Bringing about rationality: Incorporating plans into a BDI agent architecture

Abstract

Agents attempt to achieve their intentions through the use of plans, leading to further intentions corresponding to the actions and subgoals of those plans. We extend Rao and Georgeff's logic of belief, desire and intention with a logical representation of plans. This representation allows the specification of subgoals, using Segerberg's “bringing it about” modal operator. By using a plan library to semantically restrict an agent's intention-worlds, we define a framework that models the reasoning process of an intention-based autonomous agent. We show that this framework supports several desirable properties involving an agent's commitment to future intentions, based on its available plans.

The authors wish to thank the referees for several useful suggestions. This work was partially supported by the Cooperative Research Centre for Intelligent Decision Systems, Melbourne, Australia.