A Reinforcement Learning Approach for Evaluation of Disaster Relief
Gajanan Madhavrao Walunjkar, Anne Koteswara Rao · 2019
Ad hoc networks are considered more suitable for disaster scenarios due to infrastructure-less feature. Existing routing algorithms used on such mobile ad hoc networks are non-adaptive routing algorithms and based on shortest path algorithms. But shortest path may not give optimum path. If the network is heavy loaded with traffic, shortest path may not give better results. Various reinforcement algorithms are used to design such adaptive routing where the routing takes place based on the actual traffic present on a network. One of the reinforcement routing algorithms is Q routing and various variants of Q routing like CQ routing, CDRQ routing etc. In this paper, Q routing and CDRQ routing are implemented and compared using disaster area mobility model in disaster area scenario.