AROA: Adam Remora Optimization Algorithm and Deep Q network for energy harvesting in Fog-IoV network

Abstract

Electric Vehicles (EV) has gained immense popularity due to the increasing awareness amongst people regarding low carbon emission. Smart vehicles have become a central part of the Internet of vehicles (IoV) infrastructure. Whenever several vehicles distribute its tasks, then the classical centralized model meet various issues, such as security and delay in communication. This paper devises a technique for energy harvesting in the Fog-IoV network. The Fog-IoV network simulation is done for enhanced processing. The three layers, such as fog layer, cloud layer, and IoV layer are adapted for electricity trading. The power prediction is performed with a deep reinforcement learning technique, namely Deep Q network (DQN). The optimal electricity trading is done with the proposed Adam Remora Optimization Algorithm (AROA). The AROA is obtained by the amalgamation of Remora Optimization Algorithm (ROA) and Adam optimization algorithm. The EVs represents buyer of electricity that demands electricity in such a way that Road side unit (RSU) perform bidding. The fitness function is newly modelled using predicted power, price, and distance. The experimentation of the technique is done in terms of fitness, power, and pricing. The proposed AROA-based DQN offered enhanced performance with the highest power of 11.920 and the smallest pricing of 16.949%.