EAR: Energy-Aware Risk-averse Routing for Disaster Response Networks

Disaster Response Networks (DRNs) are specific delay-tolerant networks (DTNs) designed to support temporary communications during disaster recovery with limited energy resources. Existing research on DRNs routing addresses routing metrics like packet delivery delay (PDD), packet delivery ratio (PDR), and total transmission energy (TTE). However, the complex interplay among these metrics remains unclear and makes their control difficult. Moreover, the standard deviation of packet delivery delay (PDS), which is overlooked by the previous works, is an important metric in DRNs. In this paper, we propose the Energy-Aware Risk-averse routing framework (EAR) for DRNs. EAR models the mobility in DRNs using a stochastic multigraph and computes the risks associated with routing paths using the Mean-Risk and Max-Probability models. A “λ-optimal” algorithm is used for finding routing paths with the minimum risks. A parameter L restricting the number of packet replicas and a differentiated service model are utilized to improve packet delivery efficiency (PDE) while maintaining other metrics at an adequate level. We show via extensive simulations that EAR provides flexible control of the routing risks and delivers packets to the destinations in a more energy-efficient way (up to 8x higher PDE with 4% lower PDR) than well-known routing protocols Prophet, MaxProp, RAPID and Spray&Wait.