【作者】 刘爱军；

【导师】 杨育；

【作者基本信息】 重庆大学 ， 机械工程， 2011， 博士

【摘要】 车间作业调度是制造企业生产管理中十分重要且关键的环节,是实现先进制造和提高生产效益的主要方法和手段。论文针对离散制造车间多生产模式下作业调度较难优化的问题,以智能算法为主要技术手段,对其进行了系统而深入研究,提出了调度优化方案。主要研究内容如下:①对车间调度问题进行了概述,重点对调度发展过程、分类、特点、基础理论、指标体系、算法编码进行了归纳与总结;对车间调度研究现状和车间作业调度研究现状做了系统的分析;提出车间作业调度研究中的不足,明确研究的目的。②对多生产模式下Job Shop调度进行了总体研究;把车间作业调度问题细分为单目标经典生产模式下Job Shop调度问题、多目标静态柔性非模糊生产模式下Job Shop调度问题、多目标静态柔性模糊生产模式下Job Shop调度问题、多目标动态柔性生产模式下Job Shop调度问题等四个子问题;在此基础上,对论文所要研究的Job Shop调度四个子问题的总体数学模型、一般典型流程和总体技术框架进行了深入研究。③针对单目标经典生产模式下Job Shop调度问题,分别从改善算法结构和算法融合的角度出发,提出了两种优化技术:基于改进免疫克隆算法的优化技术和基于含精英策略的小生境遗传模拟退火算法的优化技术。前者,在融入并行计算和种群协同竞争思想的基础上,通过免疫记忆机制、克隆增殖、高频变异和交叉算子的操作,取得了深度搜索和广度寻优之间的平衡;后者,通过小生境技术、自适应双点交叉和互换变异策略、精英保留策略改善算法性能,并采用这两种优化技术对以最小化加工周期为目标的经典Job Shop调度问题进行了优化。④针对多目标静态柔性非模糊生产模式下Job Shop调度过程中只考虑工件和机器设备而忽略人机协同的问题,提出人机协同配置的调度优化技术。解决方案的基本思路如下:根据目前绝大多数调度理论仅关注机器设备单一资源调度的特点,提出人机双资源协同配置的多目标优化模型;并设计了可实现工艺路线和人机配合的两层柔性约束的三层编码方式;采用遗传算法非支配解集思想对多产线共存下的生产路径进行寻优。⑤针对多目标静态柔性模糊生产模式下Job Shop调度中的多资源多工艺路线优化问题,提出了两种优化技术:基于多种群遗传算法的优化技术和基于改进非支配排序遗传算法的优化技术。前者,采用多目标单一化方法处理多个需要优化的目标,建立了以最小化最大完工时间和最大化顾客满意度为目标的优化模型,提出了多种群协同进化的遗传算法,并用该算法对多目标静态柔性模糊生产模式下Job Shop调度问题进行了研究;后者,采用非支配解集思想处理多个需要优化的目标,建立了以最小化生产总流程时间、最大化客户满意度和最小化加工成本为目标的优化模型,提出了改进非支配排序遗传算法,并用该算法对多目标静态柔性模糊生产模式下Job Shop调度问题进行了优化。⑥针对多目标动态柔性生产模式下Job Shop调度中周期和事件双重扰动的优化问题,提出了基于自适应遗传算法的多目标优化技术,解决方案的基本思路如下:首先,对动态调度策略、动态调度研究方法及动态窗口规划技术进行了研究;其次,对基于周期驱动、基于事件驱动、基于周期和事件混合驱动的动态调度类型进行了研究;最后,采用自适应动态柔性多目标调度算法对周期和事件双重扰动的调度问题进行了优化,并对影响动态调度性能波动的事件因素和再调度周期进行了分析。⑦最后,对本文所做研究工作进行了总结,并对今后的研究工作进行了展望。更多还原

【Abstract】 Job shop scheduling is the very important, but weak part in the production management of manufacturing enterprises, and it is the foundation and key to realize advanced manufacture and improve production efficiency. To address the difficult problem of job shop scheduling optimization in multi-production mode for discrete manufacturing workshop, Job scheduling problem in multi-production mode for discrete manufacturing workshop are deeply studied by means of intelligent optimization algorithms, the optimization solutions are proposed respectively. The main contents of this thesis are described as follows:①An overview of the shop scheduling problem is presented. The development process, classification and characteristics of job shop scheduling (JSS) are summarized. The research status of production shop scheduling and JSS is analyzed systematically. The insufficiency of JSS research is pointed out and the purpose of this research is presented.②The classification, related theories and techniques of job shop scheduling are studied in general. First, based on analysis, the JSS problems are divided into four types, namely single objective job shop scheduling problem (STJSSP) in classic production mode, multi-objective flexible job shop scheduling problem (MOFJSSP) in static non fuzzy production mode, MOFJSSP in static fuzzy production mode, MOFJSSP in dynamic production mode. Second, regarding those four sub-problems, related general mathematical models, typical process procedures and the overall technology frameworks are elaborated.③To solve the problem in the classic JSS, two optimization techniques are proposed considering the structure and combination of algorithm, namely the optimization technique based on Immune clonal algorithm and the optimization technique based on Elite strategy with niche genetic simulated annealing algorithm. For the former, the balance of depth search and breadth search is obtained by the application of immune memory mechanism, clonal proliferation, high frequency mutation and crossover operation based on the idea of population collaborative competition and parallel computing. For the latter, algorithm performance is improved by niche technology, adaptive double point crossover and interchange mutation strategy, elitism strategy, and with the two optimization techniques the CJSSP with minimization processing cycle is further optimized . ④For static multi-objective flexible job shop scheduling only consider the process and machinery parts and ignore the problem of man-machine cooperation,the scheduling optimization technique of man-machine cooperation configuration is proposed. The basic idea of solution is as follows: according to the characteristic that most scheduling theory only pays attention to the single equipment resource scheduling currently, double resource collaborative optimization configuration multi-objective model about man and machine is put forward; and the three layer encoding method is designed about the process route and man-machine cooperation two layer flexible constraints; the production path under multiple lines is optimized by non-dominated set genetic algorithms.⑤For the optimization problem of multiple resources and process routes in static fuzzy flexible job shop multi-objective scheduling, two optimization techniques are proposed. One optimization technique is based on Multi-group Genetic Algorithm; the other technique is Improved Non-dominated Sorting Genetic Algorithm. For the former, multiple objective simplification method is used to deal with the multiple objectives which need to be optimized, the multi-objective optimization model is established with the objective of minimizing the maximum completion time and maximizing the customer satisfaction, and a Genetic Algorithm of multi-group concerted evolution is presented. With the algorithm static fuzzy multi-objective FJSSP is studied. For the latter, the non-dominated solution set mind is used to deal with the multiple objectives which need to be optimized. The multi-objective optimization model is established with the objective of minimizing the total production cycle time, maximizing the customer satisfaction and minimizing the processing cost. An Improved Non-dominated Sorting Genetic Algorithm is presented. Static fuzzy multi-objective FJSSP is optimized with the algorithm.⑥For optimization problems with the cycle and event doubly perturbed in dynamic flexible job shop multi-objective scheduling, the multi-objective optimization technique based on adaptive Genetic Algorithm is presented. The basic idea of solution is shown as follows: Firstly, the strategy of dynamic scheduling, the research method of dynamic scheduling and the planning technique of dynamic window are studied. Secondly, the dynamic scheduling type based on the cycle driving, event driving and the hybrid driving of cycle and event are researched. Finally, the Adaptive Genetic Algorithm is used to optimize the multi-objective dynamic scheduling problem based on the hybrid driving of cycle and event, and the events and dispatching cycle impacting dynamic scheduling performance fluctuation are analyzed.⑦At last, the main contents and contributions of the research are summarized, and the suggestions for further research of this topic are presented.更多还原