Q-Learning for energy balancing and avoiding the void hole routing protocol in underwater sensor networks

Abstract

In energy constraint networks, the utilization of limited node battery is very crucial to enhance the network lifespan. The imbalanced node battery dissipation greatly effects the performance of the network. In this paper, we propose Q-Learning based energy-efficient and balanced data gathering routing protocol (QL-EEBDG). The effectiveness of a forwarder node is computed based on; residual energy of the source node and group energies of the neighbour nodes. The consideration of energy parameters provides complete control on the forwarder node selection and ensures efficient energy consumptions in the network. Still, due to topology changes, void node occurs which is avoided through adjacent node technique (QL-EEBDG-ADN). This scheme finds an alternate route via neighbor nodes to provide continuous communication among the network nodes. Simulations are performed to validate the effectiveness of proposed schemes against existing scheme based on energy tax, network lifetime.