基于改进人工蜂群算法的机电产品并行拆卸序列规划研究

发布时间：2018-08-18 18:22

【摘要】：随着我国制造业的不断发展,越来越多的报废机电产品带来的资源再利用问题与潜藏的环境污染问题亟待解决,而拆卸是解决这一系列问题的基础与关键。如何对废旧机电产品进行高效的拆卸利用已经是现今“绿色技术”的研究热点,也是机电产品生命周期研究中所面临的重要课题。因此,本文致力于探索更为高效的拆卸分析及序列规划方法,以改进人工蜂群算法为基础,对复杂机电产品的并行拆卸序列规划方法进行了研究。首先,对复杂产品的拆卸信息模型及评价指标进行了探讨,为后续拆卸序列规划的研究奠定了基础。其次,针对拆卸序列规划易出现的组合爆炸问题,提出了基于人工蜂群算法的解决方案。同时,针对此算法的不足之处,提出了基于二叉树的优先约束策略,对算法中随机生成的初始种群进行约束化处理,使之能够有效提取拆卸解空间信息;定义了可行度算法与适应度算法协同搜寻蜜源,在保证种群多样性的前提下提高了改进算法的收敛速度,使之更加适合求解拆卸序列规划问题。最后,通过实例计算并与其他算法进行对比,证明了本文所述改进算法的可行性及高效性。在此改进算法的基础上,本文提出了一种适应于复杂机电产品的并行拆卸序列规划方法。通过对比分析并行拆卸与传统串行拆卸的不同之处,就并行拆卸的关键问题提出了相应的解决方案:定义了可变序列矩阵方法,由此提出了一种单一集合的并行化分析方法,用以解决并行拆卸序列长度与每步步长不确定的问题,建立了串、并行拆卸序列分析之间的联系纽带;针对并行拆卸繁多的单元选取情况,提出诱导因子方法,引领规划过程向最优方向进行演化。最后,于改进人工蜂算法的基础上,对某发动机模型进行并行拆卸序列规划求解,证明了本文所述基于改进算法的并行拆卸方法的优越性。
[Abstract]:With the continuous development of manufacturing industry in China, more and more problems of resource reuse and latent environmental pollution caused by end-of-life mechanical and electrical products need to be solved, and disassembly is the basis and key to solve these problems. How to disassemble and utilize the waste electromechanical products efficiently has become the research hotspot of "green technology" nowadays, and it is also an important subject in the research of the life cycle of electromechanical products. Therefore, this paper is devoted to exploring more efficient disassembly analysis and sequence planning methods. Based on the improved artificial bee colony algorithm, the parallel disassembly sequence planning method for complex electromechanical products is studied. Firstly, the disassembly information model and evaluation index of complex products are discussed, which lays a foundation for the research of disassembly sequence planning. Secondly, a solution based on artificial bee colony algorithm is proposed to solve the combinatorial explosion problem which is easy to occur in disassembly sequence planning. At the same time, aiming at the shortcomings of the algorithm, a priority constraint strategy based on binary tree is proposed. The initial population generated randomly in the algorithm is constrained to extract the information of disassembly solution space effectively. The feasibility algorithm and fitness algorithm are defined to search honey source in cooperation. The convergence speed of the improved algorithm is improved on the premise of ensuring diversity of population, which makes it more suitable for solving disassembly sequence planning problem. Finally, the feasibility and efficiency of the improved algorithm are proved by example calculation and comparison with other algorithms. Based on the improved algorithm, a parallel disassembly sequence planning method for complex electromechanical products is proposed. By comparing and analyzing the differences between parallel disassembly and traditional serial disassembly, the corresponding solutions to the key problems of parallel disassembly are put forward: the variable sequence matrix method is defined, and a single set parallelization analysis method is proposed. In order to solve the problem that the length of parallel disassembly sequence and the length of each step are uncertain, the link between serial disassembly sequence analysis and parallel disassembly sequence analysis is established. Leading the planning process to the optimal direction of evolution. Finally, on the basis of the improved human worker bee algorithm, the parallel disassembly sequence planning of an engine model is carried out, which proves the superiority of the parallel disassembly method based on the improved algorithm in this paper.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;X705

【参考文献】