面向供应链的复杂网络模型分析与控制研究

发布时间：2018-03-24 03:02

本文选题：供应链　切入点：复杂网络　出处：《浙江工业大学》2014年博士论文

【摘要】：全球经济一体化趋势的不断加强,政治、经济、社会环境的巨大变化,顾客消费水平的不断提高,使得企业间的竞争日益加剧。企业为了提高竞争力而采取了许多先进的技术和管理方法。企业在全球市场中不再是作为一个单个的实体,而是作为参与供应链竞争的实体,同时企业间的竞争转化成了供应链间的竞争。在合作、竞争、动态的环境中,供应链是由顾客、零售商、分销商、供应商等实体构成的动态供需网络,具有快速响应环境变化的特征。作为一个典型的复杂网络系统,供应链上的企业彼此之间相互作用,同时又受到其他供应链实体和市场环境的影响,企业间关系错综复杂。目前供应链的研究大多是研究其具体运作(微观行为)问题,缺乏对系统整体宏观演化规律和行为的研究。从系统的角度,基于复杂网络理论,探究复杂供应链网络的生长演化规律、整体宏观行为,对于供应链管理具有重大的实践价值和理论意义。本文基于复杂网络理论,对供应链网络模型、抗毁性和牵制控制问题进行了分析和研究,主要包括以下内容： 1、提出了分层供应链的演化网络模型,为本文随后章节的相关讨论提供了网络拓扑结构模型。该模型通过概率选择和节点度和边权多参数组合的竞争机制实现网络演化,并执行数值仿真加以验证,结果表明该模型具有小世界效应和无标度的特性,并且其幂律分布的指数落在(2,3)范围内,与绝大多数真实网络的幂律分布相吻合。 2、考虑节点失效性能的传递性,提出传递攻击策略。并通过数值仿真分析验证了供应链网络在受到不同干扰(随机攻击、选择攻击、传递攻击)的情形下,网络连通性、网络稳定鲁棒性和性能鲁棒性的表现。 3、提出融合全局信息和局部信息的失效节点负载分配策略,建立供应链网络级联失效模型。通过仿真实验分析了初始节点攻击方式、不同失效节点负载分配策略、以及不同节点过载承受能力等因素对供应链网络级联失效行为的影响。 4、提出了异质节点的供应链网络牵制控制模型,研究了网络中不同类型节点稳定到相同状态的控制问题。针对异质节点设计不同的局部反馈控制器,利用Lyapunov函数和线性矩阵不等式,得出该网络指数稳定条件。并通过数值仿真实验验证了该模型的有效性。
[Abstract]:With the increasing trend of global economic integration, the great changes in the political, economic and social environment, and the continuous improvement of the level of customer consumption, Enterprises adopt many advanced technologies and management methods in order to improve their competitiveness. Enterprises in the global market are no longer a single entity, but as an entity to participate in supply chain competition. In a cooperative, competitive and dynamic environment, the supply chain is a dynamic supply and demand network composed of customers, retailers, distributors, suppliers and other entities. It is characterized by rapid response to environmental changes. As a typical complex network system, enterprises in the supply chain interact with each other, and at the same time, they are influenced by other supply chain entities and market environment. The relationship between enterprises is complicated. At present, most of the researches on supply chain are focused on its specific operation (micro behavior), but lack of research on the macroscopical evolution law and behavior of the whole system. From the system point of view, based on the complex network theory, It is of great practical value and theoretical significance for supply chain management to explore the law of growth and evolution of complex supply chain network and the overall macro-behavior. Based on the complex network theory, this paper analyzes and studies the supply chain network model, survivability and containment control, including the following contents:. 1. An evolutionary network model of hierarchical supply chain is proposed, which provides a network topology model for the related discussion in the following chapters. The model realizes network evolution by means of probability selection and competition mechanism of node degree and edge weight multi-parameter combination. The numerical simulation results show that the model has small world effect and scale-free property, and the exponent of the power law distribution falls within the range of 2 ~ 3), which coincides with the power law distribution of most real networks. 2. Considering the transitivity of node failure performance, a transmission attack strategy is proposed, and the network connectivity is verified by numerical simulation under different interference (random attack, selective attack, delivery attack). Network stability robustness and performance robustness. 3. A failure node load allocation strategy combining global and local information is proposed, and a cascade failure model of supply chain network is established. The initial node attack mode and different failure node load allocation strategies are analyzed by simulation experiments. And the influence of different node overload capacity on the cascade failure behavior of supply chain network. 4. The supply chain control model of heterogeneous nodes is proposed, and the control problem of different types of nodes to the same state in the network is studied. Different local feedback controllers are designed for heterogeneous nodes. By using Lyapunov function and linear matrix inequality, the exponential stability condition of the network is obtained, and the validity of the model is verified by numerical simulation.
【学位授予单位】：浙江工业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：O157.5

【相似文献】