Asymptotic Pheromone Behavior in Swarm Intelligent MANETs

Abstract

An analytical justification is proposed for the designand global routing performance of three pheromone update methodsproposed for use in Termite, a swarm intelligent routing algorithm formobile wireless ad-hoc networks. A simple model is used in orderto determine the average amount of pheromone present on a link, aswell as some basic aspects of the pheromone dynamics. This includesa tendency towards a one-zero pheromone distribution favoring thebetter link. The pheromone update methods are investigated with theperspective that link pheromone is more an estimate of link utility thansimply a routing heuristic. This allows the routing solution to berephrased from a biological analogy to a more traditional best-metricrouting terminology. A signal estimation perspective is suggested.