A reinforcement learning optimized negotiation method based on mediator agent

作者：

Highlights：

• Proposes a bilateral multi-issue optimized negotiation model based on reinforcement learning.

• Introduces a mediator agent as the mediation mechanism.

• Uses the improved reinforcement learning negotiation strategy to produce the optimal proposal.

• Introduces a benchmark concession utility function to optimize the ability of mediation of the mediator agent.

摘要

•Proposes a bilateral multi-issue optimized negotiation model based on reinforcement learning.•Introduces a mediator agent as the mediation mechanism.•Uses the improved reinforcement learning negotiation strategy to produce the optimal proposal.•Introduces a benchmark concession utility function to optimize the ability of mediation of the mediator agent.

论文关键词：Multi-agent system,Reinforcement learning,Optimized negotiation,Mediator agent,Negotiation strategy,Adaptive learning

论文评审过程：Available online 12 June 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.06.003