Efficient energy consumption and fault tolerant method for clustering and reliable routing in wireless sensor network

A very important in Wireless Sensor Network (WSN) field are energy efficiency and network lifetime. Organizing nodes into clusters, choosing best cluster head, analyzing routing paths, preserving the clusters are the issues that significantly affect the network lifespan. In this research work, a Double deep Q-learning dynamic Arithmetic Reinforcement Learning (DQARL) approach is proposed for energy efficient clustering and scheduling based reliable routing. In first phase, Hunter Prey Optimization (HPO) algorithm is suggested to determine the optimum head. The selection takes place based on various fitness functions. In second phase, duty cycling is done for improve network lifetime. This approach schedules every node in a distribute manner, that decreases the energy consumption of each sensor nodes effectively. Considered three modes namely sleep, listen and transmit during scheduling. In third phase, Weighted Practical Byzantine Fault Tolerance algorithm (WPBFT) is proposed for reliable routing and minimizing delay of packet transmission by selecting best route between source and cluster head. Simulation result proves that proposed method is more effective for energy consumption and tolerate more than 95% of fault.