Using Jason to Develop Normative Agents

摘要

Norms have become one of the most promising mechanisms of social control to ensure a desirable social order in open multi-agent systems where autonomous, heterogeneous and independently designed entities can work towards similar or different ends. Norms regulate the behaviour of agents by defining permissions, obligations and prohibitions, and by stating stimulus to their fulfilment while defining rewards and discouraging their violation while pointing out punishments. Since goal-oriented agents' priority is the satisfaction of their own desires, they must evaluate the positive and negative effects of the fulfilment or violation of the norms before choosing to comply or not with them. In this context, we present the new functions of the Jason platform defined to support normative reasoning, i.e, to build agents able to deal with desires and norms. Agents are then able to check if they should adopt or not the norm, evaluate the effects of the fulfilment or violation of the norm on their desires, detect and solve conflicts among norms, and select desires and plans according to their choices of fulfilling or not a norm. We demonstrate the applicability of such new functions through a non-combatant evacuation scenario.