Representing and monitoring social commitments using the event calculus

摘要

Multiagent social commitments provide a principled basis for agent interactions, and serve as a natural tool to resolve design ambiguities. Indeed, they have been the subject of considerable research for more than a decade. However, the take-up of the social commitments paradigm is yet to come. To explain this negative result, we pinpoint a number of shortcomings, which this article aims to address. We extend current commitment modelling languages, thus leveraging expressive possibilities that were precluded by previous formalizations. We propose a novel axiomatization of commitment operations in a first order Event Calculus framework, that accommodates reasoning with data and metric time. Finally, we illustrate how publicly available \({\mathcal{REC}}\) implementations can be exploited for commitment monitoring purposes.