基于改进遗传算法的产品试验计划调度研究

发布时间：2018-02-16 09:57

本文关键词： 试验计划试验调度回溯算法遗传算法　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：高压电器产品试验检测在输配电装备制造行业中占有重要地位,随着国家电网对输配电装备的需求量逐年增长,国内的输配电行业试验检测业务数量急剧增加,同时客户对企业试验检测服务的能力、周期和费用等要求越来越高。由于试验检测过程的不确定性及试验过程建模的复杂性,使得试验计划调度效果普遍不理想,目前缺乏一种较为通用的产品试验计划调度解决方法,这就对企业的试验检测效率和质量提出了挑战。针对上述问题,本文分别从试验计划和试验调度两方面着手。首先,针对试验计划任务分配问题,提出了基于回溯算法的试验任务分配方法,经算法实现得到了试验任务的最佳分配方案;其次,针对产品试验调度问题,提出了基于遗传算法的试验调度方法,并对算法进行了一定的改进与实现,最终得到了满意的试验调度结果。本文的主要研究工作如下:1.介绍了高压电器产品试验检测计划调度的概念及研究现状,总结出现有试验计划调度问题解决方法的不足,提出了产品试验计划与试验调度协同优化的方法。对于产品试验计划任务分配问题,进行基于回溯算法的设计与编码实现,得到了试验计划任务的最佳分配方案。2.首先介绍了产品试验调度问题的特点和组成要素以及传统解决方法的不足;然后概述了遗传算法的概念、优缺点、执行流程以及具体操作步骤;最后针对产品试验调度问题,以试验完工时间最小和各试验机器的负载分布均衡为优化目标,进行基于遗传算法的设计与编码实现,并对所得调度结果进行一定的分析与评价。3.分析了使用遗传算法求解试验调度所存在的一些问题,提出了扩大算法搜索空间和动态确定交叉与变异概率的算法改进思想,并对遗传算法进行一定的改进与编码实现,最终得到了满意的试验调度结果,从而有力地验证了算法改进后的良好性能。
[Abstract]:The testing and testing of high-voltage electrical products plays an important role in the manufacturing industry of transmission and distribution equipment. With the increasing demand for transmission and distribution equipment in the State Grid, the number of test and detection services in the transmission and distribution industry in China has increased sharply. At the same time, customers are demanding more and more on the capability, cycle and cost of the testing service. Because of the uncertainty of the testing process and the complexity of the modeling of the test process, the scheduling effect of the test plan is generally not ideal. At present, there is a lack of a more general solution to the scheduling of product test plans, which challenges the efficiency and quality of testing and testing in enterprises. In view of the above problems, this paper starts from two aspects: test plan and test scheduling. In order to solve the task assignment problem of experimental plan, a method based on backtracking algorithm is proposed, and the optimal assignment scheme of test task is obtained by the algorithm. Secondly, aiming at the problem of product test scheduling, The experiment scheduling method based on genetic algorithm is put forward, and the algorithm is improved and implemented. The main research work of this paper is as follows: 1.Introduces the concept and research status of test test plan scheduling for high-voltage electrical products, and summarizes the deficiency of solving the problem of test plan scheduling. The collaborative optimization method of product test plan and test scheduling is proposed. For the task assignment of product test plan, the design and coding implementation based on backtracking algorithm are carried out. The optimal assignment scheme of the test plan task is obtained. Firstly, the characteristics and components of the product test scheduling problem and the shortcomings of the traditional solution are introduced, and then the concept, advantages and disadvantages of genetic algorithm are summarized. Finally, aiming at the problem of product test scheduling, the minimum test completion time and the load distribution balance of each test machine are taken as the optimization objectives, and the design and coding based on genetic algorithm are carried out. Some problems existing in solving test scheduling with genetic algorithm (GA) are analyzed, and an improved algorithm for expanding search space and dynamically determining crossover and mutation probability is put forward. The genetic algorithm is improved and coded to a certain extent, and a satisfactory experimental scheduling result is obtained, which verifies the good performance of the improved algorithm.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM51;TP18

【参考文献】