Graph-Transfromational Swarms : A Graph-Transformational Approach to Swarm Computation

Abstract

Computer systems are becoming increasingly distributed and interconnected. Various emerging notions, such as smart grids, system of systems, industry 4.0 or cyber-physical systems have gained more and more importance during the last few years. All of them propose to solve engineering problems by using several autonomous components that act in parallel and are interconnected, foremost using Internet technologies. These emerging concepts look very promising, but also exhibit various technical challenges. For instance, how is it possible to develop decentralized control mechanisms that produce a desired emerging behavior to solve a given task or how to model such solutions in order to analyze their behavior in terms of complexity and correctness? These are two major questions that this thesis attempts to answer. Indeed, it provides graph-transformational swarms as a novel concept that combines the ideas and principles of swarms and swarm computing and the formal methods of graph transformation to model distributed systems. Graph-transformational swarms captures the advantages of swarms and swarm computing and of graph transformation.