卷烟企业物流系统柔性调度管理研究

【作者】 顾红；

【导师】 邹平；

【作者基本信息】 昆明理工大学 ， 管理科学与工程， 2012， 博士

【摘要】 现代物流是指为了满足消费者需要而进行的全生命周期中的原材料中间过程与库存、最后产品和相关信息有效流动和储存的计划、实施和提高控制管理的过程。企业物流系统是指为了满足企业生产需要,有效达成物流目的的机制。资本的逐利性驱使企业必须持续获利。当从自然资源,譬如原材料、设施设备(第—利润源)和劳动力资源(第二利润源)中获利持续性难以为继时,新的利润增长点被锁定在物资的流通过程中,这里是尚未开垦的处女地,被视为新的第三利润源泉。所以如何最大限度的开发物流系统潜能,成为现在的热点问题之一。作者也在其中选取一个细微分支进行研究,以期能够做一点实在和有益的工作。本论文通过对卷烟企业物流系统中的中枢——自动化物流系统的研究,在对系统从结构到控制、调度、管理层的深入分析后,选取部分核心要素问题进行突破。主要工作内容如下：首先对整个烟草行业的物流状况进行回顾,发现目前的物流研究停留在宏观层面多,微观层面少；静态局部因素研究多,考虑物流系统的动态复杂性的少。故本论文从微观层面——卷烟企业物流系统着手进行全面系统的研究。论文共分为八章,全面系统论述了当前国内企业在全球经济一体化趋势下所面临的机遇和挑战,以及应对策略——开发“第三利润源”物流。并从卷烟企业的自动化物流系统着手,进行深入研究,主要创新工作如下：对卷烟企业物流系统的自动化立体库库的存储能力设计进行研究；并结合国家的相关行业管理要求进行生产工艺的柔性开发研究。从控制机关(controlling organ, CO)与目标系统(target system,TS)模型来进行探讨,用双向控制关系视野对系统上各层组织——计算机调度管理系统和自动控制系统柔性开发进行研究。基于随机概率对自动化立体库仓储系统的能力进行深入研究。给出表征实际工程状态的巷道堆垛机分区域能力计算的数学模型和算法。并给出能力计算的精度收敛方法和评价方法。最后通过实例,验证本文所建模型的随机概率算法的可行性和有效性。基于深入研究建立自动化立体库调度策略计算的数学模型以拓展柔性。运用机器学习和改进遗传算法理论,训练实例数据生成储存时间、储存货位的专家库知识,解决海量数据分析计算与快速实时调度矛盾；实现基于专家库自学习和遗传算法求最优解的自动化立体库调度。并通过实测数据实证了自动化立体库柔性调度研究模型的可行性和有效性。对物流系统中子系统——重要的搬运设备环行有轨自动车(The Rounding Rail Guided Vehicle, RGV)也称环行穿梭车,简称环穿,自动导引车(The Automatic Guided Vehicle,AGV)的工作特点和调度策略进行深入研究：创建了环行RGV系统搬运作业的多目标优化调度数学模型和有容积约束的AGV作业调度问题的数学模型。并给出基于自学习和改进遗传算法的环行RGV系统实时调度算法,以及基于改进遗传算法和妥协方法的有容积约束AGV作业调度算法；并讨论调度算法的编码、选择、交叉变异操作规则方法,提出改进种群和妥协解的专家库自学习方法。最后用实测数据进行验证研究证明模型是可行和有效的。为更好开发行业效益和效率,应该有效把握目前在物流领域的机遇,国内企业有很多事可做。在既定条件下,如何从物流系统中最大限度的开发尚未利用的柔性,从技术和管理的角度,提升物流效率,以提高企业的核心竞争力,是一个永恒的议题。更多还原

【Abstract】 Modem logistics is the process of planning, implementing, and controlling the efficient, effective flow and storage of goods, services, and related information from point of origin to point of consumption for the purpose of conforming to customer requirements in the whole life cycle. Logistics system can be considered as the mechanism of how to achieve logistics purpose effectively. Enterprise must continue profit base on capital nature of profit. When it becomes more and more difficult to pursue profits from nature resources managing, such as raw materials and establishments (that is, the first kind of resources)and the resource of labor forces managing (that is, the second kind of resource), many opportunities of obtaining profits are found in the process of logistics, maybe think of good scopes to explore, which are also called the space of decreasing cost, or the third kind of resources. So, how to using modem scientific theory and practice to tap logistics system has arouse more and more attention. The author chose one of the field to study in order to find something benefit to enterprise.Through in-depth study of the automatic logistics system as key in the tobacco enterprise, some problem can be solved after through the in-depth analysis from structure to control, scheduling and manage about it. The content of this thesis is described as below:This thesis is composed of nine chapters and elaborates upon issues is related to currently opportunity and challenge of the enterprise in china of the international economic incorporate. So we must tap "the third kind of resources (logistics system)".The author choses the cigarette manufacture logistics system of tobacco industry carry through elaborates research, has accomplished the following research objectives:Comprehensively reviewed theories related to logistics system in tobacco industry. The previous theories in this area need further instructions on how to guide the logistics process and management. In the context of globalization, their weakness demonstrated in two aspects:focus on framework but lack of embody; focus on static state and certain factors but lack of dynamic complexity.The quantificational study for storage capability of the automatic tiered warehouse about the finished product and auxiliary material is put up, according to the policy and rule of our state, the flexible process of it is also studied. In addition, this research results cover all the organization. All researches view the modle of Controlling Organ(CO) and Target System (TS) on framework.According to the scheduling management strategy, the random probability of actual SRM machine performance, and the working flow of SRM machine as well, are analyzed. The mathematical model and algorithm was set up on automatic logistic tiered warehouse storage system capacity based on random probability. Precision convergence method and evaluation process of capacity calculation were given as well. The discrepancy between planning and design ability and actual system ability is solved. Finally, by means of the finished product warehouse of certain cigarette factory, calculation result of estimating algorithm and random probability algorithm were compared with the actual measurement data respectively, which proved that result of random probability algorithm was closer to actual running status and it was more feasible and practicable.To improve the access efficiency, a new scheduling policy was proposed. The model is based on storage time after in-depth study on the scheduling tactics. By using the self-learning and the improved genetic algorithms, the storage time generate from the trained example data and the expert database knowledge of the storage cargo space, the conflict between the massive data analysis and rapid real-time scheduling is solved. Thus the tiered warehouse scheduling based on the expert database self-learning and genetic algorithms optimal solution is achieved. In addition, the results have been applied in the automatic tiered warehouse in a cigarette factory, its access ability is effectively improved, and the example verifies that the research is feasible and effective.Through the in-depth analysis on working trait and scheduling policy of the rounding Rail Guided Vehicle (RGV) and the Automatic Guided Vehicle (AGV), and considering the round RGV for the start-stop, waiting and complex task, a multi-objective genetic algorithm (GA) is proposed for RGV scheduling problems. Based on self-learning and the improved GA, the real-time scheduling policy of round RGV is provided. And considering the capacitated of AGV transportation, a new scheduling policy was proposed for its transit the scrap in cigarette factory following one by one manufacture unit, a capacitated genetic algorithm model is proposed for AGV scheduling problems.The rule of encoding, selecting, crossover selection and variation probability was studied. The way of improving first strip based on self-learning expert database and searching the optimal compromising solutions was given to enhance calculation efficiency. In conclusion the research results of the simulation calculation and application in the cigarette factory have been effectively improved, and the example proves that the research is feasible and effective.We should hold the good opportunity of the logistics to meet the need of enterprise. We can do much thing about how to maximum tap the flexible of the logistics system under certain term. And we can advance logistics industry in the economic systems effectively to enhance improve the core competitiveness in tobacco industry from the point virew for technology and management. This is a topic full of eternal fascination.更多还原