“CA结构的智能Agent”方法的构建与应用  

The Design and Application of "Agent with the CA's Structure" in Stock Market

作　　者：刘春霞[1] 孙绍荣[2] 

机构地区：[1]上海电视大学管理系,上海200433 [2]上海理工大学管理学院,上海200093

出　　处：《数学的实践与认识》2012年第16期29-36,共8页Mathematics in Practice and Theory

基　　金：国家自然科学基金(70871080;71171134);上海市哲学社会科学规划项目(2011BGL006);上海市教委科研创新重点项目(11ZS138)

摘　　要：智能体和元胞自动机是基于复杂适应系统理论(CAS)框架下应用较成熟的两种行为模拟方法,在对它们的优缺点及适用领域进行比较分析的基础上,构建了一种更适合人类复杂行为模拟的新方法——"CA结构的智能Agent".并将其应用于股市投资者行为的模拟研究中,模拟结果验证了股市投资行为中的羊群行为和复杂系统特性.此外,模拟还发现股市达到均衡的时间和状态取决于投资者的类型和比例.Agent and Cellular Automata are two kinds of simulation methods which are used broadly and maturely under the framework of CAS. After analyzing the advantages and disadvantages of each simulation method, we put forward an improved model- ＂Agent with the CA＇s Structure＂, which is more suitable to simulate the human beings＇ behavior. Simultaneously we apply it in simulating the investors＇ behavior in stock market. The output convinces the herd behavior phenomenon in stock market and its system complexes, it also tell us that the number of different kinds of investors can affects the equilibrium of stock market and its time to achieve it.

关 键 词：行为模拟 智能体 元胞自动机 CA结构的智能Agent 

分 类 号：F830.59[经济管理—金融学] F224