基于不确定优化方法的供应链企业间协同决策研究

发布时间：2018-04-26 03:03

本文选题：供应链库存协同 + 不确定需求　；参考：《北京邮电大学》2014年硕士论文

【摘要】：近年来,随着信息经济的繁荣发展,客户需求日渐全球化,供应链中各企业已无法再单纯依靠自身能力来应对经济全球化的挑战了。市场不确定性的日益增加要求供应链中各企业必须实现信息共享,共同建立完善的协同机制,以增强供应链的核心竞争力,才能在激烈的市场竞争中占据优势地位。供应链协同管理涉及面众多,而库存协同是供应链协同管理的基础,合理有效的库存协同机制能帮助企业在有效满足用户需求的基础上,减少库存成本,最大化供应链收益。库存决策受用户市场的需求驱动,需求的不确定性使得库存协同机制的研究极为困难,已有研究大都采用平均需求或者假定需求服从某一特定的概率分布类型的方法来简化需求的影响,使得研究成果较为理想化,缺乏普遍性和实际应用性。因此,为解决需求的不确定性问题,本论文将采用蒙特卡洛仿真技术来对市场需求的不确定进行统计建模,构建通用的库存协同决策模型。然而,蒙特卡洛仿真技术是基于概率统计的方法,在执行过程中需要付出巨大的计算代价,因此,它在带来模型通用性的同时也带来了资源消耗过大的新问题。为此,本文将着重探索如何在保证蒙特卡洛仿真模拟结果准确度的基础上尽力减少计算资源的消耗,缓解结果准确度和计算消耗量之间的矛盾。在不确定需求的供应链库存协同策略优化问题中,本论文将首先采用蒙特卡洛仿真模拟技术建立了库存协同策略优化的基本模型,其次采用粒子群算法进行库存协同最优策略的搜索,在粒子群算法的适应度评价过程中同时引入了适应度遗传概念和自适应采样技术,分别从粒子进行适应度评价的次数和粒子进行适应度评价过程中的采样次数两方面入手,大力减少了蒙特卡洛模拟计算的资源消耗,从而达到了结果准确度和资源消耗程度之间的平衡。此外,本文还将对粒子群算法的适应度遗传技术及其供应链协同策略的自适应采样算法进行多方位探讨,比较多种影响因子的作用,有效提高供应链库存协同的效率,达到决策通用性,准确性和高效性的统一。
[Abstract]:In recent years, with the prosperous development of information economy and the increasing globalization of customer demand, the enterprises in supply chain can no longer rely solely on their own ability to meet the challenge of economic globalization. With the increasing uncertainty of the market, the enterprises in the supply chain must share information and establish a perfect coordination mechanism to enhance the core competitiveness of the supply chain, so as to occupy the dominant position in the fierce market competition. Supply chain collaborative management involves many aspects, and inventory coordination is the basis of supply chain collaborative management. Reasonable and effective inventory coordination mechanism can help enterprises to reduce inventory costs and maximize supply chain benefits on the basis of effectively meeting the needs of users. Inventory decision is driven by the demand of the user market, and the uncertainty of the demand makes the research of inventory coordination mechanism very difficult. Most of the previous studies have used the method of average demand or assuming demand service from a particular probability distribution type to simplify the influence of demand, which makes the research results more idealized, lacking of universality and practical application. Therefore, in order to solve the uncertainty of demand, this paper uses Monte Carlo simulation technology to model the uncertainty of market demand and construct a general inventory collaborative decision-making model. However, Monte Carlo simulation technology is based on probability and statistics, and it has to pay a huge computational cost in the process of execution. Therefore, it brings not only the generality of the model, but also the new problem of excessive resource consumption. Therefore, this paper will focus on how to reduce the consumption of computing resources on the basis of ensuring the accuracy of Monte Carlo simulation results, and to alleviate the contradiction between the accuracy of the results and the computational consumption. In the inventory coordination strategy optimization problem of supply chain with uncertain demand, this paper first uses Monte Carlo simulation technology to establish the basic model of inventory coordination policy optimization. Secondly, Particle Swarm Optimization (PSO) algorithm is used to search inventory cooperative optimal strategy, and the fitness genetic concept and adaptive sampling technique are introduced in the process of PSO fitness evaluation. Starting from two aspects: the number of particle fitness evaluation and the sampling number in the process of particle fitness evaluation, the resource consumption of Monte Carlo simulation calculation is greatly reduced. Thus, the balance between the accuracy of the results and the degree of resource consumption is achieved. In addition, in this paper, the fitness genetic technology of PSO and the adaptive sampling algorithm of supply chain coordination strategy are discussed in order to improve the efficiency of inventory coordination in supply chain. Achieve the unity of decision generality, accuracy and efficiency.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F274;TP18

【参考文献】