基于区间理论的不确定集成式工艺规划与车间调度问题研究

发布时间：2018-06-05 01:52

本文选题：集成式工艺规划与车间调度问题 + 不确定问题　；参考：《华中科技大学》2014年硕士论文

【摘要】：制造系统中，工艺规划和车间调度分别承担重要的功能，二者有着紧密的联系。但是在以往的研究中，通常独立对两个系统进行优化研究。虽然两个系统都可以通过技术手段进行优化，减少生产过程中的冲突与浪费，但是将两个系统进行集成，将更大程度的提高系统的生产效率。本文针对集成式工艺规划与车间调度（Integrated Process Planning and Scheduling,IPPS）这一类经典的车间调度问题开展研究，同时重点考虑实际生产过程中广泛存在的不确定事件。这些事件主要是因外部因素而造成的加工环境的不确定，比如，运输时间的不确定、夹具和刀具装卸时间的不确定、准备时间的不确定、生产过程中的设备故障、能源短缺等，都会累积造成工序调度过程中的不确定性等。不确定IPPS问题能更好的处理这些事件，使得调度结果能更好地反映实际生产状况。确定性的IPPS问题已被证明为NP-Complete问题，不确定IPPS问题由于考虑了更多的不确定事件，所以建模与求解更加复杂，这也导致国内外鲜有针对不确定IPPS问题的研究。本文重点针对不确定IPPS的建模与优化方法开展研究。首先采用改进的区间理论对问题进行建模，主要内容包括利用区间理论和区间数表征不确定加工时间。给出区间数的操作法则，并针对当前区间数的比较方法进行了研究，提出了一种新的基于可能度和偏序率的区间数比较方法以提高区间数比较的精度。最终将改进的区间理论应用在不确定IPPS问题的建模上，提出了基于区间数的不确定IPPS问题模型。本文针对不确定IPPS问题的特性，首先提出了基于遗传算法的不确定IPPS问题求解方法。在工艺规划和车间调度部分，分别采用集成式编码和基于工序的编码方式，并设计了高效的遗传操作方法。基于标准IPPS算例将确定的加工时间进行区间化形成不确定IPPS问题算例，，算例测试结果验证了提出算法的有效性。由于单一算法求解简单问题比较有优势，而面对复杂的实际生产环境，需要拥有适应大规模问题求解的算法，基于此，本文提出了一种基于粒子群优化混合算法的不确定IPPS问题求解方法。在该方法中，对于基础粒子群算法进行了重新定义，以使其适用于非连续优化问题。同时在该方法中引入了遗传操作，提高了该方法处理组合优化问题的能力。最后对测试实例进行测试，验证了混合算法在求解大规模不确定IPPS问题上的卓越性能。最后，对本文的研究工作进行了总结，展望了下一步的研究工作。
[Abstract]:In manufacturing systems, process planning and job shop scheduling play important roles, which are closely related to each other. However, in previous studies, two systems are usually optimized independently. Although both systems can be optimized by technical means to reduce conflict and waste in the production process, the two systems are integrated. This paper focuses on the classical job shop scheduling problems such as integrated process planning and shop scheduling, such as integrated process planning and workshop scheduling. At the same time, it focuses on the uncertain events that exist widely in the actual production process. These events are mainly due to the uncertainty of the processing environment caused by external factors, such as the uncertainty of the transportation time, the uncertainty of the handling time of fixtures and cutters, the uncertainty of the preparation time, the failure of equipment in the production process, the shortage of energy, etc. Will accumulate to cause the process scheduling process uncertainty and so on. The uncertain IPPS problem can deal with these events better, so that the scheduling results can better reflect the actual production situation. The deterministic IPPS problem has been proved to be NP-Complete problem. This also leads to the lack of research on uncertain IPPS at home and abroad. This paper focuses on the modeling and optimization methods of uncertain IPPS. Firstly, the improved interval theory is used to model the problem. The main content includes the use of interval theory and interval number to characterize the uncertain processing time. The operation rule of interval number is given, and the comparison method of interval number is studied. A new method of interval number comparison based on probability and partial order rate is proposed to improve the accuracy of interval number comparison. Finally, the improved interval theory is applied to the modeling of uncertain IPPS problem, and a model of uncertain IPPS problem based on interval number is proposed. Firstly, an uncertain IPPS problem solving method based on genetic algorithm is proposed. In the part of process planning and job shop scheduling, integrated coding and process-based coding are adopted, and an efficient genetic operation method is designed. Based on the standard IPPS example, the fixed processing time is interlaced to form an example of uncertain IPPS problem. The experimental results show that the proposed algorithm is effective. Because a single algorithm has the advantage of solving simple problems, but in the face of complex actual production environment, it is necessary to have an algorithm suitable for large-scale problem solving. In this paper, a method for solving uncertain IPPS problems based on particle swarm optimization (PSO) hybrid algorithm is proposed. In this method, the basic particle swarm optimization algorithm is redefined to make it suitable for discontinuous optimization problems. At the same time, genetic operation is introduced into the method, which improves the ability of the method to deal with combinatorial optimization problems. Finally, a test example is tested to verify the excellent performance of the hybrid algorithm in solving large-scale uncertain IPPS problems. Finally, the research work of this paper is summarized, and the next research work is prospected.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP18;TB497

【参考文献】