带有返工工件的单机重调度问题

发布时间：2018-12-21 19:50

【摘要】：针对复杂多变的车间生产环境,制定行之有效的生产调度计划是企业生产管理的核心需求和学术研究的热点问题。车间生产中经常会遇到由于有突发事件的干扰,破坏了最初生产调度的最优性,甚至使其成为不可行调度。此时,需要针对干扰做出必要的反应,在满足车间生产的各种约束条件下,通过重调度的方法修复初始调度,最终制定适合实际生产要求的最优或者近优调度计划。对于重调度的研究不但具有重要的理论价值而且具有现实的应用意义。在单机环境下的离散制造车间实际生产中,经常有一类不合格工件通过返回到该工位进行简单修复即可成为合格工件的情况,因此需要在初始调度的基础上和满足各种约束的前提下调度这些返工工件。本文研究了针对带有返工工件的单机重调度问题(RRSM-Rescheduling for rework jobs on single machine)。针对RRSM本身问题的特点,分别设计了基本的遗传算法、规则引导的遗传算法和规则引导的自适应遗传算法,三种智能算法来求解该问题,并对该问题进行了大量的仿真实验,验证了三种算法的求解性能。首先在小规模的工件总数下,通过均匀设计试验方案和大量的实验算例确定三种遗传算法的最优参数组合。然后通过均匀设计试验方案选取小规模工件总数下不同的算例参数组合,利用已确定的最优参数组合下的三种遗传算法对不同的算例进行仿真实验。最后在给出一个代表性的算例参数组合下,增大问题规模,进行三种遗传算法的仿真实验。实验结果表明,规则引导的自适应遗传算法求解的效率和解的质量明显的优于前两种遗传算法。为进一步验证规则引导的自适应遗传算法的有效性,通过对给出的多个算例进行仿真实验,将规则引导的自适应遗传算法和已有的启发式算法和分支定界算法所得的结果进行了对比和分析。同时,在工件总数不同的规模下统计分析了规则引导的自适应遗传算法的稳定性。实验结果表明,规则引导的自适应遗传算法的求解效果较好并且表现出了很好的稳定性。
[Abstract]:In view of the complex and changeable workshop production environment, making effective production scheduling plan is the core requirement of enterprise production management and a hot issue of academic research. Because of the disturbance of unexpected events in workshop production, the optimization of initial production scheduling is destroyed, and even becomes infeasible scheduling. In this case, we need to make the necessary response to the interference and repair the initial scheduling through rescheduling under various constraints of the workshop production. Finally, the optimal or near-optimal scheduling plan suitable for the actual production requirements is formulated. The research on rescheduling not only has important theoretical value but also has practical significance. In the actual production of discrete manufacturing workshop in a single machine environment, there are often a class of unqualified workpieces which can become qualified workpieces by returning to the station for simple repair. Therefore, it is necessary to schedule these rework jobs on the basis of initial scheduling and satisfying various constraints. In this paper, we study the single machine rescheduling problem with rework workpiece (RRSM-Rescheduling for rework jobs on single machine). According to the characteristics of RRSM itself, the basic genetic algorithm, rule-guided adaptive genetic algorithm and three intelligent algorithms are designed to solve the problem, and a large number of simulation experiments are carried out to solve the problem. The performance of the three algorithms is verified. Firstly, the optimal parameter combination of the three genetic algorithms is determined by uniform design test scheme and a large number of experimental examples under the small total number of jobs. Then, through the uniform design test scheme, the parameters of different examples are selected under the total number of small scale workpieces, and three genetic algorithms are used to simulate the different examples by using the determined optimal parameter combination. Finally, a typical example is given to increase the scale of the problem, and the simulation experiments of three genetic algorithms are carried out. The experimental results show that the efficiency and quality of the rule guided adaptive genetic algorithm is better than that of the first two genetic algorithms. In order to further verify the validity of the adaptive genetic algorithm guided by rules, the simulation experiments are carried out on several given examples. The results obtained from the rule guided adaptive genetic algorithm and the existing heuristic algorithm and branch and bound algorithm are compared and analyzed. At the same time, the stability of the adaptive genetic algorithm guided by rules is statistically analyzed under the different size of the total number of jobs. The experimental results show that the rule guided adaptive genetic algorithm is effective and stable.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB497

【相似文献】