Analyzing stigmergetic algorithms through automata games Host Publication: Workshop on Knowledge Discovery and Emergent Complexity in BioInformatics (KDECB 2006) Authors: P. Vrancx, K. Verbeeck and A. Nowé Publisher: Springer Publication Year: 2007 Number of Pages: 12 ISBN: 978-3-540-71036-3
Abstract: Abstract. Stigmergy describes a class of mechanisms that mediate an-
imal to animal interaction through the environment. Recently this con-
cept has proved interesting for use in multi-agent systems, as it provides a
simple framework for agent interaction and coordination. However, deter-
mining the global system behavior that will arise from local stigmergetic
interactions is a complex problem. In this paper stigmergetic mecha-
nisms are modeled using simple reinforcement learners, called learning
automata.We show that using automata to model stigmergy, the learn-
ing problem can be asymptotically approximated by an automata game.
Existing convergence results for automata games enables us to under-
stand these stigmergetic methods and predict their global behavior. A
simple multi-pheromone example is described and analyzed through its
corresponding automata game.Abstract. Stigmergy describes a class of mechanisms that mediate an-
imal to animal interaction through the environment. Recently this con-
cept has proved interesting for use in multi-agent systems, as it provides a
simple framework for agent interaction and coordination. However, deter-
mining the global system behavior that will arise from local stigmergetic
interactions is a complex problem. In this paper stigmergetic mecha-
nisms are modeled using simple reinforcement learners, called learning
automata.We show that using automata to model stigmergy, the learn-
ing problem can be asymptotically approximated by an automata game.
Existing convergence results for automata games enables us to under-
stand these stigmergetic methods and predict their global behavior. A
simple multi-pheromone example is described and analyzed through its
corresponding automata game.
|