Deliver More For Less Energy

Most existing MANET Proactive routing protocols rely heavily on hop count evaluation. Although this is simple and efficient, it sacrifices the potential performance gains obtainable by considering dynamic network conditions such as congestion levels and link stability. Remedies of this issue have been presented by considering various predictive dynamic routing metrics mostly related to bandwidth and energy consumption. However, those predictions do not fully appreciate the nonstationary nature of the dynamic routing metrics and possible heterogeneous power consumption of the nodes. With these added concerns, in this work, in-depth investigation is conducted on prediction and evaluation of dynamic routing metrics including queuing delay, energy cost, and link stability. Artificial neural network models and advanced statistical methods are used for prediction of these metrics. Queuing delay and energy cost are investigated individually and then evaluated compositively togeth er with predicted link stability. Their effects on multiple routing objectives are extensively simulated and analyzed.