基于粒子群算法的多产品批处理生产调度问题研究

发布时间：2018-08-19 16:05

【摘要】：流程工业在世界经济中占很大比例,是许多国家的基础产业和支柱产业。随着市场需求的变化,其生产过程逐步向多产品批处理生产方式转化。多产品批处理方式固有的灵活性使得其效率和效益很大程度上依赖生产计划的编制和调度方案的确定。然而,其生产调度问题是复杂的优化问题,现有的理论和方法不能很好的解决,迫切需要研究多产品批处理调度问题的建模和优化技术,探索寻找更好的调度方法,以指导企业实践、降低企业运营成本、提高企业管理水平。多产品批处理生产调度中各产品的物料按照相同的生产工艺依次连续通过各阶段的设备进行加工,当各阶段连续生产时类似于离散过程经典Flow-Shop调度问题。本文在以往研究的基础上对多产品批处理生产调度问题用PSO算法求解方面做了相关探索性研究并将研究应用于铝锭生产制造过程当中,探求解决实际问题的方法,为企业生产管理系统提供理论基础。首先,将各阶段由单一设备连续加工的多产品批处理调度问题转化为经典一般Flow-shop调度问题,对其建模方法进行研究,然后运用三种方法求解问题。其中,前两种方法是以往学者基于连续建模思想建立混合整数规划模型求解的方法。第三种方法是提出一种改进粒子群算法求解,算法中引入双向搜索策略改善了PSO易陷入局部最优而丧失种群多样性的缺陷。通过算例验证,传统的模型在求解小规模问题性能较好,而在大规模问题时,PSO算法是求解问题比较有效的方法。其次,在以上研究的基础上将问题进一步向实际生产环境拓展,研究了各阶段由并行设备协调生产的多产品批处理调度问题。由于加入了并行设备的选择而使问题相对于单一设备问题更加复杂。将其转化为经典的混合Flow-shop调度问题,在一般Flow-shop调度问题建模的基础上建立模型,并提出一种结合单纯形搜索和粒子群算法优势提高算法求解能力和效率的改进粒子群算法。并用一类经典实例测试验证,将算法的计算结果与文献中的模型计算结果比较,得出随着问题规模的增大,传统模型难以求解且解的质量不高,而在处理大规模的问题中,PSO体现出了优良性能。再次,基于以上研究对现代铝工业生产进行了简要描述和分析,根据实际生产流程对生产工艺过程简化处理,提取某铝厂由铝土矿生产工业用铝锭的生产过程,分析处理后将其抽象为一个各阶段并行设备协调生产的多产品批处理生产模式,然后用批处理方法对产品各阶段进行分批、计算确定对应批次的加工时间和能耗,最后将其转化为经典混合Flow-shop调度问题,从节能角度出发建立以最小能耗为调度目标的数学规划模型,调用与单纯形法混合的改进PSO进行求解,将求解结果与基本GA结果进行对比分析,进一步验证了PSO的优越性和其解决此类调度问题的能力。最后,总结了全文,展望了所研究问题未来的发展和应用。
[Abstract]:The process industry, which accounts for a large proportion of the world economy, is the basic and pillar industry in many countries. With the change of market demand, its production process is gradually transformed into multi-product batch production mode. However, the production scheduling problem is a complex optimization problem, and the existing theories and methods can not be well solved. It is urgent to study the modeling and optimization technology of multi-product batch scheduling problem, and explore a better scheduling method to guide enterprise practice, reduce enterprise operating costs and improve enterprise management level. In batch production scheduling, the material of each product is processed successively through the equipment of each stage according to the same production process. When each stage is continuous production, it is similar to the classical Flow-Shop scheduling problem of discrete process. Based on the previous research, this paper uses PSO algorithm to solve the multi-product batch production scheduling problem. The related exploratory research is carried out and applied to the aluminum ingot production and manufacturing process to explore the method to solve practical problems and provide a theoretical basis for the enterprise production management system. The first two methods are based on the idea of continuous modeling. The third method is to propose an improved particle swarm optimization algorithm to solve the problem. The introduction of two-way search strategy to improve PSO easy to fall into local optimum and loss of population diversity. A numerical example shows that the traditional model has better performance in solving small-scale problems, and the PSO algorithm is a more effective method for solving large-scale problems. Secondly, based on the above research, the problem is further extended to the actual production environment, and the multi-product batch production coordinated by parallel equipment in each stage is studied. The problem is more complicated than a single device problem because of the choice of parallel devices. It is transformed into a classical hybrid Flow-shop scheduling problem. Based on the modeling of the general Flow-shop scheduling problem, a model is established, and an improved algorithm is proposed which combines the advantages of simplex search and particle swarm optimization. An improved particle swarm optimization algorithm for force and efficiency is proposed. A class of classical examples are used to test and verify the proposed algorithm. The results are compared with those of the model in the literature. It is concluded that the traditional model is difficult to solve and the quality of the solution is not high with the increase of the scale of the problem. This paper briefly describes and analyzes the production of modern aluminum industry, simplifies the production process according to the actual production process, extracts the production process of industrial aluminum ingot produced by bauxite in an aluminum plant, and abstracts it into a multi-product batch production mode coordinated by concurrent equipment in each stage after analysis and treatment, and then uses batch production. Processing method is used to calculate and determine the processing time and energy consumption of corresponding batches in batches. Finally, it is transformed into a classical mixed Flow-shop scheduling problem. From the point of view of energy-saving, a mathematical programming model with minimum energy consumption as the scheduling objective is established. The improved PSO mixed with simplex method is invoked to solve the problem, and the solution results and basis are obtained. The GA results are compared and analyzed to further verify the superiority of PSO and its ability to solve such scheduling problems. Finally, the full text is summarized, and the future development and application of the research problems are prospected.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP18;F273;F426.32

【参考文献】