作　　者： ; ; ; ; ;

机构地区： 北京科技大学计算机与通信工程学院

出　　处： 《哈尔滨工程大学学报》 2006年第B07期592-596,共5页

摘　　要： 首先提出了一个双边多主题的协商模型，给出了协商模型定义、协商要素及一个软件人双边协商框架，并详细讨论了协商主题、协商协议、效用函数、协商资源及协商策略，然后利用BP神经网络学习算法对协商策略中关键的最终结果集进行预测和学习，实现了软件人协商的自学习过程．与其他的相关研究相比，论文针对不完整信息环境下的协商交互过程，利用神经网络具有的网络自适应、自学习的能力，对协商最终结果集进行预测，提高了协商的效率．在协商交互中采用该框架的个体将得到更为有利的协商结果，因此更加适用于不完整信息环境下的协商． A bilateral multi-subject negotiation model is given in this paper. Negotiation model definition, negotiation element and a Softman bilateral negotiation framework are presented. And then negotiation subject, negotiation protocol, utility function, negotiation resource and negotiation strategy are discussed respectively. At last, it forecasts and learns key result collection with BP neural network learning algorithm and implements self-learning process of Softman negotiation. Compared with other relative research, this paper aims at negotiation process in an environment with incomplete information, and makes use of the ability of network self-adaptation and self-learning to predict last negotiation result set, the efficiency of negotiation has been improved. The individual adopting this framework will get more favorable negotiation results, so it suits more negotiation efficiency in incomplete information environment.

关 键 词： 软件人 协商 自学习 神经网络

领　　域： [自动化与计算机技术] [自动化与计算机技术]