Norm representation and reasoning: A formalization in event calculus

Abstract

Norms play an important role in coordinating, regulating and predicting agents' behavior in open multi-agent societies. Much work has been done on modeling and developing normative multi-agent systems. Norms in open multi-agent societies are not fixed, they might emerge, change or vanish; therefore agents need a mechanism to adapt their behavior accordingly. Using Event Calculus we propose a formal representation of prohibition and obligation norms. This includes the norm's context, rewards and sanctions. Using this formalization we propose a technique for BDI agents to reason at run time about their behavior taking into consideration current norms and past actions performed by the agent. In this work, we assume that the best behavior of an agent is the behavior with maximum utility. Our technique has been applied to a simple mining simulation. © 2013 Springer-Verlag.