Discrete event modeling of swarm intelligence based routing in network systems

摘要

Simulation remains attractive for performance and scalability analysis and/or design of networks. This paper presents a biologically inspired discrete-event modeling approach for simulating alternative computer network protocols. This approach identifies and incorporates the key attributes of honeybees and their societal properties into simulation models that are formalized according to the Discrete Event System Specification (DEVS) formalism. We describe our approach with particular emphasis on how to model the individual honeybees and their cooperation. These models, collectively referred to as SwarmNet, support routing algorithms akin to honeybees searching for and foraging on food. Adaptation and probabilistic specifications are introduced into honeybee (BEE) and Routing Information Protocol (RIP) routing algorithms. A set of simulation experiments are developed to show the biologically inspired network modeling with the BEE routing algorithm, as compared with the RIP routing algorithm, offers favorable throughput and delay performance and also exhibit superior survivability against network load surges. The paper concludes with some observations on the SwarmNet modeling approach and outlines some future research directions.