Resource and Job Aware Coordination in Multi Agent Filtering Framework

S. Albayrak and D. Milosevic (Germany)


Intelligent Agents, Intelligent Information Systems, Prob abilistic Reasoning, Multi Agent Filtering Framework, Distributed Artificial Intelligence


In nowadays information reach society, filtering strategies have to be combined respecting both the availabilities of resources and the properties of assigned jobs. The major drawback of many existed systems, which try to make different synergies between known filtering strategies, is usually concerned with not taking care of resource avail ability, being critical for the presenting of successful de ployments. The cornerstone of the presented solution is both in the encapsulation of many existed searching algo rithms inside separate filtering agents, and in the integra tion of comprehensive resource and job aware coordina tion mechanisms into one manager agent. The flexibility of the realised coordination scheme in addressing a trade off between response time and result relevance is practi cally demonstrated in a system serving as intelligent per sonal information assistant (PIA). Experimental results, obtained during a two week internal PIA usage, show the elimination of long-lasting jobs with duration over 1000s and an increase of up to 12% in a received feedback value.

Important Links:

Go Back