基于Agent的人工股票市场的模拟与分析

发布时间：2018-06-22 06:52

本文选题：人工股票市场 + 前景理论　；参考：《郑州大学》2014年硕士论文

【摘要】：20世纪七十年代计算机科学技术得到较快发展,尤其是90年代以后,随着计算机应用技术的日趋成熟,国内外金融学领域的专家们开始了计算机与金融结合的探索。试图通过计算机技术来模拟现实中金融市场,并通过研究金融市场微观结构——投资者的行为模式来揭示宏观的金融市场规律,即基于Agent的计算金融学(Agent-based Computational Financial)的研究由此进入了人们的视野。其中最具有代表性的就是美国圣塔菲研究院的基于Agent的人工股票市场(Artificial Stock Market)。随着研究的不断深入,学者们发现,合理地、有效地对Agent进行分类和搭建,尤其是能够准确地对Agent行为进行描述是整个研究中最为关键的一部分,也是计算机技术应用的关键。此外,行为金融学作为一门依托于人工金融市场领域的学科,也一直得不到重用。在当前人工股票市场的研究中,关于心理因素对投资决策的影响的研究并不突出,因此可以说,利用行为金融学的前景理论基础,将各类投资者的心理因素对投资决策过程的影响引入到人工股票市场,更加真实地模拟现实的股票市场将是一次大胆地尝试和创新。作者在前人的研究基础之上,采用自底向上的计算金融学建模思想构建了一个简易的人工股票市场,通过对底层的分析推进整个模型的分析和模拟。首先根据各投资者的投资行为特征及其对股票价格的影响作用,将其划分为机构投资者和个人投资者两大类,在对机构投资者建模的过程中,根据人工股票市场的运行特性特别选用改进的遗传算法对机构投资者的学习过程进行创新。此外在对个人投资者进行分析时,还特别引入了前景理论思想,对个人投资者进行建模,最后对仿真实验结果进行分析,在强调投资者组成结构对股票市场的影响的同时,并动态地再现了银行利率调整对股票市场的影响及涨跌幅限制制度对股票市场的影响。实验结果分析表明：实验数据从总体特性上来说与现实的股票市场的分析结果是相符的。虽然目前无法完全模拟现实股票市场的复杂性,但是对于一些制度制定和政策的有效性分析也是非常适用的。同时计算金融学这种全新的金融思维模式给金融领域的研究带来了一种全新的方式方法。
[Abstract]:Computer science and technology developed rapidly in the seventies of the 20th century, especially after 1990s, with the maturation of computer application technology, the experts in the field of finance at home and abroad began to explore the combination of computer and finance. This paper attempts to simulate the real financial market by computer technology, and to reveal the macro financial market law by studying the microstructure of the financial market-the behavior pattern of the investors. That is, the research of Agent-based Computational Financial based on Agent has come into people's view. The most representative one is the Artificial Stock Market based on Agent. With the deepening of the research, scholars have found that the reasonable and effective classification and construction of Agent, especially the accurate description of Agent behavior is the most critical part of the whole study, and also the key to the application of computer technology. In addition, behavioral finance, as a discipline relying on artificial financial markets, has not been reused. In the current research of artificial stock market, the research on the influence of psychological factors on investment decision is not prominent, so it can be said that the perspective of behavioral finance theory foundation, It will be a bold attempt and innovation to introduce the psychological factors of all kinds of investors into the artificial stock market and to simulate the real stock market more truthfully. On the basis of previous studies, the author constructs a simple artificial stock market by adopting bottom-up computational finance modeling idea, and promotes the analysis and simulation of the whole model through the analysis of the bottom layer. Firstly, according to the characteristics of investors' investment behavior and their influence on stock prices, they are divided into two categories: institutional investors and individual investors, and in the process of modeling institutional investors, According to the operating characteristics of artificial stock market, the improved genetic algorithm is used to innovate the learning process of institutional investors. In addition, in the analysis of individual investors, the perspective theory is especially introduced to model the individual investors. Finally, the simulation results are analyzed to emphasize the influence of investor composition structure on the stock market. It also dynamically reproduces the influence of bank interest rate adjustment on stock market and the effect of limit system on stock market. The experimental results show that the experimental data are consistent with the real stock market analysis results in terms of overall characteristics. Although it is not possible to fully simulate the complexity of the real stock market at present, it is also very suitable for the analysis of the effectiveness of some systems and policies. At the same time, computational finance, a new mode of financial thinking, has brought a new way to the study of finance.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F830.91

【相似文献】