不确定条件下SMT产品生产规划的建模及GA求解  

On SMT Production Schedule Under Uncertainty

作　　者：刘电霆[1] 周德俭[2] 

机构地区：[1]桂林理工大学机械与控制工程学院,广西桂林541004 [2]广西工学院机械工程系,广西柳州545006

出　　处：《桂林理工大学学报》2011年第2期278-285,共8页Journal of Guilin University of Technology

基　　金：广西自然科学基金项目(桂科自0991251);广西制造系统与先进制造技术重点实验室开放基金项目(09-007-05_011);广西科学研究与技术开发计划项目(桂科攻0992006-8)

摘　　要：针对PCBA组装企业的SMT产品生产规划问题,给出了一种考虑预估订单和生产线故障率等不确定条件的生产任务排程方法。该方法将预估订单的批量、来料时间和交货期等用区间数表示,建立其不确定型调度模型,并采用可能度水平将该模型转换为确定型,然后用遗传算法进行求解。给出了该遗传算法的结构,并提出了"染色体按生产线分段、段内按产品加工序排列"的8进制编码、基因交叉和段长改变的变异操作等具体方法。实验结果验证了其可行性,具有较好的参考价值。Facing the problem of the SMT （surface mount technology） production plan in PCB （printed circuit board） assembly firm, a kind of new method of arranging the assembly tasks to the production lines under uncertain condition, where the customs＇orders predicted and the fault rates of the production lines are considered, is presented. Its corresponding mathematic model, where the scales of orders predicted and the arriving time of materials and the supplying time of orders predicted are described as interval numbers, is set up. The model is converted into an ascertain one by of this GA, especially, the method possibility-level, and then solved by GA （genetic algorithm）. The structure of octal chromosome coding, are proposed in this paper. And the mutation operations include the swap of the genes on two points and the length of the segments is changed randomly. A computational example is studied, the results reveal that the model and the converting method and the GA are correct. The method presented has high reference value in solvine similar problem.

关 键 词：任务规划 不确定 建模 遗传算法 区间数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]