基于离散类电磁机制算法的装配序列规划研究

发布时间：2018-07-14 14:08

【摘要】：装配序列规划(Assembly Sequence Planning,ASP)问题一直是装配工艺中需要考虑的主要问题之一。优秀的装配序列规划方案能够有效减少产品生产装配过程中的冲突,提高生产效率,增强企业的市场应变能力,帮助企业更好地适应日益多样化的市场需求。本文研究了利用离散类电磁机制(Discrete Electromagnetism-like Mechanism,DEM)算法解决ASP问题。首先,概述了装配序列规划问题和类电磁机制算法及其研究现状。其次,介绍了基于连接体的装配序列规划。ASP问题属于组合优化问题,其复杂程度随着所研究对象的信息数量水平的提高而大幅度提高。这就使得求解具有一定复杂度的ASP问题时,需要花费较多时间来寻找合适的最优解。为了在考虑更多约束信息的情况下,降低ASP问题的复杂程度,本文在ASP问题的建模过程中引入了连接体概念,有效地降低了ASP问题的复杂程度。接着,介绍了EM算法的原理,并提出离散EM算法。EM算法是一种基于种群的随机全局优化方法。该算法通过模拟电磁场中带电粒子之间“吸引—排斥”机制,引导带电粒子向最优点处移动,因此称为类电磁机制算法。ASP问题属于离散问题,不适合用传统EM算法直接求解。本文主要研究了该算法,通过重新定义解的距离和移动、引入记忆机制、改进电量和合力的计算公式等途径,将针对连续优化问题的类电磁机制算法进行离散化,后使用该算法解决基于连接体的装配序列规划问题。实验表明,离散类电磁机制算法与引导式遗传算法、基因算法和人工免疫算法相比,在求解精度和求解时间上都有一定优势。然后,使用离散EM算法计算了三个实例,通过与其它优化算法的对比分析,验证了离散EM算法求解ASP问题的有效性。最后,总结全文。
[Abstract]:Assembly sequence planning is one of the main problems in assembly process. The excellent assembly sequence planning scheme can effectively reduce the conflicts in the process of product production and assembly, improve the production efficiency, enhance the market adaptability of enterprises, and help enterprises better adapt to the increasingly diverse market demand. In this paper, the application of discrete electromagnetic mechanism (discrete electromagnetic mechanism) to solve ASP problem is studied. First, the assembly sequence planning problem, electromagnetic mechanism algorithm and its research status are summarized. Secondly, it is introduced that the assembly sequence planning based on connectors. ASP is a combinatorial optimization problem, and its complexity increases greatly with the increase of the information quantity level of the objects studied. In order to solve the ASP problem with certain complexity, it takes more time to find the appropriate optimal solution. In order to reduce the complexity of ASP problems, the concept of connectors is introduced into the modeling process of ASP problems, which effectively reduces the complexity of ASP problems. Then, the principle of EM algorithm is introduced, and the discrete EM algorithm. EM algorithm is a stochastic global optimization method based on population. By simulating the mechanism of "attraction and repulsion" between charged particles in electromagnetic field, this algorithm can lead charged particles to move to the best point. Therefore, the problem called electromagnetic mechanism algorithm. ASP is a discrete problem, which is not suitable to be solved directly by traditional EM algorithm. This paper mainly studies this algorithm. By redefining the distance and movement of the solution, introducing the memory mechanism, improving the calculation formula of electric quantity and resultant force, the algorithm of electromagnetic mechanism for continuous optimization problem is discretized. Then the algorithm is used to solve the assembly sequence planning problem based on connectors. The experimental results show that the discrete electromagnetic mechanism algorithm has some advantages in solving precision and time compared with the leading genetic algorithm, genetic algorithm and artificial immune algorithm. Then, three examples are calculated by using the discrete EM algorithm, and the effectiveness of the discrete EM algorithm in solving ASP problem is verified by comparing with other optimization algorithms. Finally, the full text is summarized.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH182

【参考文献】