MOBILE CLOUD OFFLOADING FRAMEWORK

Abstract

Mobile Cloud Offloading (MCO) addresses resource limitations of mobile devices by migrat- ing compute-intensive tasks to powerful remote servers. This paper presents a predictive MCO framework that incorporates a machine learning-based decision engine to intelligently select the optimal execution environment among local, edge, and cloud resources. The framework is evalu- ated using diverse workloads including matrix multiplication and image processing. Experimental results demonstrate that the ML-driven predictive approach consistently achieves lower execu- tion latency compared to static and reactive offloading strategies, validating the effectiveness of context-aware, proactive decision-making in mobile computing environments.