Adaptive Multi-Objective Reinforcement Learning-Based Energy-Efficient Routing in Wireless Sensor Networks

Abstract

Wireless Sensor Networks (WSNs) are widely used in monitoring and IoT applications, where efficient routing is critical due to limited node energy and dynamic network conditions. Traditional routing techniques such as greedy routing suffer from issues like local minima, lack of energy awareness, and uneven load distribution. To address these challenges, this paper proposes an adaptive multi-phase routing framework integrating cost-based optimization, multi-objective decision-making, and reinforcement learning (RL). Initially, greedy routing is implemented as a baseline to highlight fundamental limitations. A cost-based routing approach is introduced to improve link quality by incorporating distance and communication constraints. Further enhancement is achieved through a multi-objective routing model that considers residual energy and node load, ensuring balanced and energy-efficient data transmission. Finally, a Q-learning-based routing mechanism enables nodes to learn optimal routing paths dynamically. Simulation results demonstrate that the proposed RL-based approach significantly improves packet delivery ratio, enhances network lifetime, and achieves better adaptability compared to conventional routing methods.