Adaptive power-saving mode control in NB-IoT networks using soft actor-critic reinforcement learning for optimal power management

In the evolving landscape of the Internet of Things (IoT), optimizing power efficiency in Narrowband IoT (NB-IoT) networks is crucial for extending device lifetimes while maintaining performance. This research leverages the Soft Actor-Critic (SAC) reinforcement learning algorithm to intelligently manage power-saving modes in NB-IoT devices. The study compares SAC with Proximal Policy Optimization, and Deep Q-Network. The methodology involves simulating an NB-IoT environment and evaluating performance using metrics such as total reward, overall energy efficiency, power consumption, mode count and duration, and duty cycle percentage. The SAC-based approach demonstrated significant improvements in power efficiency, achieving balanced enhancements in power conservation and network performance. These findings suggest that reinforcement learning techniques like SAC can play a pivotal role in advancing the efficiency and sustainability of NB-IoT networks, leading to prolonged device operation, reduced costs, and enhanced overall performance, thus paving the way for more resilient and scalable IoT deployments.