基于深度学习的智能交易Agent的交易行为研究

Study on Transaction Behavior of Artificial Intelligent Agents Based on Deep Learning

基于Agent的模型（ABM）在许多领域取得了显著研究成果,在Agent设计等方面的改进也层出不穷。由于在真实市场中Agent之间拥有不同的学习能力、不同的学习速度、不同的社交网络,因此不同的Agent设定使得模型结果也不一致。为了得到更一般的结论,文章将在深度学习的基础上融合行为金融学,设定各种类型的Agent模拟股票市场。实证结果表明文章中的模型能够很好地反映真实股票市场的运行情况,表明智能Agent交易行为的变化和股市动态具有较强的相关性。通过规范市场Agent相关行为可以起到规范市场的作用。

Significant research results about the agent-based model （ABM） have been achieved in many fields, and the im- provement on the Agents＇ design emerges in endlessly. Since different Agent in the real world has different learning abilities, dif- ferent learning speed and even different social networks, different Agent settings also have different results. In order to get a more general conclusion, this paper integrates behavioral finance on the basis of deep learning, and sets various types of Agents to simu- late the stock market. The empirical results show that the model reflects the performance of real stock market very well, and indi- cate the change of Agent trading behavior and the dynamics of stock market have strong correlation, and that regulating market Agent＇s related behaviors helps regulate the market.