基于多Agent的股民行为仿真研究-管理科学与工程专业论文.docxVIP

基于多 基于多 Agent 的股民行为仿真研究 MULTI-AGENT SIMULATION STUDY OF INVESTORS BEHAVIOR ABSTRACT The stock market is a typical complex adaptive system with the characteristics of open, complexity, non-linear. It is difficult to describe and analyze them with traditional methods based linear, static, and balanced structure. However,The modeling method based multi-Agent break through the single-level, single view of the limitations, which is a bottom-up approach, for the description and study of complex adaptive systems such as the stock market provides a more realistic, more effective modeling framework.It will greatly enhance people of understand,research, manage and control the complex adaptive system of capabilities.Therefore, this paper introduces the modeling method based Multi-Agent from a new angle to examine the behavior of investors in the stock market. Firstly, this paper learns from predecessors through their literatures, according to the theory of complex adaptive system and Multi-Agent carried out the complexity of the stock market analysis, and for the Chinese stock market investors behavior analysis. The relevant data conclude that Chinas stock market transactions clearly demonstrate the characteristics of individual investors, most of the Chinese investors have herd and other psychological bias. Previous studies showed that these factors will lead to the formation of investors herd behavior, so, this third chapter examines the existence of our investors present herd behavior by CCK test, for the stock market as the three stages of a whole,rise and fall of the Chinese investors behavior, respectively conduct an empirical study on herd. The empirical results show that Chinas stock market both in the rising stage and on the decline stage there is a obvious herd behavior. Finally,for the complexity of the investors behavior, proposed a specific method of based Multi-Agent Modeling and Simulation, in StarLogo modeling and simulation of the MAS platform, established th