Energy efficient task offloading from IoHT in delay sensitive edge networks
Amit Kumar, Jitendra Kumar Samriya, Ashok Bhansali, Meena Malik, Varsha Arya, Wadee Alhalabi, Eaman Alharbi, Brij Bhooshan Gupta · Alexandria Engineering Journal · 2025
The edge server connected to the Internet of Things (IoT) is designed to provide optimized resource allocation and task processing, particularly for delay-sensitive applications. This becomes especially challenging when certain tasks demand lower response times compared to others. In this paper, we develop a deep learning model to optimize the task offloading problem in a multi-user Internet of Health Things (IoHT) edge environment. The proposed model, based on a Deep Reinforcement Learning (DRL) architecture, is trained on various input data to manage task offloading across data centres within the edge network. It effectively addresses the challenges of optimizing both offloading probability and power allocation. Simulations are conducted to evaluate the energy efficiency of IoHT devices and their computational offloading capabilities in the cloud. The results demonstrate that the proposed method achieves a higher response rate and greater energy savings compared to existing approaches.