Trading financial indices with reinforcement learning agents

作者：

Highlights：

• Proposed reinforcement learning (RL) agents for individual retirement portfolio.

• Tested the framework using real world datasets.

• Results indicate that a dynamic RL agent portfolio performs the best.

摘要

•Proposed reinforcement learning (RL) agents for individual retirement portfolio.•Tested the framework using real world datasets.•Results indicate that a dynamic RL agent portfolio performs the best.

论文关键词：Reinforcement learning,Multi-agent systems,Markov decision process,Portfolio management

论文评审过程：Received 16 July 2017, Revised 22 February 2018, Accepted 23 February 2018, Available online 6 March 2018, Version of Record 10 March 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.02.032