A formal performance modeling framework for bio-inspired ad hoc routing protocols
Proceedings of the 10th annual conference on Genetic and evolutionary computation
Bee-sensor: a step towards meta-routing strategies in hybrid ad hoc networks
ANTS'10 Proceedings of the 7th international conference on Swarm intelligence
Hi-index | 0.00 |
A biologically inspired algorithm is presented to route messages in mobile wireless ad-hoc networks. Such computer networks are primarily characterized by their quickly changing topologies due to frequent node mobility. The principles of swarm intelligence are used to define a probabilistic algorithm for which routing through paths of maximum utility is an emergent property. This adaptive algorithm, dubbed Termite, uses stigmergy to reduce the amount of control traffic needed to maintain performance. Strong routing robustness is achieved through the use of multiple paths; each packet is routed randomly and independently. Once the basic operation of Termite is verified, alternative metric estimation techniques are tested via simulation. Optimal system parameters are selected by testing over orders of magnitude. A simple analytical model is built in order to explain the simulation results. The model also used to propose two heuristics for determining the optimal pheromone decay rate. All of the enhancements to Termite, now known as ReTermite, are consolidated and tested against Ad-hoc On-demand Distance Vector (AODV), a leading ad-hoc routing algorithm. ReTermite is shown to be superior in many primary metrics and the reasons for this explained. Previous heuristic models are also compared to simulation results.