【作者】 何龙飞；

【导师】 赵道致；

【作者基本信息】 天津大学 ， 工业工程， 2010， 博士

【摘要】 随着传统高度纵向集成式产业链向新型多维扩散式产业链的演变,如何实现供应网络中分散决策环境下的链协调和全局优化,以提升供应链系统绩效和盈利能力,逐渐成为理论界和实务界共同关注的研究课题。尽管在供应链协调领域已有一系列研究成果,但对于若干特定链结构的供应链协调和优化问题迄今研究不足。本文重点讨论了多样化链结构下供应链协调机制的设计和网络优化,论文主要研究内容和创新性工作如下:(1)针对供应网络由多个供应商和少数零售商组成的力量不对等供应链结构,分别建模分析得到完全信息和不完全信息下的静态定价博弈和跟随定价博弈均衡,给出外生机制设计实现系统收益Pareto改进的条件,基于Rubinstein讨价还价博弈给出外力干预实现系统绩效Pareto改进的作用区域和约束。(2)以具有动态博弈过程的两级供应链结构为对象,研究给出完整的旁支付自执行契约的设计方法,使原契约演化为新的动态博弈;SSEC设计使得分散决策的各方在追求个体利益最优的行为却取得了集中决策下的集体理性效果,其满足激励相容、参与约束和自发性稳定性原则,丰富了供应链协调机制设计理论。(3)针对由风险中性供应商、风险厌恶分销商和风险中性零售商组成的三层供应链结构,建模分析在VMI模式下同时考虑风险约束和需求对价格敏感性等前提下的供应链协调契约设计,基于下侧风险度量给出实现系统协调的条件以及达到全局最优的必要条件。(4)基于随机最优控制建模分析,以多层供应链结构为对象,考虑创新型产品特点,分别研究在市场需求随机波动下的该类产品运输与库存的最优控制策略和多节点库存运输的供应链协调。以扩散控制模型和考虑常规和紧急两种运输方式分析求解出该产品的最优库存即时控制策略,最小化长期平均运营成本;分别用联盟博弈和模糊合作博弈证明在确定环境和模糊环境下的多层供应链联盟核和模糊核存在的可能性及预测相应的联盟和模糊联盟中的成本分担。(5)分析了具有特殊需求关联的供应链结构,根据网络内各环节产品的物流方向将网络内的产品需求分为网络内部成员企业之间需求(中间需求)和网络外部需求(最终需求),并用投入产出矩阵描述需求关系,建立网络外部需求随机波动情况下单周期与有限多周期供应网络的动态投入产出模型。运用定义于? n空间的n维K凸性给出该供应网络多产品有限多周期情形下的动态产量(库存)最优的联合优化策略。更多还原

【Abstract】 Accordingly with highly integrated industry chain evolving to be newly multi- dimensionally dispersed, it has gradually become one of the hottest research topics in both academia and industry that how to achieve supply chain cooordination(SCC) and obtain global optimization of the networks in a decentralized-decision situation for improving performance and enhancing profitability of the supply chain. Although there exists a lot of research done in the field of SCC, it’s obvious that the SCC and optimization problems on some specified supply chain structure are still open. In this thesis, we mainly study how to make SCC mechanism design and supply network optimizaiton for diverse specified supply chain structure with the research results and inovations as follows:(1) With a bargaining power unblanced supply chain structure consisting of many suppliers approximating competely competitive market and several monopolistic channel retailers, we model and analyse this structure in static and dynamic pricing games respectively to get the equilibria, upon which we obtain the fundamental conditions for mechanism design of the exogenrous force to get pareto improvement of the system payoff, and then the empact domain and constraints of this exogenous forces are interpreted exploiting Rubistein bargaining game.(2) Fousing on the general two-stage supply chain structure frequently described by dynamic game, we devise a complete mothedology of side-payment self-enforcing contract (SSEC) design, which enable the original game to evolve to a new dynamic game. SSEC design stimulates every participant in the decentralized-decision situation behaves in a way optimizing individual payoff, but surprisely achieving a collective-rational effect. It naturally meets principles of incetive compatibility, participation constraints and self-enforced stability, which enrich the supply chain coordination mechanism design theory.(3) On a triple-echelon supply chain structure comprising a risk-neutral supplier, a risk-averse distributor and a risk-neutral retailer, we refine and model this supply transaction process considering simultaneously the risk constraints and pricing dependent demand sensitivity, and design a new contract to get chain wide coordination preconditioning satisfactions of risk constraint and global optimization.(4) Considering a multi-echelon supply chain structure with innovative products, we exploit the tool of stochastic optimal control to model the problem of minimizing long run average system operation cost. Further we provide an optimal inventory and transportation control policy clearly defined by several parameters, which can optimize the total operational cost when the multi-echelon supply chain is suffering demand fluctuation and volatility. What’s more, we also utilize coalitional game and fuzzy coalitional game to analyze and prove the stability of coalition and fuzzy coalition in deterministic and fuzzy environment, respectively, depending on which we also conjecture the cost allocation in both coalition and fuzzy coalition.(5) For a supply chain structure possessing special demand correlations among different participants, we sort the product demand into independent demand for inter-network and intra-network, and intermediate demand which can be combined to an input-output matrix. We further model a single-horizon and a multiple period input-output relationships to grasp the dynamics of inventories in the network subject to stochastically independent demands. Finally, we give out a jointly optimal policy to monitor inventory levels systematically and locally by making use of the n dimensional K -convexity in finite horizons.更多还原