Multi-issue negotiation with deep reinforcement learning
作者：
Highlights：
• Deep reinforcement learning adapts to time-based and behavior-based opponents.
• Acceptance accuracy can be improved through analysis of marginal utility.
• The Cauchy distribution is suitable for offer strategies.
• Neural agents learn to cooperate against relative tit-for-tat and in self-play.
• Cooperation during self-play resembles ultimatum games in evolutionary game theory.
摘要
•Deep reinforcement learning adapts to time-based and behavior-based opponents.•Acceptance accuracy can be improved through analysis of marginal utility.•The Cauchy distribution is suitable for offer strategies.•Neural agents learn to cooperate against relative tit-for-tat and in self-play.•Cooperation during self-play resembles ultimatum games in evolutionary game theory.
论文关键词：Deep reinforcement learning,Negotiation,Game theory
论文评审过程：Received 10 July 2020, Revised 15 September 2020, Accepted 13 October 2020, Available online 22 October 2020, Version of Record 28 October 2020.
论文官网地址：https://doi.org/10.1016/j.knosys.2020.106544
