FE4NorOMAS: A Distributed Fuzzy Enforcement Approach for Normative Open Multi-Agent Systems

Abstract

In the Normative Multi-agent Systems (NorMAS) field, norms are designed and specified as regulatory mechanisms intending to outline the ideal and appropriate behavior for the agents. To that end, norms are enforced via sanctions, considering that agents can choose to comply with the norm. Enforcement of norms responds to norm violations (i.e., punishment) or compliance (i.e., reward), which has recently been accepted as a sanction. Also, fuzzy enforcement refers to norm enforcement based on provided fuzzy reasoning process outputs. In this paper, we propose a novel fuzzy enforcement approach called FE4NorOMAS for distributed Fuzzy Enforcement FOR Normative Open Multi-Agent Systems aiming to make sanctions more flexible in the sense that they will be performed over several levels of restrictions. Flexible sanctions allow autonomous agents to accommodate their behaviors according to the norm. The more the behavior follows the norm, the greater the reward. On the other hand, the greater the violation of the norm, the greater the penalty. The feasibility of the proposed approach is tested using several traffic signing system scenarios on the MaDKit agent platform.