A Knowledge-based Approach for Resource Discovery and Allotment in Swarm Middleware

This paper proposes a semantic-based solution to resource discovery, allotment and sharing in swarm intelligence scenarios. The envisioned framework allows a novel and advanced retrieval of resources in highly dense contexts, based on semantics of the annotation they convey. In fact, logic-based inferences permit to determine the compatibility between a request and available resources so enhancing the quality of discovery. In order to do that, a vocabulary (ontology) sharing is unavoidable to properly annotate resources. The approach proposed here bypass centralized and encumbering Knowledge Bases (KBs) through an ontology scattering and rebuilding when needed. On the other hand, factual knowledge follows the material resources and is distributed by nature. The proposed framework has been implemented in a prototype to prove correctness of the approach and obtain an early performance evaluation. The adopted communication solution is represented by Bee Data Distribution System (Bee-DDS), a message-oriented middleware exploiting a publish-subscribe model. It provides affordable interaction among loosely-coupled resources in the scenario to support advanced discovery and sharing functionalities.