A Comprehensive Analysis of KNN and Order Preserving Techniques for Offloading Efficiency in MEC Networks

This paper explores the performance of K-Nearest Neighbors (KNN) versus Order Preserving (OP) algorithms of optimizing dynamic computation offloading in wireless-powered mobile-edge computing (MEC) networks. The KNN significantly outperforms over OP in terms of normalized computation rate and training losses, those offering better adaptability for changing wireless channel conditions. The proximity-based decision-making of KNN enables faster convergence and improved computational efficiency compared to OP focus on task order. As a result, KNN is more effective for real-time applications such as augmented reality (AR), virtual reality (VR), and autonomous systems. The presented effectiveness of KNN in this context suggests its broader applicability in future heterogeneous networks, including smart cities and healthcare, where real-time computational efficiency is critical.