多Agent过滤框架中协作社团选择的研究

Research on the Cooperative Community Selection in Multi-Agent Filtering Framework

分布式信息过滤解决方法的核心是在信息资源的周围安装过滤社团,并建立一个综合的协作机制,既考虑每种特定资源的动态性,又尽力在运行时提高自身的性能。实验结果表明,集成的合作机制能够平均提高用户10%以上的满意度,且能够将平均过滤延时控制在4 s以内。

The essence of a proposed solution for performing distributed filtering is in both installing filtering communities around information sources and setting a comprehensive cooperation mechanism,which both takes care about the dynamics of each particular source and tries to improve itself during a runtime. Experimental results show that integrated cooperation mechanisms manage to enlarge the average user satisfaction for more than 10; while increasing the average duration of filte- ring for less than four seconds.