Predictive Multiple Metrics in Proactive Mobile Ad Hoc Network Routing

In this paper we present a MANET proactive routing enhancement scheme by comprehensive evaluation of multiple dynamic routing metrics, including delay, energy cost, and link stability. We developed efficient routing metric prediction methods: predicting delay and energy using double exponential smoothing, and predicting link stability using a heuristic based on the normal-like distributions of the link lifetimes in typical MANET mobility scenarios. The routing metrics predictions are incorporated with regular routing information exchanges. On each node, routing table is constructed by a modified version of Dijkstra's algorithm, which evaluates the predicted metrics values compositively. We integrated such a multi-metric prediction/evaluation mechanism into OLSR and name it OLSR MC. We show by simulation that OLSR MC is more adaptive to the network dynamics and therefore is able to improve performance significantly on multiple routing objectives, including higher packet delivery ratio, shorter average end-to- end delay, and prolonged network energy lifetime.