Throughput and Lifetime Enhancement of WSNs Using Transmission Power Control and Q-learning

In this paper, a Q-learning algorithm is proposed to improve the performance of the routing in the Wireless Sensor Networks (WSNs). We have also combined transmission power control (TPC) method with Q-learning to further improve the performance. In the proposed method, each sensor node is treated as an agent which uses Q-learning for routing decisions in distributed manner and employs TPC for transmission of data packets. In the network, agents with higher residual energy and smaller hop distance to sink are given priority to forward packets to the next hop. A convex energy function is used to calculate the effective distance, which is then used for deciding the power level to be used in sending of a packet. We have also computed and presented the time and space complexity of the proposed QL-TPC protocol. This protocol has been simulated using NS3. The results have been obtained as an average of ten simulation runs. The simulation results show improvement in network performance in term of throughput, end-to-end delay and network lifetime for different network size, packet size and propagation models. The performance of the proposed model is compared with the other existing protocols QLRP, Q-Routing, RBLR and AODV. It is observed that the proposed QL-TPC protocol outperforms all other protocols. Further, the scalability of the protocol is also investigated and observed that our proposed protocol is scalable.