中国股市仿真系统建模及其非线性特征研究  

Modeling Agent-Based Artificial Chinese Stock Market and Nonlinear Characteristic Analysis

作　　者：邹琳[1] 马超群[1] 李红权[2] 

机构地区：[1]湖南大学工商管理学院,长沙410082 [2]湖南师范大学商学院,长沙410081

出　　处：《系统管理学报》2008年第4期385-389,共5页Journal of Systems & Management

基　　金：国家自然科学基金资助项目(70471030)

摘　　要：作为新兴的股票市场,中国股票市场还不成熟和完善,用传统方法很难建模。在交易者模型中引入了噪声交易者模型,并加入政策因子,应用遗传算法来进化预测规则,建立了Agent的价格预期模型。同时根据中国股利收益率偏低的特点,建立了中国的股利动力学模型。在此基础上建立了中国股票市场仿真系统。对该系统进行模拟及分析,发现仿真系统与真实市场同样主要的非线性动力学特征——分形和混沌动力学特征。这一研究对挖掘中国股票市场的演化规律和混沌动力学产生的关键因素具有重要意义。As an emerging booming stock market, Chinese stock market has not mature nor perfect. It is difficult to model it by using traditional methods. In this paper, we divide traders into rational traders and noise traders and introduce a model of noise traders. Joining policy factor and evolving forecasting rules with GA, we modeling the formation of Agent＇s price expectations based on character of Chinese stock market. At same time, we build the dividend dynamic model with the character that the dividend distribution ratios are very low for Chinese stocks. Based on these models＇ we simulate Chinese stock market. Then we compare the characteristic of real stock market and of artificial stock market＇ and find real stock market and artificial stock market are of same primary nonlinear characteristic-fractal and chaos. It is significant to study evolving rules of Chinese stock market and order parameters that lead stock market come into being chaos.

关 键 词：中国股市 仿真系统 非线性动力学 遗传算法 

分 类 号：F830.9[经济管理—金融学]