15th European Conference on Artificial Intelligence
|July 21-26 2002 Lyon France|
Carles Sierra, Jordi Sabater, Jaume Agusti, Pere Garcia
This paper explores the existing gap between multiagent specification and implementation and the potential help that evolutionary programming techniques can bring in. We present a methodology to help the programmer in the transition from a set of desired global properties expressed as an equation-based model that a Multi-Agent System (MAS) must fullfil to an actual society of interacting agents. The evolutionary techniques are seen, within this methodology, as a procedure to tune the parameters of the population of agents in order that their aggregated behaviour maximaly approaches the desired global properties.
Keywords: Multi-Agent Systems, Design, Genetic Algorithms
Citation: Carles Sierra, Jordi Sabater, Jaume Agusti, Pere Garcia: Evolutionary Computation in MAS design. In F. van Harmelen (ed.): ECAI2002, Proceedings of the 15th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2002, pp.188-192.