基于遗传算法的可重构制造系统多零件流水线构形优化  

Optimization for multi-part flow-line configurations of reconfigurable manufacturing systems based on genetic algorithm

作　　者：窦建平[1] 戴先中[2] 李俊[2] 

机构地区：[1]东南大学机械工程学院,江苏南京211189 [2]东南大学自动化学院复杂工程系统测量与控制教育部重点实验室,江苏南京210096

出　　处：《计算机集成制造系统》2010年第7期1390-1400,共11页Computer Integrated Manufacturing Systems

基　　金：国家973计划资助项目(2002CB312204);东南大学新进博士科研启动基金资助项目(9202000023)~~

摘　　要：获取各生产周期内的最优和K-1个次优(K优)多零件流水线构形是可重构制造系统运行阶段的一个重要优化问题。给定各零件的工序优先图、工序和工位操作的关系以及各工位操作的可选设备,多零件流水线构形优化问题即为确定工作站数量、各工作站内并列放置机床的类型和数量以及选择和分配各零件的工位操作,以最小化构形的资本成本。为获得K优构形,首先放宽现有模型对工位操作分配的限制,建立了构形优化问题的0-1非线性规划模型,扩展了可行解空间。随后提出一种面向可行工位操作分配的遗传算法,从可行解空间中快速获取K优解。案例研究表明,该方法能获得优于现有模型最优解的解,同时也验证了所建模型和优化方法的有效性。Obtaining the optimal and K-1 near-optimal （K-best） Multi-Part Flow-Line （MPFL） configurations as candidates for each demand period was an important optimization problem for reconfigurable manufacturing system in the operational phases. Given the operation precedence graph for each part, relationship between sequence and Operation Setups （OSs） as well as machine options for each OS, the problem was to determine the MPFL configuration＇s parameters in order to minimize capital cost of MPFL configurations. The parameters included number of workstations, number of parallel machines and machine type as well as assigned OSs for each workstation. To generate K-best MPFL configurations, firstly a generic 0-1 NonLinear Programming （NLP） model which widened the solution space was developed by relaxing the limitation of the assignment of OSs in existing models. Then, a Feasible OS Assignment Oriented Generation Algorithm （FOAOGA） was proposed to efficiently find K-best MPFL configurations from the solution space of the 0-1 NLP model. A case study showed that the optimum found by FOAOGA was better than the optimum obtained by existing approach, and also demonstrated the effectiveness of the proposed model.

关 键 词：可重构制造系统 构形优化 多零件流水线 0-1非线性规划 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]