主体异质性、复杂网络与网络效应下的标准竞争——基于计算经济学的研究  

Agent Heterogenous,Complex Network and Standard Competition with Network Effect——Research Based upon Computational Economics

作　　者：鲜于波[1] 梅琳[1] 

机构地区：[1]中山大学岭南学院,广州510275

出　　处：《系统管理学报》2008年第2期225-234,共10页Journal of Systems & Management

基　　金：国家自然科学基金资助项目(7057312570673118);中山大学"985工程"产业与区域发展研究哲学社会科学创新基地资助项目

摘　　要：分析了Agent的网络效应异质性和预期异质性对网络效应下标准竞争和扩散的影响。采用复杂网络"小世界效应"模型和基于Agent的计算机建模技术,并结合行为博弈学习理论EWA算法,由此对复杂网络下标准竞争过程中的主体异质性对标准扩散的影响进行计算机建模研究。基于异质性个体在局部网络中的交互作用和Agent行为的学习演化与预期调整,通过计算机仿真,发现网络效应异质性对标准扩散有正面促进作用。本文研究也发现预期的强度和预期调整因子的取值大小对标准竞争扩散的速度和结果有非常明显的影响。在一定情况下,异质性的预期能够克服先发优势;Agent的异质性预期也能帮助实现新旧技术的转换。This paper analyses the role of agent heterogeneous（ including network effect heterogeneous and expectation heterogeneous） in standard diffusion with the network effect from the point of agent-based computational modeling. At the same time, this paper uses the theory of complex network. It also includes the result achieved in behavior game theory especially the EWA algorithm proposed by Camerer. Based on the local interaction of agents and the new agent learning algorithms, this paper analyses the agent heterogeneous and its influence in standard diffusion process. The simulation result shows agent heterogeneous will be helpful in standard diffusion process. This paper also found that the strength of expectation and adjustment speed of expectation will have obvious effect on standard diffusion process. If the expectation is high and the network effect is strong, agent heterogeneous will help the new standard replace the old one.

关 键 词：网络效应 标准竞争 复杂网络 计算经济学 Agent异质性 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置] N94[自动化与计算机技术—控制科学与工程]