【作者】 鞠全勇；

【导师】 朱剑英；

【作者基本信息】 南京航空航天大学 ， 机械电子工程， 2007， 博士

【摘要】 本文研究了智能制造系统的生产计划和车间优化调度问题,主要工作和创新点如下:根据生命科学中免疫系统的信息处理机制,将免疫计算和改进的遗传算法相结合,建立了一种用于车间调度的免疫遗传算法。针对作业车间调度问题,设计了免疫遗传计算中疫苗的提取和接种方法,通过作业车间调度十个典型标准问题验证,文中所述免疫遗传算法可行,较现有免疫算法、一般遗传算法及一些传统优化设计方法在收敛效率和准确性等方面有很大改进与提高。在研究双资源、多工艺路线作业车间调度的基础上,从实际作业车间调度系统存在大量不确定因素的情况出发,建立了模糊调度的数学模型。基于模糊理论和自适应原理,对算法中初始种群的构造、适应度计算、模糊遗传操作等方面进行了研究,以最小完工时间和平均满意度最大为优化目标,应用改进的模糊遗传算法,求解出最优调度工序。提出了面向车间调度的动态、分布式工艺计划与车间调度集成模型,深入构建、研究了集成模型的层次结构。将工艺计划与基于周期和事件驱动的动态优化调度有机地相结合,使集成系统能适应连续加工过程中复杂的环境变化并高效地完成实时处理,减少突发事件造成的工序大范围的重新设计。把二倍体混合遗传算法引入动态车间优化调度运算,从而使集成模型中动态生产调度与控制功能得以实现。实例验证了集成系统和算法的可行性和有效性。研究了批量生产中以生产周期、最大提前/最大拖后时间、生产成本、以及设备利用率指标:机床总负荷和机床最大负荷为调度目标的柔性作业车间优化调度问题。提出了批量生产优化调度策略。建立了多目标优化调度模型。结合多种群粒子群搜索与遗传算法的优点提出了具有倾向性粒子群搜索的多种群混合算法,以提高搜索效率和搜索质量。仿真结果表明该模型及算法较目前国内外现有方法更为有效和合理。本课题的研究受到国家自然科学基金重大项目“支持产品创新的先进制造技术中的若干基础性研究”(项目编号:59990470)的支持,课题的研究已经通过国防科工委的专家鉴定。专家认为:“研究成果具有开创性,整个研究成果属于国际先进水平,根据查新报告,其中多工艺多资源的动态优化生产调度技术属于国际首创”。更多还原

【Abstract】 The scheduling optimization of the job shops is a core of advanced manufacturing and modern managing technology. The problem of job shop scheduling has become one of key bottle-neck in production process. The subject is studied in this thesis and several innovations are presented.According to the information processing mechanism of immune system in biotic science, a new approach of immune genetic algorithm for job shop scheduling is proposed through combining immune algorithm with improved genetic algorithm. Aiming at the problem of job shop scheduling, the approach of distilling and injecting vaccination is solved, which is difficulty in immune algorithm. It is testified that convergence efficiency and accuracy of the immune genetic algorithm in solving ten standard job shop scheduling problems. The results indicate the proposed algorithm is competitive, being able to produce better solutions then other approach.Based on the foundation of studying job shop scheduling with dual-resource and multiple process plans, from the circs of mass incertitude complications existing in practical job shop scheduling system, the model of fuzzy job shop scheduling is formulated. The minimum fuzzy completion time or maximum average agreement index is taken as the target. An improved genetic algorithm is presented to attain the best strategy, in which coding; fitness counting, fuzzy algorithm operation is included.An integrated model is proposed, which is an integrated model of distributed and dynamic process planning & job shop scheduling. The integrated model’s hierarchy structure is constructed and researched in deeply. Process planning is combined with dynamic rolling windows scheduling based on period and event-driven. The integrated system can adapt to continuous processing in a changing environment and finish the disposal in time, and reducing redesign of process planning in large scale due to outburst events. The particle swarm algorithm is introduced to job shop scheduling operation. Consequently, the function of job shop scheduling and control of the integration module is realized. Feasibility and validity of the integrated system is validated by examples.The problem of multi-objective flexible job shop scheduling optimization of batch production is studied, where multi-objects of makespan, earliness/tardiness, production cost and equipment utilization rate: total and maximum machine tool loads are concerned. The strategy of job shop scheduling optimization of batch production is proposed. The model of multi-objective scheduling optimization is set up. Aiming at improving searching efficiency and searching quality, multiple population hybrid algorithm combining both advantages of particle swarm optimization and genetic algorithm is presented. A simulation experiment is carried out to illustrate that the proposed model and algorithm is more efficiency and feasible than that used in home and abroad in existence at present.The study is supported by key project of National Natural Science Foundation, and is authenticated by the specialists of Commission of Science Technology and Industry for National Defenses. The specialists have declared:“achievements of the study are creative and in the lead of the world. The investigation shows static and dynamic scheduling problems on multi-resources constraints are firstly solved in the world”.更多还原