离散制造车间作业调度系统的研究与实现

发布时间：2018-07-16 13:42

【摘要】：随着以动态化、全球化和用户驱动为显著特征的市场竞争的加剧，离散制造系统的复杂性与动态性不断增加，因此对车间的生产过程管理提出了更高的要求。车间作业调度是离散制造业生产管理的核心内容之一，它对资源优化利用，提高工作效率，节省成本都有非常重要的作用。但经调研发现，目前大部分离散车间作业调度系统与实际生产环境脱节，调度系统难以根据车间现场的变化及时修正，导致调度结果的误差不断积累。为解决目前调度系统存在的问题，本课题研究工作的主要内容包括以下四个方面： (1)针对车间生产过程中数据信息不准确、反馈实时性差的问题，采用RFID技术和Zigbee技术实现对离散制造过程和设备、工人双资源状态的唯一标识和追踪，采集车间现场加工过程动态数据，并实时反馈至调度系统，从而形成闭环的调度过程。针对Zigbee网络在车间的应用特点，提出了一种基于网格和虚拟力导向蚁群算法的高能效路由策略，使其在降低能耗和延长网络生存时间上有更好的表现。 (2)针对离散制造车间充满动态性的特点，利用多Agent技术具有自主性、协作性的特点，建立了混合型多Agent作业调度系统模型，整个调度系统建立在管理Agent、计划Agent、库存Agent、任务Agent、工艺Agent、数据采集Agent、设备Agent、工人Agent、调度Agent的基础上，研究了各Agent的功能、结构模型及基于合同网的Agent交互行为。在此基础上，介绍了作业调度系统中在设备故障和订单插入动态事件扰动下，多Agent的协作过程。 (3)针对离散制造车间设备和工人双资源约束的车间作业多目标调度问题，提出一种基于模糊Pareto支配的生物地理学优化算法封装在调度Agent中，详细描述了该算法的在车间作业调度中的应用流程。通过仿真实验，验证了该算法可以在平稳状态和设备故障扰动状态下，完成车间作业的调度和设备、工人双资源的有效分配。 (4)最后，，针对某离散制造模具车间作业调度现状与需求，结合车间实际需求及多Agent作业调度系统的特点，对系统的体系结构、功能模块、数据采集终端的选取、数据库的设计几个方面进行了讨论，在此基础上，开发并实现了基于多Agent的作业调度系统。
[Abstract]:With the aggravation of market competition, which is characterized by dynamic, globalization and user driving, the complexity and dynamics of discrete manufacturing system (DMS) are increasing, so higher requirements are put forward for workshop production process management. Job shop scheduling is one of the core contents of production management in discrete manufacturing industry. It plays an important role in optimizing the utilization of resources, improving work efficiency and saving costs. However, it is found that most of the discrete job shop scheduling systems are out of touch with the actual production environment, and it is difficult for the scheduling system to be corrected timely according to the changes in the job shop site, resulting in the accumulation of errors in the scheduling results. In order to solve the problems existing in the current scheduling system, the main contents of this research work include the following four aspects: (1) aiming at the problem of inaccurate data information and poor real-time feedback in the process of workshop production, RFID technology and Zigbee technology are used to realize the unique marking and tracing of discrete manufacturing process and equipment and workers' dual resource state. The dynamic data of workshop field processing process are collected and fed back to the scheduling system in real time to form a closed loop scheduling process. According to the application characteristics of Zigbee network in workshop, an efficient routing strategy based on grid and virtual force-oriented ant colony algorithm is proposed. It has better performance in reducing energy consumption and prolonging the network lifetime. (2) aiming at the dynamic characteristics of discrete manufacturing workshop, using multi-Agent technology has the characteristics of autonomy and cooperation. A hybrid multi-Agent job scheduling system model is established. The whole scheduling system is based on the management, planning, inventory, task, process, data acquisition, equipment, worker and Agent scheduling of Agent. Structure model and Agent interaction behavior based on contract net. On the basis of this, this paper introduces the dynamic event disturbance of equipment failure and order insertion in the job scheduling system. The collaborative process of multiple Agent. (3) aiming at the multi-objective scheduling problem of job shop with the constraints of discrete manufacturing workshop equipment and workers, a biogeographic optimization algorithm based on fuzzy Pareto domination is proposed and encapsulated in the scheduling Agent. The application flow of the algorithm in job shop scheduling is described in detail. The simulation results show that the algorithm can complete the scheduling of workshop jobs and the efficient allocation of workers' dual resources in stationary state and equipment fault disturbance state. (4) finally, According to the present situation and demand of job scheduling in a discrete manufacturing mould shop, combined with the actual demand of workshop and the characteristics of multiple Agent job scheduling system, the system structure, function module and data acquisition terminal are selected. Several aspects of database design are discussed. On this basis, a job scheduling system based on multiple Agent is developed and implemented.
【学位授予单位】：江南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB497

【参考文献】