【作者】 陈华；

【导师】 张岐山；

【作者基本信息】 福州大学 ， 管理科学与工程， 2011， 硕士

【摘要】 供应链配送网络是供应链的重要组成部分。供应链配送网络优化问题是供应链管理在战略层面上的一个最基本的核心问题，是物流和供应链研究领域中的一个重要课题。近年来不确定性的供应链配送网络优化成为研究的一个热点。目前对不确定性供应链配送网络优化的研究，主要集中在随机性和模糊性方面，而面向灰性的供应链配送网络优化研究较少。在现实中，供应链配送网络中的不确定信息除了随机信息、模糊信息外，还含有灰信息，如灰色需求信息等。灰色需求不确定性客观存在，但在理论研究上并没有得到足够的重视。因此，灰色需求下供应链配送网络优化问题的研究有重要的理论价值和现实意义。在理论上，本文进一步补充和完善供应链配送网络优化的研究，在实践上，本文能够提高供应链配送网络构建过程中处理灰色需求信息的能力。首先，论文研究了供应链配送网络优化中的节点降维方法。供应链配送网络优化问题随着问题规模的扩大，维数灾难成为瓶颈。本文给出了基于灰关联分析的Topsis方法，该方法可缩减配送网络设施节点的数量，实现降维目的，大大减少模型求解算法的计算量。通过供应商选择的实例验证了该方法的有效性。其次，针对供应链中比较典型的多供应商、多工厂、多分销中心和多销售中心组成的四级配送网络结构，考虑了多原材料多产品的情况，重点研究灰色需求下的供应链配送网络优化问题，建立了供应链配送网络灰色优化模型。该优化模型主要是选择设施（工厂和分销中心）以及确定配送方案，以最小成本满足需求。根据模型的特点，为解决模型的优化求解问题，提出了先将灰色优化模型转化成能求解的灰色机会约束规划模型，并设计了基于混合微粒群算法的两种求解算法对其进行求解。（1）基于混合智能的灰色优化模型求解算法，该算法采用了双层求解机制，上层是基于模拟退火算法的微粒群算法求解供应链配送网络设施选址问题，下层将灰色模拟和微粒群算法相结合，解决配送网络的配送运输问题。（2）基于定位系数优化的灰色优化模型求解算法，该算法将微粒群算法引入到灰数白化的定位系数优化中，并以灰色模拟检验灰色机会约束，丰富了灰数白化的方法。两种求解算法有效地提高了求解能力和效率。通过实验算例说明了模型的有效性，并给出结果分析。最后，对本文的研究成果进行简要总结，并指出需要进一步研究的问题。更多还原

【Abstract】 A supply chain distribution network is an important component in a supplychain. In supply chain management, the optimization problems of distributionnetworks have always been a key strategic issue. It was considered as animportant research in the field of logistics and supply chain management. Inrecent years, the uncertain optimization of distribution networks in supply chainhas become a focused research. At present, domestic and foreign scholarshave done extensive and in-depth researches to the randomness and fuzziness,but have not yet worked on the greyness in the optimization problems ofdistribution networks in the supply chain. In actuality, there is plenty ofinformation about the uncertainty in the supply chain distribution network, whichincludes random information, fuzzy information, and also includes greyinformation, for instance grey customer demand. Grey demand uncertaintyexists objectively,but it doesn’t get enough attention in the theoreticalresearch.Thus, the study on the optimization problem of distribution networks inthe supply chain under grey demand will undoubtedly be important fortheoretical study and practical application.The research results are asupplement and perfection of supply chain distribution network optimization,and improve in practice the capacity of dealing with uncertain information ofgrey demand in the process of building a distribution network in supply chain.Firstly, this thesis researches dimensional reduction algorithm for nodes inoptimization of distribution networks in the supply chain.The dimension disastercaused by the exponential increase of dimensions of nodes,becomes abottleneck of supply chain distribution network optimization problem.Themethod of weighted Topsis based on grey correlation analysis isestablished.The number of distribution network facility nodes can be reduced toan acceptable level, thus greatly reducing the computational complexity ofsolution algorithm for distribution network optimization model.The effectivenessof this method is finally proved by example of supplier selection.Secondly, to focus on typical four-stage distribution network of supply chain with many suppliers, plants, distribution centers and sales centers, distributionnetwork optimization model considering many raw materials and products ispresented,which is based on the problem of supply chain distributionoptimization under grey demand. The tasks of this problem involve the choice ofthe facilities (plants and distribution centers) to be opened and the distributionnetwork designed to satisfy the demand with minimum cost. According to thespecialty of the grey optimization model, the model is transformed into the greychance-constrained model,and then two hybrid PSO intelligent algorithmsbased on grey chance-constrained programming are proposed for it. One issolution algorithm for distribution network grey optimization model based onhybrid intelligent algorithms，which adopted two-layer solution mechanism.Theupper layer mixed with simulation annealing algorithm in the method forimproving the algorithm’s performance to solve the choice of the facilities(plants and distribution centers) to be opened.The lower layer develops PSOand grey simulation to solve the distribution network design problem to satisfythe demand. The other algorithm is solution algorithm for distribution networkgrey optimization model based on optimization of the position coefficient. PSOis used in the process of optimizing the position coefficient in grey number’swhitening transformation,and grey simulation technology is used to process thecomplex opportunity restriction.So the methods of optimizing of grey number’swhitening transformation have been enriched.The computer simulation showsthat the model is feasible and the algorithm is effective. And the solutionanalysis is provided.Finally, this thesis summarizes results obtained, and then presents someproblems for further study.更多还原