零售网络的结构建模与演化分析  

The Structure of the Retail Networks:Simulation Modeling and Evolution Analysis

作　　者：刘小峰[1] 陈国华[1] 李真[1] 

机构地区：[1]南京大学工程管理学院,南京210093

出　　处：《管理科学》2009年第4期23-30,共8页Journal of Management Science

基　　金：国家自然科学基金(70731002;70401013);"十一五"国家科技支撑计划(2006BAH02A06)~~

摘　　要：为探寻零售网络的演化规律,基于复杂自适应系统理论构建零售网络结构模型。通过对agent、环境以及交互规则的描述和分析,描述了零售网络结构;利用Multi-agent仿真方法模拟零售网络在平稳需求、递增需求、近似实际数据需求3种需求情况下的动态演化过程,分析零售网络零售商数目、平均适应度、均价、死亡率与新增率的演化规律。模拟结果表明,零售网络系统沿演化路径经过若干个周期循环之后达到一种稳定的状态,在近似实际数据需求和递增需求驱动下,系统表现出较大的波动性;零售商的平均适应度与产品成交均价不存在相关性,用演化的观点阐述长期价格战的不合理性;另外,零售商的死亡率与新增率的演化特征解释了现实零售网络在市场成熟时期"有退难进"的竞争特征。To explore the laws of evolution of retail networks, the retail network model is build based on the theory of complex adaptive system （CAS）. The research framework of the theory of evolution has been developed by the agent, the environment and the description of the rules of interaction. The Multi-agent simulation is an effective method for investigating the growth of retail network in a specific environment and the interaction under some rules. It demonstrates the dynamics of the retail networks in steady demand, periodic pulse demand and demand of approximated from real demand data. Some evolution characteristics of variables are analyzed, such as the number of retailers, average fitness of retailers, average price, mortality and birth rate of the network. Experimental results show that the retail networks system achieve a stable state along the evolutionary path through several cycles, however, it shows greater volatility in the periodic pulse demand and demand of approximated from real demand data. There is no correlation between average fitness of retailers and average price. It reflects the long-term price war is irrational from the perspective of evolution. In addition, the features of the mortality and birth rate of retailers explain the special phenomenon of the real retail network in a mature competitive market.

关 键 词：零售网络 Muhi-agent 演化 复杂自适应系统 

分 类 号：F252[经济管理—国民经济]