基于做市商交易机制的人工股票市场建模仿真研究

发布时间：2018-03-05 01:23

本文选题：Agent　切入点：人工股票市场　出处：《华中科技大学》2012年硕士论文　论文类型：学位论文

【摘要】：做市商交易机制在西方国家的金融市场运用的十分广泛，，几乎所有的股票市场都是采用做市商交易机制。近年来，我国学者也开始关注做市商交易机制，准备把其引入我国股票市场。随着实验金融学和行为金融学的出现和计算机技术的发展，新兴的基于Agent建模的计算机实验金融学越来越多的被运用到金融学领域。本文的研究进一步推广应用了计算机实验金融学理论，同时对金融市场微观结构理论的发展提供良好的实验平台，对社会经济和科技发展具有较好的理论与应用价值。本项研究的主要思路和研究方法可以推广到针对其它金融模式下的仿真实验研究方面，同时对一般金融研究也有一定的借鉴意义。本文首先从金融市场微观结构基础知识出发，考虑了买卖价差形成原因和市场中的不对称信息的状态和做市商的存货风险等，结合做市商的存货模型和信息模型构建了一个新的做市商模型。然后利用基于Agent建模的实验金融学方法模拟了市场中的各类Agent，其中做市商Agent具有自适应学习能力，会根据市场的状态调整自身的行为以达到多目标任务，可模拟出实际市场的动态变化过程。进一步通过计算机仿真反复多次实验研究做市商交易机制下的人工股票市场，通过调整市场环境参数和投资者学习模型中参数等来研究做市商人工金融市场复杂系统。分析了对做市商的存货成本和交易量等量之间的关系，证明了做市商agent心理始终拥有一个最佳存货量。最后，文章进一步通过对做市商的买卖价差建模，分析了买卖价差构成各部分的比例，又通过计算机仿真实验研究买卖价差成分与交易量之间的关系，和做市商的存货量和价差之间的关系。所有的实验都是为了有针对性地调整做市商的行为，保证市场上交易量和价差达到最佳状态的目的。
[Abstract]:The market maker trading mechanism is widely used in the financial markets of western countries. Almost all stock markets adopt the market maker trading mechanism. In recent years, Chinese scholars have also begun to pay close attention to the market maker trading mechanism. With the emergence of experimental finance and behavioral finance and the development of computer technology, new computer experimental finance based on Agent modeling has been applied to finance more and more. The research in this paper further popularizes and applies the theory of computer experimental finance, and at the same time provides a good experimental platform for the development of financial market microstructure theory. It has good theoretical and practical value for the development of social economy and science and technology. The main ideas and research methods of this study can be extended to the simulation experiments under other financial models. At the same time, the general financial research also has certain reference significance. Based on the basic knowledge of the microstructure of financial market, this paper first considers the reasons of the spread of buying and selling price, the state of asymmetric information in the market and the inventory risk of the market maker, etc. A new market maker model is constructed by combining inventory model and information model of market makers. Then the experimental financial method based on Agent modeling is used to simulate all kinds of agents in the market. The market-maker Agent has adaptive learning ability and adjusts its behavior to achieve multi-objective tasks according to the market state. Can simulate the dynamic change process of the actual market. Furthermore, the artificial stock market under the market maker trading mechanism is studied repeatedly by computer simulation, and the complex system of market maker artificial financial market is studied by adjusting the market environment parameters and the parameters in the investor learning model. The relationship between inventory cost and trading volume of market makers is analyzed. It is proved that the agent psychology of market makers always has an optimal inventory quantity. Finally, the paper further analyzes the proportion of each part of the price difference by modeling the market maker's price difference, and studies the relationship between the component of the price difference and the trading volume through computer simulation experiment. All experiments are aimed at adjusting the behavior of market makers in order to ensure that the trading volume and the spread in the market reach the best condition.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F830.91;F224

【参考文献】