金工论坛第八讲
报告题目(Title): Reinforcement learning in continuous double auction markets: an experimental study
时间:2018年12月25日周二,12:20-13:20
地点:对外经济贸易大学博学楼925
主讲人(Speaker):张军欢
个人介绍Biography:
张军欢,北京航空航天大学金融系副教授,伦敦国王学院计算金融学博士,美国佐治亚大学人工智能博士后。研究领域主要为人工智能和区块链在金融和经济领域的应用,尤其是计算与行为金融,区块链金融和人工智能。研究成果发表于Review of Quantitative Finance and Accounting等国际杂志。
摘要(Abstract):
This paper investigates the influence of reinforcement learning and individual rationality in continuous double auction (CDA) markets. Two reinforcement learning methods with epsilon-greedy and SoftMax strategies are used to describe traders’ decision-making processes by learning from the experiences of themselves. Also, a random learning strategy is considered as a benchmark to compare with other reinforcement learning methods. The performance of CDA markets includes market price, number of trades, buyers’ profit, sellers’ profit, and market efficiency.