Energy-efficient computation offloading and resource allocation for delay-sensitive mobile edge computing

Abstract

Mobile edge computing is an emerging computing paradigm to augment computational capabilities of mobile devices by offloading computation-intensive tasks from resource-constrained smart mobile device onto edge clouds nearby with potential computation capability. However, in general, edge clouds have limited computation resource and energy. Thus it is critical to achieve high energy efficiency while ensuring satisfactory user experience. In this paper, we first formulate the computation offloading problem into the system energy-efficiency cost minimization problem by taking into account the completion time and energy. Then, by solving the optimization problem, we propose a distributed algorithm consisting of clock frequency configuration, transmission power allocation, channel rate scheduling and offloading strategy selection. In this algorithm, the clock frequency for local execution and the transmission power in edge cloud execution are optimized jointly. Furthermore, to optimize the queue delay, we formulate an M/M/n queue model with individual capacity between different windows and propose an algorithm for optimal task offloading rate. Finally, the experimental results show that our algorithm can achieve energy-efficient offloading performance compared to other existing algorithms.