Energy-Efficient Task Offloading and Resource Allocation in Mobile Cloud Computing Using Edge-AI and Network Virtualization

Abstract

In the emerging landscape of Mobile Cloud Computing, energy efficiency and resource optimization are very vital challenges because most of the cloud and edge resources are increasing the execution of tasks in mobile devices. The focus of this paper is to propose a new energy-efficient task offloading and resource allocation framework in Edge-AI enabled network virtualization for dynamic management of computational tasks in mobile cloud environments. The framework allows real-time decisions of task offloading by comparing energy consumption of local execution versus edge processing and further looks at the received performance gains in executing that way. Then, based on the savings in energy and availability of the edge resources, it grades tasks for offloading. Network virtualization optimizes edge resource use by allocating resources depending on demand from tasks, leading to a reduction in latency for increased processing efficiency. The simulation results proved that our approach could really cut down energy consumption on mobile devices, with low latency and high rates of task success, better than cloud-only offloading through static edge computing methods and traditional dynamic programming.