不确定需求环境下汽车供应网络的牛鞭效应研究

The structure and behavior of supply networks are becoming increasingly complex as the economical globalization. Small fluctuations of customer demand will be amplified in a supply network, which brings the bullwhip effect problem and thus causes great economical losses for companies. During the recent years, all kinds of emergencies occur frequently and the global economy downturn goes on. Thus, the research on the bullwhip effect of supply network under uncertain demand is of great practical value for improving the capability of operations and management under dynamic environments and improving the flexibility of supply networks. By considering order collaboration, information sharing, lead times, and decentralized control, this project will firstly develop difference equation models by analyzing the structural and functional characteristics of automative supply networks. Effective inventory control policies are solved or generated using advanced robust control theory via designing the bullwhip metrics. The influences of network structure, lead time, and demand forms on the bullwhip effect will be systematically studied by conducting numerical simulations. Secondly, multi-agent simulation platforms for automative supply network will be developed based on the theories of complex adaptive systems and complex networks. Based on the simulation platforms, the impact of firm bankrupts, supply relationship changes, and the adjustments of inventory policies on the bullwhip effect of supply network will be investigated. This project will finally bring original research achievements on the modeling, simulation and analyzing for the dynamics of automative supply networks, which will certainly contribute to the fields of supply chain management and complex sciences.

经济全球化以及生产外包使得供应网络的结构与行为变得越来越复杂, 从而产生严重牛鞭效应问题，对企业造成不可估量的经济损失。近年来，随着全球经济的不断下滑以及突发事件的频繁发生，研究不确定需求环境下供应网络的牛鞭效应对于提高企业的动态运作管理能力、提高供应网络的灵活性具有重要的现实意义。本课题以汽车行业为背景，首先分析汽车供应网络的结构特性与功能特性，针对订单合作、信息共享、多提前期等不同的情况分别建立动力学模型，利用时滞系统理论、鲁棒控制理论及数值仿真实验揭示网络结构、提前期、顾客需求等因素对牛鞭效应的影响规律；然后，基于复杂适应系统理论以及复杂网络理论建立供应网络结构演化的多主体仿真实验平台，研究企业破产、供应关系改变以及库存策略调整对供应网络牛鞭效应的影响。本课题力图在汽车供应网络的动态建模、仿真与分析等方面取得原创性成果，对供应链管理理论和复杂性科学的研究有重大促进作用。

不确定需求环境给复杂供应链网络的运作与管理增添了困难。本项目以汽车供应链网络为背景，建立了汽车供应链备件需求预测模型，综合应用复杂网络、鲁棒控制、多主体仿真等理论与方法，分别研究了不确定需求环境下考虑多提前期、闭环供应链、订单合作、横向调货等问因素的供应链网络的牛鞭效应问题，创造性地提出了状态空间增维技术，解决了多时滞系统鲁棒控制器无法求解等问题。通过多主体数值仿真实验，发现了适度的订单合作与横向调货有利于抑制供应链网络的牛鞭效应，而网络结构是影响供应链牛鞭效应的一个非常重要的因素。本项目的研究也拓宽了应用领域，包括不同布局下的仓库运作与管理、绿色供应链的行为偏好与激励机制等问题。本项目的研究对于揭示供应链的复杂性并提出管理控制手段具有重要意义。