ENERGY EFFICIENT GEOCASTING BASED ON Q-LEARNING FOR WIRELESS SENSOR NETWORKS

In this paper, we propose two energy efficient geocasting protocols based on Q-learning for wireless sensor networks (WSNs), called FERMA-QL and FER-MA-QL-E. We utilize the theorem of Fermat point to find Fermat points in geocasting, the node which is the closest to the Fermat points is selected as the relay nodes. Then, we establish the shared path among gateways, relay nodes and base station by Q-learning for data transmission. In FERMA-QL, the reward is given by the reciprocal of the distance between the received node and the destination node In FERMA-QL-E, the reward is given by the remaining energy of the received node divided by the distance between itself and the destination node. Sensors utilize the shared path to forward their data to achieve goal of reduce energy consumption. Simulation result shows that the proposed FERMA-QL and FERMA-QL-E can efficiently extend the life-time of the WSN.