基于元胞自动机的股票市场投资策略演化研究

发布时间：2017-12-31 15:38

本文关键词：基于元胞自动机的股票市场投资策略演化研究　出处：《广东工业大学》2012年硕士论文　论文类型：学位论文

【摘要】：传统金融学的研究方法侧重于总体把握,而容易忽视个体行为,特别是缺乏个体与环境之间的互动。作为典型的复杂自适应系统,金融市场具有内部结构层次多、参与要素多且性质复杂、变量关系多、内部因果关系多等特点,其时间上或空间上的分离性(滞后性)、强耦合性等特性决定了经典金融理论的线性视点与均衡分析不能充分揭示市场规律。在此基础上,人工股票市场中的投资者的心理情感范式设计以及智慧程度刻画日益成为研究的热点问题,作为探索复杂性与复杂系统的有力工具,元胞自动机模型正被广泛应用与股票市场研究。本文以投资者的策略演化行为作为研究对象,结合传统资产定价模型框架,建立了新的基于元胞自动机的投资策略演化模型,对于投资者心理情感范式的设计及其演化方式、智慧程度,本文给出了创新性的改进。首先,本文对目前基于复杂自适应系统的各类股票价格模型进行了回顾,对其分析方法及分析思路,尤其是演化机制的设计及智慧程度刻画进行了介绍,提出当前研究的缺陷,进而说明了本文的选题依据。其次,基于经典的资产定价模型框架及元胞自动机理论,(1)设计了“单一策略”及“混合策略”两种投资者策略演化方式；(2)利用非线性动力学的相关理论对模型进行了稳定性、分叉分析；(3)分别将“策略演化效用函数”及“元胞遗传算法”引入两种基于元胞自动机的策略演化方式中,提高了投资者的智能化程度。第三,利用matlab对上述设计及分析进行计算机仿真模拟,对实验结果进行横向及纵向比较,得到“混合策略”及“单一策略”两种演化方式对价格行为、利润、策略选择人数比例等市场特征的影响。最后,在实证部分,结合真实股票市场中的实际状况及市场上常见的大众投资者心理情感范式,对模型的适用性做进一步阐述。在完成了上述四项工作之后,提出本文结论：(1)投资者的“从众”行为未必一定会导致市场价格波动异常；(2)技术面投资策略易造成价格波动,但投资者在不同策略之间的转换过程也是造成股市震荡的原因；(3)引入策略演化效用函数或元胞遗传算法后,既在一定程度上保证了利润,也同时维持了市场稳定；(4)散户之间刻意的学习行为并不一定会获得超额利润,而操纵股价的机构正是成功的掌握了“大众心理”,从而获得超额利润。
[Abstract]:As a typical complex adaptive system, traditional research methods of finance focus on overall grasp, but tend to ignore individual behavior, especially the lack of interaction between individual and environment. The financial market has many characteristics such as multiple levels of internal structure, more participating elements and complex nature, more variables, more internal causality and so on, and its separation in time or space (lag). The characteristics of strong coupling determine that the linear view and equilibrium analysis of classical financial theory can not fully reveal the market law. The design of investor's psychological and emotional paradigm and the characterization of the degree of intelligence in artificial stock market have increasingly become a hot issue in the research, as a powerful tool to explore the complexity and complexity of the system. Cellular automata model is widely used in stock market research. Based on the traditional asset pricing model, a new evolutionary model of investment strategy based on cellular automata is proposed in this paper. For the design and evolution of investor's psychological emotion paradigm and the degree of intelligence, this paper gives some innovative improvements. Firstly, this paper reviews all kinds of stock price models based on complex adaptive system, and introduces the analysis methods and ideas, especially the design of evolution mechanism and the description of intelligence. Put forward the defects of the current research, and then explain the basis of this topic. Secondly, based on the classical asset pricing model framework and cellular automata theory, this paper designs two kinds of "single strategy" and "mixed strategy". 2) the stability and bifurcation of the model are analyzed by using the theory of nonlinear dynamics. 3) the strategy evolution utility function and the cellular genetic algorithm are introduced into two kinds of strategy evolution methods based on cellular automata, which improves the intelligence of investors. Thirdly, matlab is used to simulate the above design and analysis, and the experimental results are compared horizontally and longitudinally. The effects of "mixed strategy" and "single strategy" on market characteristics such as price behavior, profit and the proportion of people selected by strategy are obtained. Finally, in the empirical part, combined with the actual situation in the real stock market and the popular investor psychological emotion paradigm, the applicability of the model is further elaborated. After completing the above four tasks, the conclusion of this paper is that "herd" behavior of investors may not necessarily lead to abnormal market price fluctuations; (2) Technical investment strategy is easy to cause price fluctuation, but the conversion process between different strategies is also the cause of stock market volatility; 3) after introducing the strategy evolution utility function or cellular genetic algorithm, the profit is guaranteed to a certain extent, and the market stability is maintained at the same time. (4) deliberate learning behavior among retail investors does not necessarily lead to excess profit, and the institution that manipulates the stock price successfully grasps the "popular psychology" and thus obtains the excess profit.
【学位授予单位】：广东工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.51;F224

【相似文献】