【作者】 任鸣鸣；

【导师】 杨超；

【作者基本信息】 华中科技大学 ， 管理科学与工程， 2008， 博士

【摘要】 在经济全球化快速推进、市场竞争日益加剧和客户对供货时间越来越敏感的背景下,企业之间的竞争转向供应链与供应链之间的竞争。供应链系统的网络拓朴结构是供应链整体运作的实体支撑,对供应链整体的响应速度、服务质量、运作成本、运作效率和一体化运作的流畅性等有重要影响作用,因此,优良的网络结构是供应链系统合理化和科学性运作的前提和保证。本文运用供应链管理理论和优化决策理论,对现有服务设施选址的研究现状与存在问题进行认真分析的前提下,依据供应链不同层次结构节点设施(如生产设施、物流设施、零售设施等)的功能与特性,对供应链系统节点设施选址模型与算法做了一些基础性和创新性研究。首先,分析了供应链系统设施选址中设施容量确定的局限性,把设施容量也作为决策变量,研究了静态环境和需求不确定环境下的供应链系统节点设施选址和容量的一体化决策问题,以固定成本和运营成本最小为目标,分别建立相应的数学模型,此处的固定成本随容量扩大而增加,运营成本随生产量的增加而减小,设计模型求解的Benders分解算法,并用算例进行仿真计算,说明算法的有效性和实用性。其次,论文分析了供应链系统动态多变特性对供应链终端设施选址的影响,不同阶段客户需求分布和设施运营成本有较大差异,研究了适应供应链环境变动的多阶段竞争性设施选址问题。设计了较为全面的设施引力评价指标体系,提出了设施引力的测定方法,建立客户选择服务设施的效用函数,在此基础上构建多阶段竞争性设施选址模型。分析模型特点,设计了模型求解的分散搜索算法。一组算例分别用Lingo和分散搜索算法计算比较,检验分散搜索算法的求解速度和收敛性。第三,论文分析了时间在供应链竞争中的作用,研究了基于时间约束的多级供应链系统物流中心调整问题以及工厂和物流中心一体化的调整问题,对这两个问题分别建立混合整数规划模型,考虑供应链中物流资源的整合利用以及关闭物流中心的成本,通过引入参数,把模型中的时间约束问题转化成为无时间约束问题,以降低模型的求解难度,并以实例分析基于不同时间约束的各成本及在总成本中所占比重。第四,生产者责任延伸制下,供应链系统不仅要承担产品的库存、运输和配送等物流任务,废弃产品的回收再资源化也是供应链管理的重要内容。本文分析了单个企业独立实施废弃产品回收存在的问题,以多企业联合回收废弃产品为着眼点,研究了多回收源点、多种废弃产品的处理站选址问题,以回收成本最小为优化目标,分别建立了每种废弃品有最大处理容量限制以及有最大和最小处理容量限制的处理站选址模型,分析两个模型特点,设计拉格朗日松驰启发式算法,并用一组算例对算法的可行性进行检验。本文所做的一些研究是对现有设施选址模型与算法的深入和发展,弥补了供应链系统设施选址研究方面的不足,使选址模型和算法能更好地应用于解决实际问题。本文所做的研究有利于优化供应链系统的服务设施规划和进行战略调整,实现客户需求分布与设施布局的有效均衡,提高对客户需求的反应速度,减少物流成本,增强供应链整体的竞争能力。更多还原

【Abstract】 The competition among supply chains substitutes for the competition among corporations under the surroundings of global economy, increasingly intense competition of marketing and decreasing response time from customers sending out demand to meeting their demand. The network framework of supply chain system that consist of facilities is playing an import work for the supply chain operation such as clipping response coming from customs’demand, improving service, decreasing cost, upgrading work efficiency, and smoothing manipulate in the integrated supply chain. Thus an excellent design of supply chain network is basic guarantee for the scientific and economic operation of the supply chain system.In this paper, several supply chains facilities location models and efficient algorithms for solving these models are developed by the facilities’characteristic and function at the special sites in the supply chain, which large numbers of existing location articles are studied and some out of line for supply chain facilities location are analyzed before writing this paper. Supply chain management theory, optimization and decision-making science are basic tools for these researches. Firstly, the limitation of decided capacity for production and logistics facilities in location research is analyzed and then the problems of facilities locations and capacity decision-marking are studied for supply chain system for certain and uncertain surroundings. Two mixed-integer programming models which the capacities are also decision-making variable are built for certain and uncertain surroundings, Benders decomposed algorithms are developed, and a set of examples are solved for testing these algorithms. Secondly, the dynamically changing trait of supply chain systems for the effect of terminal facilities is analyzed. There are considerable difference about demand and running cost at different stages so multistage competitive facilities locations problem in supply chain terminal is researched. A criterion system for facilities’attraction is designed, a measure for facilities attraction is proposed, and an extended utility function is developed. Based on these studies, a multistage competitive facilities locations model is developed, an effective scatter search heuristic algorithm is devised according to the model property, and also compares the results of scatter search heuristic algorithm solving examples for Lingo solving to same examples so that can show the solving speed and convergence for the provided algorithm. Thirdly, detailed study the time function in the competitive supply chains. The problems of logistics facilities locations and integrated logistics and plants facilities locations based on time restriction for multilayer supply chain system are studied. Two mixed-integer programming models with time restriction are proposed. These researches take into account integrating supply chain system resource, closing logistics facilities cost, transform the model to no time restriction by introducing parameters so that reducing the complicacy of solving models and examples give costs and each cost proportion in the total cost for different time restriction. Fourthly, with extended producer responsibility implementing, supply chain system not only want to produce, store, transport, allot products to customers, but also want to take back used products from customers. The problem for single plant taking-back is analyzed. According to consortium of manufactures takeback used products together; the disposal sites locations and flows distribution problem for used productions are researched. Two models which have multiple initial collection sites, multi-manufacturer, multi-used-product and multi-disposal-site are developed for only involving upper bound of disposing capacity as well as involving lowest and the utmost disposing capacity for each disposal facility at every disposal site. Due to the complexity of models, the heuristic algorithms based on Lagrangean relaxation for these problems are elaborated.The developed models and algorithms in the dissertation succeed and improve the existing researches about facilities location so that these results can be put into facilities locations practice in supply chain system. These works support optimization of facilities locations and relocations in supply chain system and can urge the equilibriums between customers’demand distributing and facilities sites, improve the answer for customers demand, decrease logistics cost and strengthen competition for total corporations in the supply chain system.更多还原