Multi-Agent Simulation Based Research On IPO Initial Day Price Performance

Due to the limitation of traditional"Reduction-based"economic study methods, they can not explain many modern economic phenomenons, especially for the financial stock market, in which the participators'psychology factor affects a lot. With the development of system engineering, a new CAS (Complex Adaptive System) theory was advanced and agent-based financial markets are universally applied in economy study which is far from traditional notion of equilibrium. By studying the famous agent-based model ASM (Artificial Stock Market), an IPO model was set up, upon which some simulation study was then emplemented.By analyzing the current research condition in the agent-based economic research area in China and abroad, the author found the current weaknessan in the IPO research with this method, and presented the meaning of the study in this paper, then had an analysis of China's 2004 IPOs data, upon which found the relationship between some factors and the IPO day close-price, then an IPO simulation model was set up. Here the CAS theory is applied and referring to the UML method, the main charts of the system are drawn to direct the realization of the Model on computer. With the realized IPO simulation model, some simulations of the IPO market are done and some results are therefore gain.The research in this paper is a successful practice of the agent-base simulation method in IPO market study, which put forward the foundamental studies of former researchers into a specified finance study field. Meanwhile, the agent-base IPO simulation model set up in this paper proves valuable references for studying the phenomenons in the IPO market and also helpful for the government to make suitable policies to guide the IPO market's development.