基于贝叶斯的多议题协商优化

An Integrated-Utility Based Bayesian in Multi-Issues Negotiation

在限时条件下的Agent之间的多议题协商中,虽然最差的结果是没有达成协定,而达成了一个使自己潜在利益受损的协定未必就是好的选择。在很多情况下,由于推理策略和交互机制的不完善使得Agent个体失去自己应得的利益。论文使用贝叶斯方法对协商对手进行预测,尽量使自己的初始信念准确反映对手的意识形态;并在此基础之上提出了一个优化的协商交互模型。在此模型中,Agent个体充分利用自己的预测结果,在协商成功的基础上获得尽可能多的利益。

In time-limited multi-issues negotiation among multi-Agents,although the worst result is no trade-off,but a result which lost its profit is not a good choose.In many case,the Agent lost the profit because the illation strategy and negotiation mechanism are not perfect.This paper forecasts opponent＇s belief,and then gets an optimized negotiation model.In this model,the Agent makes use of the forecast to get more profit based on negotiation.