成批服务排队网络下的虚拟单元构建与调度问题研究

发布时间：2018-05-10 07:22

本文选题：虚拟单元构建 + 跨单元调度　；参考：《江苏科技大学》2017年硕士论文

【摘要】：虚拟单元制造系统(Virtual Cellular Manufacturing System,VCMS)是单元生产方式在制造企业应用的进一步发展,得到了越来越多国家的关注和重视,然而由于工件到达的随机性,以及虚拟单元物理位置不改变而带来的复杂的车间流问题,虚拟单元制造系统中往往存在工件在机器前较长的排队等待情况。鉴于此,本文采用成批服务排队网络策略,着重考虑排队等待因素的影响,从而进一步优化复杂的单元路径选择和减少零件排队等待时间,显著提高设备利用率,缩短零件生产周期,企业快速响应顾客需求。当工件以一定批量随机达到以及在设备上加工时间不确定时,如果现有的虚拟单元不能够满足待加工工件的生产需求,该批量的工件需要排队等待加工。因此成批量服务的虚拟单元调度问题分为两个阶段,第一个阶段是在复杂的车间流中选择合适的加工路径进行虚拟单元构建,以减少排队等待的问题研究,第二阶段是在构建完成基础上,虚拟单元中排队等待加工的工件动态调度问题研究。关于虚拟单元构建问题,考虑车间复杂的车间流即生产路径,本文基于数学规划及网络图论等相关理论,综合零件的多工艺生产路线、生产批量、加工时间、排队等待费用、设备服务强度等因素,提出了一种成批服务排队网络下最优路径选择方法。在零件成批服务的策略下,构建了以零件加工成本最小化和排队等待费用最小化,以及设备服务强度最小化的多目标模型,提出了新的最优路径决策方法,筛选出了最优的加工路径,避免了零件排队等待。该方法实现了减少零件排队等待同时保证了单元车间复杂流的简单化。基于路径选择的基础上,进行生产制造资源的聚类分析,设计了基于马氏距离的零件族相似划分算法,实现了虚拟单元的构建。关于虚拟单元调度问题,本文将每个虚拟单元看作一个节点,考虑零件的排队等待,研究了成批服务排队网络下由于多工艺引起的跨单元生产调度问题。利用排队网络理论将每一个虚拟单元看作是一个节点,将虚拟单元制造系统作为一个整体排队网络,通过网络图论的方式将各单元节点等效为一个M/M/C排队模型,通过外部到达率计算网络内部各节点的输入参数,在工件成批服务的策略下,以零件跨单元加工的总加工流程时间最小、设备最大负荷最小化作为调度目标,构建了成批服务的排队网络模型。通过免疫遗传算法求解模型,实现了跨单元总流程加工时间最短,加工路径最优,减少了零件排队等待。最后,本文在对HDZH公司作业班组的机加工生产现状及存在问题进行详细分析的基础上,运用基于成批服务排队网络下虚拟单元构建和跨单元调度的方法,对造船企业的机加工轴系和舵系的生产过程进行分析,以验证本文所提研究方案在实际企业中的应用价值及意义。研究结果表明,本文所提出的研究方案具有较高的可行性和有效性。
[Abstract]:Virtual Cellular Manufacturing system (VCM) is the further development of the application of cell production in manufacturing enterprises, and has been paid more and more attention by more and more countries. However, due to the randomness of workpiece arrival, And the complex workshop flow problem caused by the unchanging physical position of the virtual cell. In the virtual cell manufacturing system, there is always a long queue waiting in front of the machine. In view of this, this paper adopts batch service queuing network strategy, emphasizes the influence of queuing and waiting factors, and further optimizes the complex cell path selection and reduces the queuing waiting time of parts, and improves the equipment utilization rate significantly. Shorten parts production cycle, enterprises respond to customer needs quickly. If the existing virtual unit can not meet the production demand of the workpiece to be processed, the batch workpiece needs to be queued for processing when the workpiece is reached randomly in a certain batch and the processing time on the equipment is uncertain. Therefore, the virtual cell scheduling problem of batch service is divided into two stages. The first stage is to select the appropriate processing path in the complex shop flow to construct the virtual cell, so as to reduce the problem of queue waiting. In the second stage, the dynamic scheduling problem of the jobs queued and processed in the virtual cell is studied on the basis of the completion of the construction. Regarding the problem of virtual cell construction, considering the complex workshop flow, that is, the production path, based on mathematical planning and network graph theory, this paper synthesizes the multi-process production route of parts, production batch, processing time, queue waiting cost, etc. In this paper, an optimal path selection method for batch service queuing network is proposed. Under the strategy of part batch service, a multi-objective model based on the minimization of parts processing cost and queue waiting cost, as well as the minimization of equipment service intensity, is constructed, and a new optimal path decision method is proposed. The optimal processing path is selected to avoid waiting in queue. This method can reduce the queue waiting of parts and ensure the simplification of complex flow in the unit workshop. On the basis of path selection, the clustering analysis of manufacturing resources is carried out, and the similarity partition algorithm of parts family based on Markov distance is designed to realize the construction of virtual cell. For the problem of virtual cell scheduling, this paper considers each virtual cell as a node, considers the queue waiting of parts, and studies the cross-cell production scheduling problem caused by multi-process in batch service queueing network. Each virtual cell is regarded as a node by using queuing network theory, and the virtual cell manufacturing system is regarded as a whole queueing network. Each cell node is equivalent to a M/M/C queuing model by network graph theory. The input parameters of each node in the network are calculated by the external arrival rate. Under the strategy of batch service of the workpiece, the minimum processing time and the minimum maximum load of the parts across the unit are taken as the scheduling objectives. The queuing network model of batch service is constructed. The immune genetic algorithm (IGA) is used to solve the model, which realizes the shortest processing time and the optimal machining path, and reduces the queue waiting of the parts. Finally, based on the detailed analysis of the production status and existing problems of job teams in HDZH Company, this paper applies the method of virtual cell construction and cross-cell scheduling based on batch service queuing network. The production process of machined shafting and rudder system in shipbuilding enterprises is analyzed in order to verify the application value and significance of the research scheme proposed in this paper in practical enterprises. The results show that the proposed scheme is feasible and effective.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F426.474;O226

【参考文献】