A.M.: Emergence in cognitive multi-agent systems

Abstract: Emergence is a hot topic of the present research in the domain of complex systems, especially connected with the modelling and simulation of large multi-agent systems. Emergent properties have been obtained in systems inspired from nature – ants, wasps, etc – but also in simple artificial systems like Cellular Automata – obtaining gliding structures – and Transporting Agents – obtaining traffic directions. Most of the systems designed to manifest (truly) emergent properties are made of reactive or mostly reactive agents, with a limited set of actions that they apply as a direct consequence of certain stimuli. As more computational power is available and cognitive multi-agent systems tend to become a mature technology, the study of emergence in the context of such systems is a promising direction of development. Starting from the analysis of emergence in reactive systems, this paper is an attempt to define the different forms and requirements of emergent properties in the context of cognitive multi-agent systems and highlights different settings in which such properties may significantly enhance the overall performance of the system