Strategy coordination approach for safe learning about novel filtering strategies in multi agent framework

Abstract

In commercial and information reach society, the properties of novel filtering strategies have to be explored without dramatically increasing response time while trying to combine them to effectively use available system resources. The major drawback of many existing systems, which try to make different synergies between filtering strategies, is usually concerned with not taking care of the availability of resources, being especially critical for the realisation of successful commercial deployments. The essence of a presented solution is both in the encapsulation of many known searching algorithms inside separate filtering agents, and in the integration of flexible resource aware coordination mechanisms into one manager agent. The flexibility of a realised coordination scheme in facilitating an easy integration of novel strategies is practically demonstrated in an intelligent personal information assistant (PIA). Experimental results, obtained during a 2 week internal PIA usage, show the elimination of jobs longer than 1000s together with an increase of up to 10% in a received feedback values. © Springer-Verlag Berlin Heidelberg 2006.