| 基于SVM的多Agent协商伙伴选择 |
| |
| 引用本文： | 谷琦松,刘胜全. 基于SVM的多Agent协商伙伴选择[J]. 计算机应用与软件, 2012, 29(6): 131-134 |
| |
| 作者姓名： | 谷琦松 刘胜全 |
| |
| 作者单位： | 1. 新疆大学信息科学与工程学院 新疆乌鲁木齐830046
2. 新疆大学信息科学与工程学院 新疆乌鲁木齐830046;新疆大学现代教育技术中心 新疆乌鲁木齐830046 |
| |
| 基金项目： | 新疆维吾尔自治区科技厅科研项目(200931103) |
| |
| 摘 要： | 根据多Agent协商问题的交互特点,引入SVM(Support Vector Machine)分类方法对Agent的协商历史信息进行学习,从Agent的协商历史信息中提取样本来训练SVM,结合模拟协商过程和己方的决策信息,预测与特定伙伴协商时可能出现的结果以及相应的协商收益,根据Agent的自利性原则,选择最合适的协商伙伴.最后,通过仿真实验验证了所提出方法的有效性和优越性.
|
| 关 键 词： | 多Agent协商 效用 伙伴选择 支持向量机 |
SVM-BASED MULTI-AGENT NEGOTIATION PARTNER SELECTION |
| |
| Gu Qisong , Liu Shengquan. SVM-BASED MULTI-AGENT NEGOTIATION PARTNER SELECTION[J]. Computer Applications and Software, 2012, 29(6): 131-134 |
| |
| Authors: | Gu Qisong Liu Shengquan |
| |
| Affiliation: | 1,2 1(School of Information Science and Engineering,Xinjiang University,Urumqi 830046,Xinjiang,China) 2(Modern Education Technology Center,Xinjiang University,Urumqi 830046,Xinjiang,China) |
| |
| Abstract: | According to the interactive features of Multi-agent negotiation problem,SVM(Support Vector Machine) classification method is involved in to study the Agent’s negotiation history information,extract samples from the Agent’s negotiation history information to train SVM,and combine the simulated negotiation process with one’s decision-making information to predict possible results when negotiating with a particular partner and the corresponding negotiation revenue.Thus,depending on the Agent’s self-interest principle,the most appropriate negotiation partner is selected.Finally,the effectiveness and superiority of the method presented in this paper are verified through simulation experiments. |
| |
| Keywords: | Multi-agent negotiation Utility Partner selection Support vector machine |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
