Bayesian Classifier model for reactive routing improvement in mobile ad-hoc networks

This paper introduces a Bayesian Classifier model to improve the performance for reactive routing protocols in mobile ad-hoc networks (MANETs). The model can help to increase the network throughput and reduce end-to-end delay by controlling the broadcast area. Additional information related to hop-counts and node densities are used to support routing protocol in broadcasting. It is shown that the achieved throughput is increased twice and the delay is reduced by 50% in comparison to existing conventional routing protocols. The number of control packets disseminated during route discovery process decreased significantly in comparison with conventional scheme.