Towards modelling of reactive, goal-oriented and hybrid intelligent agents using P systems

Abstract

Intelligent agents are classified into various types depending on whether they just react to the stimuli they perceive (reactive) or they develop plans to solve their own goals (proactive or goal-oriented). In practice, agents are a mixture of two layers since they perform reactive or proactive tasks depending on what is the most appropriate at a given time (hybrid agents). Bearing in mind the dynamic organisation of a multi-agent system consisting of any of the above types, it is only natural to consider Population P Systems as a suitable candidate for modelling. In this paper, we describe preliminary work done towards modelling of MAS which include all types of agents. An initial attempt is made to tackle certain issues that have to do with the objects and rules that define each agent operation. Alongside the alternative solutions, we present a concrete example to demonstrate our findings and raise discussions. © 2010 Springer-Verlag.