基于混合离散微粒群算法求解复杂并行机生产调度问题

发布时间：2018-05-17 04:16

本文选题：并行机调度 + 到达时间　；参考：《昆明理工大学》2015年硕士论文

【摘要】：近年来,作为研究热点的非传统生产调度的相关问题,如复杂并行机生产调度问题等,受到了研究学者的广泛关注。离散微粒群(Discrete Partial Swarm Optimization Algorithm, DPSO)作为一种简单有效的人工智能算法,已受到机械加工和钢铁冶炼等多个工业领域得到了成功应用,对于并行机生产调度问题的DPSO已经成为国际上前沿研究课题。因此,本论文对两类重要并行机生产调度问题进行基于DPSO的求解算法研究。论文的主要工作归纳如下：(1)针对最大完成时间(makespan)指标下的带多工序和加工约束并行机调度问题,通过分析问题的结构特性,重新设计了一种微粒的位置更新公式,进而与DPSO的全局搜索有机结合,得到IDPSO,通过随机生成测试数据进行仿真实验并与其他算法进行比较验证IDPSO的有效性。(2)针对makespan指标下的带到达时间、多工序、加工约束和序相关设置时间的复杂并行机生产调度问题,将(1)中的微粒位置更新方法应用到该问题中,并加入首次改进跳出策略和基于Interchange和Insert的局部搜索方法,进而结合IDPSO的全局搜索机制,提出HDPSO,通过比较仿真实验结果,验证了HDPSO相对于其他算法的高效性和有效性(3)针对makespan指标下的(2)中的复杂并行机生产调度问题,分析DPSO中惯性权重、学习因子对算法的影响,提出一种自适应微粒群算法,进而与(2)中提出的HDPSO有机的融合,从而得到一种AHDPSO,仿真实验验证了AHDPSO加入自适应的必要性和有效性。由文献调研可知,特别是基于DPSO算法求解复杂并行机生产调度问题的研究十分有限,有些较为复杂的问题甚至处于空白状态。本文针对上述调度问题的数学模型将已有的DPSO算法进行改进,因此针对上述的复杂并行机生产调度的研究具有实际的工程价值和学术价值。
[Abstract]:In recent years, non-traditional production scheduling issues, such as the production scheduling of complex parallel machines, which are the focus of research, have been widely concerned by researchers. As a simple and effective artificial intelligence algorithm, discrete particle swarm optimization (Partial Swarm Optimization Algorithm, DPSO) has been successfully applied in many industrial fields such as mechanical processing and iron and steel smelting. DPSO for parallel machine production scheduling problem has become an international frontier research topic. In this paper, two important parallel machine scheduling problems are studied based on DPSO. The main work of this paper is summarized as follows: (1) aiming at the parallel machine scheduling problem with multiple working procedures and processing constraints under the maximum completion time (MCP) index, a new updating formula for the position of particles is designed by analyzing the structural characteristics of the problem. Then, combining with the global search of DPSO, we get IDPSO. through random generating test data for simulation experiment and comparing with other algorithms to verify the effectiveness of IDPSO. 2) aiming at the makespan index with time of arrival, multi-working procedure, In the production scheduling problem of complex parallel machines with processing constraints and order correlation setting time, the particle position updating method in F-1) is applied to this problem, and the first improved jump out strategy and local search method based on Interchange and Insert are added. Combined with the global search mechanism of IDPSO, HDPSO is proposed. By comparing the simulation results, it is verified that the efficiency and effectiveness of HDPSO compared with other algorithms are 3) for the complex parallel machine scheduling problem under the makespan index. After analyzing the influence of inertia weight and learning factor on the algorithm in DPSO, an adaptive particle swarm optimization algorithm is proposed, which is combined with the HDPSO proposed in HDPSO. The simulation results show that it is necessary and effective to add adaptive AHDPSO. According to the literature investigation, especially the research on the production scheduling problem of complex parallel machines based on DPSO algorithm is very limited, and some more complex problems are even in blank state. In this paper, the existing DPSO algorithm is improved for the mathematical model of the above scheduling problem, so the research on the production scheduling of the complex parallel machines has practical engineering value and academic value.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F425;TP18

【参考文献】