Asset - Market Multi - Agent Model Based On Continuous Bidirectional Auction Mechanism

In this paper, we establish a pricing model under continuous double auction mechanism in limit order markets. And we conduct a dynamic analysis of this model with the help of MATLAB tool to complete the simulation work. After conducting simulations, we analysis and discuss the results in details, and compare them with the stylized facts of the real financial market as well. In our model agents are heterogeneous,we divide them into two main types: Mean reversion traders and Chartists. According to their relative performance, Mean Reversion traders mean-revert to the weighted average and Chartists choose the moving average rules.The main body of this paper consists of four parts: The first part, we give an outline of traditional economic and financial theories, as well as their defects. Meanwhile, we state new financial theory and method, and talk about the history and present situation of this study. We make some explanation on the limit order markets, the choice of traders, and continuous double auction mechanism. Also, we put forward a clear research subjects, and emphasize the research significance of this research subjects. The second chapter is the establishment of the model, the price forming mechanism in the model are introduced in detail, that is, continuous double auction mechanism, but not the traditional market-makers mechanism. We need to notice that the continuous double auction mechanism is a many-to-many trading mechanism. This part provides detail information on trading strategies and the whole trading process of Mean reversion traders, Chartists, and noise traders. In the third chapter in this paper, some important parameters are set in the model. On the basis of the establishment of the model, we use MATLAB for simulating models, get the model simulation results, and have a detailed analysis and discussion on the simulation results. In addition, we compare them with the results of the CHP model and the stylized facts of the real financial market as well,especially the volatility of the asset price and the statistical property of the rate of return.We can deduced from the model results that, the model reproduces most of the stylized facts of the real financial market, including volatility clustering, insignificant autocorrelations(ACs) of returns, leptokurtic and fat-tailed of returns distribution, and significant slowly decaying ACs of the absolute returns. At the same time, it seems that the results suggest us some important message, that is to say, we could analysis the returns properly and have possibility to get a comprehension explanation of its severalproperties in the special framework, where both realistic microstructure and behavioral traits have a role. In the final, we consider establishing pricing model of stock market with the existence of credit transactions, in order that we could study the influence of credit transactions behavior of market price.