Mrs. MORDJANE Maissoun Mrs. MEBARKI Aridje2025-04-092025-04-092024http://dspace.univ-khenchela.dz:4000/handle/123456789/8662In an era marked by rapid technological advancement, smartphones have become integral to our daily lives, transforming how we engage with the world. Despite their indispensability, these powerful devices struggle with a variety of performance issues. Among these, limited battery life stands out as a particularly critical challenge. Mobile Cloud Computing offers a promising solution by leveraging remote servers to offload resource-intensive tasks. However, existing approaches grapple to adapt to the dynamic nature of mobile environments, leading to inefficient energy consumption. To address this and enhance the overall system efficiency, this work is dedicated to proposing an approach that integrates machine learning techniques such as Reinforcement Learning and K-means clustering. This approach dynamically partitions mobile applications into offloadable tasks for remote execution on cloud servers and unoffloadable tasks for local execution on the mobile device. Through experimentation and evaluation, we demonstrate the efficacy of our approach in improving energy efficiency and performance.enThesis