基于离散微粒群算法和混合差分进化算法的复杂生产调度问题求解

发布时间：2018-01-08 06:31

本文关键词：基于离散微粒群算法和混合差分进化算法的复杂生产调度问题求解　出处：《昆明理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：车间生产调度问题一直是学术界和工业界广泛关注的热点问题,随着近期的发展,由于其NP-hard性、种群规模巨大、约束性强等特性,求解过程的限制问题也越来越突出。而针对此问题的相关智能调度算法研究也逐渐引起两界重视。离散微粒群算法(Discrete Particle Swarm Optimization, DPSO)和差分进化算法(Differential Evolution, DE)都是应用颇多且效果很好的智能算法,目前已备受学者们的青睐,并成功运用于多个生产研究系统。本论文针对两类重要的流水线车间调度问题,采用基于不同改进方案的两种算法进行了研究,其中包括基于混合差分进化算法的求解研究和基于混合离散微粒群算法的求解研究。论文的主要工作归纳如下：(1)针对可重入流水线调度问题,通过分析问题的求解过程,开发了一种基于交换邻域搜索的局部搜索机制,将其与离散微粒群算法的全局搜索策略有机结合,并通过仿真实验加以证明,得到结果比较验证了所提算法的优越性。(2)针对重入次数可变的流水线调度问题,根据问题模型特点,将重入次数设为可变,增加问题难度,利用差分进化算法的全局搜索机制,结合前端省略搜索机制,提出了一种高效混合差分进化算法,仿真实验和结果比较验证了所提算法的有效性和稳定性。针对广泛存在于流程工业的可重入流水线调度问题,目前尚无基于差分进化算法和离散微粒群算法的相关研究,因此,本论文将已有差分进化算法和自已独立设计的局部搜索融合在一起,同时也将已有离散微粒群算法进行改进,其研究具有重要学术价值和工程实用价值,具有一定创新性。
[Abstract]:The workshop production scheduling problem has been a hot issue of widespread concern in academia and industry, with the recent development of NP-hard, because of its huge population, and constrained capability, limited problem solving process is more and more outstanding. And Research on Intelligent Scheduling Algorithm for this problem has gradually attracted the attention of two. Discrete particle swarm algorithm (Discrete Particle Swarm Optimization, DPSO) and differential evolution algorithm (Differential Evolution DE) is a lot of intelligent algorithms and the effect is very good, has been praised by the scholars of all ages, and successfully applied in production of system. This dissertation focuses on two kinds of important flow shop scheduling the problem, using two kinds of algorithm based on improved scheme is studied, including the solution of the algorithm based on mixed discrete and mixed difference based on Particle Swarm Optimization for Study on the solution. The main work is summarized as follows: (1) the problem of re entrant pipeline scheduling, analyzing the solving process problems, the development of a local search mechanism exchange neighborhood search based on the discrete particle swarm algorithm combining global search strategy, and proved through the simulation experiment, the superiority of the the result is verified by the algorithm. (2) according to the flow shop scheduling problem into the number of variable, according to the characteristics of the model, the revisiting times for variable, increase the difficulty, using differential evolution algorithm global search mechanism, combined with the front end omitting search mechanism, this paper presents an efficient hybrid differential evolutionary algorithm, simulation experiment and comparison results show that the proposed algorithm is effective and stable. According to the widely existed in process industry re entrant problem there is no pipeline scheduling, based on differential evolution The related research, discrete particle swarm optimization algorithm and therefore, this paper will have poor local search algorithm and their own independent design together, at the same time will also be improved discrete particle swarm optimization algorithm, has important academic value and practical value of the research, which has certain innovation.

【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】：TP18;TB497

【共引文献】