备可靠性与生产效率,针对具有多品种混线生产和机器产能限定的半导体封装测试柔性流水车间,开展了生产调度与设备维护集成优化研究。构建了一个多目标集成优化数学模型,该模型考虑了有限缓冲区和交货期时间窗刚柔混合约束,并以最小化最大完工时间、平均流经时间、提前/拖期完工惩罚和机器负荷极差为优化目标。为实现高效求解,提出一种改进的双种群协同进化算法,该算法采用混合初始化方法生成高质量初始种群,并利用结合变邻域搜索和正交学习的混合局部搜索策略提高算法性能。以成都某半导体企业的封装测试产线为案例开展数值仿真实验,并通过对比分析验证了改进算子和所提算法的有效性与优越性。

关键词: 柔性流水车间, 生产调度与设备维护集成优化, 有限缓冲区, 交货期时间窗, 改进双种群协同进化算法