A Stochastic Algorithm for Self-Organization of Autonomous Swarms
Date: December 12 - December 15, 2005
In earlier work of the authors simulation results indicated the possibility of achieving self-organization of autonomous vehicles through Gibbs sampler-based simulated annealing. However, the dynamic graph structure associated with the network evolution presents challenges in convergence analysis. In this paper a novel algorithm is presented and shown to yield desired global configurations with primarily local interactions. Its convergence speed is provided in terms of the Gibbs potential function. The analytical results are further verified through simulation.