【作者】 曾强；

【导师】 杨育；

【作者基本信息】 重庆大学 ， 机械工程， 2010， 博士

【摘要】 车间调度是集成制造系统的重要而薄弱的环节。论文针对离散制造企业批量生产车间调度的特点,以智能优化算法为主要技术手段,对批量生产的等量分批问题、平顺移动下等量分批多目标车间调度问题、多工艺流程下等量分批多目标车间调度问题、基于准时交货要求的批量生产柔性作业车间调度问题进行了深入研究,提出了系统化的优化方案。具体研究内容如下:①对车间调度的分类、特点、发展、方法进行了系统的归纳与总结;对批量生产车间调度的研究现状进行了分析;指出现有研究存在的问题,明确了研究的目的。②对离散制造企业批量生产车间调度问题进行了描述;针对批量生产车间调度特点,提出了一个包括目标层、准则层、影响因素层、变量层、方案层、技术层六层结构的批量生产车间调度智能优化技术框架;对技术框架中的基础理论及关键技术进行了深入研究。③针对批量生产等量分批问题,提出了两种优化技术:基于Witness组合仿真优化的等量分批优化技术(WSC-ELS)和基于NSGA II算法的等量分批优化技术(NSGAII-ELS)。前者基于分步优化策略,是一种局部优化方法,计算速度快,可求解大规模等量分批调度问题;后者基于集成优化策略,是一种全局优化方法,计算速度相对慢,适用于求解中小规模等量分批调度问题。④针对平顺移动下的等量分批多目标车间调度问题,提出了两种智能优化技术:平顺移动下等量分批JSP多目标优化技术(PO-MJSP)和平顺移动下等量分批FJSP多目标优化技术(PO-MFJSP)。解决方案的基本思路如下:根据问题特点建立多目标优化模型;提出并设计了改进的NSGA II算法对模型进行求解,在算法中采用了平顺移动、时间分离、相同作业、间隙挤压四种精细化调度技术缩短完工时间;通过案例分析得出了研究结论。⑤针对多艺流程下等量分批多目标车间调度问题,提出了两种智能优化技术:多工艺流程下等量分批JSP多目标优化技术(MPF-MJSP)和多工艺流程下等量分批FJSP多目标优化技术(MPF-MFJSP)。解决方案的基本思路如下:以完工时间最短和生产成本最低建立了多目标优化模型;提出并设计了改进的NSGA II算法对模型进行求解,算法中引入了工艺流程编码,用于实现各加工批次的工艺流程优选;通过案例分析得出了研究结论。⑥针对基于准时交货要求的批量生产柔性作业车间调度问题,提出了两种智能优化技术:基于准时交货要求的批量生产FJSP单目标优化技术(JIT-SFJSP)和基于准时交货要求的批量生产FJSP多目标优化技术(JIT-MFJSP)。对于前者,建立了以加权平均隶属度最大为目标函数的单目标优化模型,提出并设计了一种多阶段混合变异的改进禁忌搜索算法;对于后者,建立了以加权平均隶属度最大和流程时间价值总量最小为目标函数的多目标优化模型,提出并设计了一种改进的NSGA II算法,算法中引入了各加工批次最早允许开工时刻以消解准时交货要求和快速生产之间的矛盾;通过案例分析得出了研究结论。⑦最后,对本文研究工作进行了总结,并对批量生产调度问题的进一步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 integrated manufacturing system. In view of the Job-shop scheduling characteristics for batch production of discrete enterprises, intelligent optimization algorithms taken as the main technical measure, the following four issues about Job-shop scheduling for batch production are deeply studied and its optimization solutions are proposed: equal lot splitting problem, multi-objective optimization problem for equal batch splitting Job-shop scheduling under parallel and ordinal shift mode, multi-objective optimization problem for equal batch splitting Job-shop scheduling under multiple process flows, multi-objective optimization problem for batch production FJSP based on JIT delivery. The main contributions of this thesis are shown as follows:①The classification, characteristic, development and method of Job-shop scheduling is systematically induced and summarized. The research status for Job-shop scheduling of batch production is analyzed. The existing problems and the research object are presented.②Description is given to Job-shop scheduling problem for batch production of discrete enterprises. Considering the characteristics of Job-shop scheduling for batch production, a six-hierarchy technical framework is proposed, including object hierarchy, rule hierarchy, influence factor hierarchy, variable hierarchy, scheme hierarchy and technique hierarchy. The basic theories and critical techniques in the framework are thoroughly studied.③For the equal lot splitting problem of batch production, two optimization techniques are proposed: one is the equal lot splitting technique based on Witness combinational simulation optimization technique (WSC-ELS), the other is that based on NSGA II(NSGAII-ELS). The former is based on decomposition optimization strategy, belonging to a local optimization method, calculating quickly, fit for large-scale equal lot splitting Job-shop scheduling problem. The latter is based on integrated optimization strategy, belonging to a global optimization method, calculating relatively slowly, fit for modest-scale equal lot splitting Job-shop scheduling problem.④For the multi-objective optimization problem of the equal batch splitting Job-shop scheduling under parallel and ordinal shift mode, two intelligent optimization techniques are put forward: one is the multi-objective optimization technique of equal lot JSP under parallel and ordinal shift mode (PO-MJSP) , the other is that of equal lot FJSP parallel and ordinal shift mode (PO-MFJSP). The basic solution thought is demonstrated as follows: Firstly, a multi-objective optimization model is established; Secondly, an improved NSGA II is presented and designed to solve the model, in which four delicacy scheduling techniques are used to reduce the makespan, including parallel and ordinal shift, time separation, similar operation and interval squeezing; Thirdly, conclusion is drawn by case study.⑤For the multi-objective optimization problem of the equal batch splitting Job-shop scheduling under multiple process flows, two intelligent optimization techniques are presented: one is the multi-objective optimization technique of equal lot JSP under multiple process flows (MPF-MJSP), the other is that of equal lot FJSP under multiple process flows (MPF-MFJSP). The basic solution thought is shown as follows: Firstly, a multi-objective optimization model is established with the objective to minimize the makespan and the manufacturing cost; Secondly, an improved NSGA II is proposed and designed to solve the model, in which process flow code is introduced to implement the optimal selection of process flow for every process batch; Thirdly, conclusion is drawn by case study.⑥For the multi-objective optimization problem of batch production Flexible Job-shop scheduling based on JIT delivery, two intelligent optimization techniques are proposed: one is the single-objective optimization technique of batch production FJSP under JIT delivery (JIT-SFJSP), the other is multi-objective optimization technique of batch production FJSP under JIT delivery (JIT-MFJSP) . For the former issue, an optimization model is established with the objective to maximize the weighted average membership degree, and an improved multi-stage hybrid mutation taboo search algorithm is put forward and designed to solve the model. For the latter issue, a multi-objective optimization model is constructed with the objective to maximize the weighted average membership degree and minimize the total flow time value , and an improved NSGA II is presented and designed to solve the model, in which the earliest allowable begin time of every process batch is introduced to eliminate the contradiction between the JIT delivery demand and quick production.⑦Finally, the whole research work of the dissertation is summarized, and the future work of batch production scheduling is given.更多还原