Automata with Multiplicities as Behaviour Model in Multi-Agent Simulations

摘要

In this paper, we present how the use of automata is able to produce tools for multi-agent simulations. First we present the agent behaviour problematic and the necessity of automatic processing for collective behaviour to take into account macro-micro interactions which lead us to use a lot of dynamic entities. In the second part of our presentation, we show that automata can give computable tools to respond to these different problematics. We present some computable architectures based on automata with multiplicities representing agent behaviours, in particular for Eco-Problem-Solving. We finally explore a domain of application in ecosystems modelling which takes into account natural and dynamic structural organisations detection and their evolution by using the previously described formalism.