基于人工鱼群算法的柔性作业车间调度研究

发布时间：2017-12-27 15:18

本文关键词：基于人工鱼群算法的柔性作业车间调度研究　出处：《大连理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：车间调度是通过合理安排各种生产资源以满足企业生产的某些性能指标,它是制造企业提升自身市场竞争力的关键因素。相对于传统调度问题,柔性作业车间调度问题增加了加工机器柔性的特性,使其更贴近企业的现实生产模式,因而对它的研究更具实际应用价值。本文以一种新型的群智能算法—人工鱼群算法为基本优化算法,分别针对柔性作业车间调度中的单目标和多目标两类问题模型展开讨论,本文的主要工作概述如下：(1)对于柔性作业车间调度问题,加工机器选择子问题的解决会影响到工序的加工顺序子问题的求解,反之亦然,因此两个子问题之间是相互制约和相互影响的。本文提出了前置安排策略和后置安排策略,它们分别以不同的先后顺序处理两个子问题从而产生不同的调度方案,保证了种群的多样性。(2)在求解单目标柔性作业车间调度问题时,本文设计了一种基于分布估计的人工鱼群算法,该算法是对基本人工鱼群算法的一种改进：为提高算法搜索的导向性设计了带有分布估计能力的觅食行为,为加强算法的全局搜索能力提出了人工鱼吸引行为,加入了基于关键路径的局部搜索以均衡算法探索和开发能力。使用160个经典用例对提出的算法进行实验,通过与其他优化算法地比较,证明了算法求解单目标问题的有效性。(3)针对最大完工时间、最大机器负载、总机器负载三个目标的柔性作业车间调度模型,受协同进化思想地启发,提出了一种协同混合人工鱼群算法；该算法在求解过程中通过鱼群的多种群协同进行全局搜索,并与模拟退火算法协同增强局部搜索能力,另外针对多目标问题设计了改进的ε—Pareto支配策略对适用度值进行评价,且在算法中采用拥挤距离和精英保留策略保持鱼群中个体的多样性；最后通过实验验证了该算法可以得到更优质的非劣解。
[Abstract]:Job shop scheduling is a key factor for manufacturing enterprises to enhance their market competitiveness by arranging various production resources to meet certain performance indicators of enterprises. Compared with traditional scheduling problem, flexible job shop scheduling problem increases the flexibility of machine processing, making it closer to the real production mode of enterprises, so the research on it is more practical. In this paper, the basic algorithm uses a novel swarm intelligence algorithm artificial fish swarm algorithm for the model, respectively, for the flexible job shop scheduling in single and multi objectives, two kinds of problems are discussed, an overview of the main work of this paper are as follows: (1) for the flexible job shop scheduling problem, machine selection method, the processing sequence of the sub problems will affect the process of the problem and vice versa, so between the two sub problems are interdependent and mutual influence. This paper puts forward the strategy of pre arrange and post arrange. They deal with two sub problems in different order, so as to generate different scheduling schemes and ensure the diversity of population. (2) in solving the multi-objective flexible job shop scheduling problem, this paper designs a kind of artificial fish swarm algorithm based on estimation of distribution, the algorithm is an improvement to the basic artificial fish swarm algorithm: design ability of foraging behavior with estimation of distribution oriented to improve the search algorithm, in order to strengthen the global search algorithm the ability to put forward the artificial fish attracting behavior, adding to the exploration and development of equalization algorithm based on local search ability of critical path. 160 classical use cases are used to experiment with the proposed algorithm, and the effectiveness of the algorithm is proved by comparing with other optimization algorithms. (3) for the flexible job shop scheduling model of the maximum completion time, the maximum machine load, the total machine load the three target, the idea of co evolution inspired, this paper proposes a collaborative hybrid artificial fish swarm algorithm; the algorithm through a variety of fish swarm CO with global search in the solution process, and simulated annealing algorithm enhance the ability of local searching, in addition to the multi-objective design e - Pareto improved control method to evaluate the fitness value, and the crowding distance and the elitist strategy to keep the diversity of the fish in the individual in the algorithm; it is proved by experiments that the algorithm can get better solution pareto.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP18;TB497

【共引文献】