Community Swarm Optimization

Rawan Ghnemat (1) , Cyrille Bertelle (1) , Gérard Henry Edmond Duchamp (2)

1 LITIS - Laboratoire d'Informatique, de Traitement de l'Information et des Systèmes
2 LIPN - Laboratoire d'Informatique de Paris-Nord

The development of distributed computations and complex systems modelling leads to the creation of innovative algorithms based on interacting virtual entities, specifically for optimisation purposes within complex phenomena. Particule Swarm Optimisation (PSO) and Ant Colony Optimisation (ACO) are two of these algorithms. We propose in this paper a method called Community Swarm Optimisation (CSO). This method is based on more sophisticated entities which are defined by behavioral automata. This algorithm leads to the emergence of the solution by the co-evolution of their behavioral and spatial characteristics. This method is suitable for urban management, in order to improve the understanding of the individual behaviors over the emergent urban organizations.

Mots clés: swarm optimization; community detection; self-organization; automaton; evolutionary methods; geographic systems; resource management

Domaines: Adaptation et Systèmes auto-organisés [nlin.AO]; Intelligence artificielle [cs.AI]

Rawan Ghnemat, Cyrille Bertelle, Gérard Henry Edmond Duchamp. Community Swarm Optimization. M.A Aziz-Alaoui and Cyrille Bertelle. From System Complexity to Emergent Properties, Springer-Varlag, pp.195-207, 2009, Understanding Complex Systems, ⟨10.1007/978-3-642-02199-2⟩. ⟨hal-00430547⟩