论文部分内容阅读
如何高效地使用Agent学习机制进行在线协商,已经成为经济学家和计算机学者共同研讨的一个主要方向。,为了当协商进入僵持状态时参与协商的买卖双方能确定是否进行妥协,从而使协商继续进行下去,本文在限时条件下的多议题协商中和贝叶斯学习的基础上提出了基于不妥协度的协商策略。实验表明在协商过程中进行学习可以提高对方私有信息的预测精确度,缩短了协商时间,提高了协商效率。
How to effectively use Agent learning mechanism to conduct online consultation has become a major direction for economists and computer scientists to discuss together. In order to determine whether or not to compromise the buyers and sellers participating in the negotiation when negotiations enter into a stalemate, so that the negotiation can be continued. Based on the multi-topic negotiation under limited conditions and Bayesian learning, Negotiation strategy. Experiments show that learning during the negotiation process can improve the prediction accuracy of private information of each other, shorten the negotiation time and improve the efficiency of the negotiation.