Reinforcement Learning based Hybrid Optimum Route Identification Scheme in WSN

Wireless Sensor Networks (WSNs) has gained lot of interest for researchers especially in monitoring and surveillance applications. Due to limited energy in sensor nodes has created challenge in transmitting the sensed data to base station by developing appropriate routing scheme. Considering the major issues of minimizing energy consumption and maximizing network lifetime, this work proposes a hybrid energy efficient Reinforcement Learning (RL) based optimal route identification scheme for WSN. RL based Q-learning algorithm is used to determine the probable next hop in the route considering degree of connectivity, link quality and available bandwidth. Gravitational Search Algorithm is applied to fine tune the routing choices using fitness function. Q value, residual energy and distance from base station to cluster head are used to derive the fitness function for GSA. Performance effectiveness of the proposed work is validated through simulation results.