Ontology-based dynamic role interaction control in multi-agent systems

For more practical uses of multi-agent systems, the security on agents' interactions is essential to allow only authorized interactions in multi-agent systems. In this paper, we propose a fine-grained Dynamic Role Interaction Access Control (DRiBAC) model to support exclusive authorization for an actual partnering agent of an interaction and dynamic control on the assignment of roles and permissions and their activation. To do this, we extend an earlier proposed RiBAC model with new concepts of role interaction, interaction assignment, and context-aware constraints. To describe the proposed model and its context and support semantic reasoning, we propose DRiBAC ontology. To verify the usability and efficiency of the ontology based DRiBAC, we implement a prototype system based on an example of an online tutoring system.