DQNOCHN: Design of an Efficient Dyna Q Network for Enhancing Onboard and Offboard Charging Performance of Energy Harvesting Networks

Designing high-efficiency energy harvesting networks with support for onboard and offboard charging is a multidomain task that involves optimization of charging component ratings, design of efficient harvesting strategies, and continuous feedback optimizations. Existing harvesting models either do not support onboard and offboard charging, or showcase lower efficiency when used for multisource energy harvesting scenarios. To overcome these issues, this paper proposes the design of an efficient Dyna Q Network (DQN) for enhancing the onboard and offboard charging performance of energy harvesting networks. Energy harvesting networks are gaining popularity due to their ability to harness energy from the surrounding environment sources. However, their performance can be limited by the availability of energy and the efficiency of their charging process. The proposed Dyna Q Network utilizes reinforcement learning techniques to optimize the charging performance of energy harvesting networks. The network is designed to adapt to different energy harvesting scenarios and to make decisions that maximize the amount of energy stored in the harvesting node’s battery sets. This is achieved by using a combination of Q-learning and Dyna algorithms, which enable the network to learn from its environment and harvest energy from Radio Frequency (RF) sources. The proposed network is evaluated using simulation experiments, and the results show that it outperforms existing charging algorithms in terms of both charging efficiency and battery capacity levels. Furthermore, the network is able to adapt to different energy harvesting scenarios, making it a versatile solution for a wide range of applications.