Self improving coordination in multi agent filtering framework

Abstract

In order to obtain acceptable quality of filtering services in real time conditions trade-off between result relevance and response time has to be addressed. Ignoring resource availability is a major drawback for many existed systems which try to boost quality by making different synergies between filtering strategies. The essence of the proposed solution for combining filtering strategies is in a comprehensive coordination which both takes care about current resource usability and tries to improve itself during a runtime. The applicability of the presented coordination between filtering strategies is illustrated in a system serving as intelligent personal information assistant (PIA). Experimental results show that long lasting jobs with duration over 1000 seconds are eliminated and that at the same time jobs, being shorter than 10 seconds, can be effectively used for adaptation.