System delay optimization for Mobile Edge Computing

Abstract Mobile edge computing (MEC) has emerged as an effective paradigm that delivers cloud services and functions to edge devices, with the objective to further enhance quality of service (QoS) of terminal users by offloading their computation-intensive tasks. In this article, a multi-user and multi-server MEC system is considered and each user can choose one MEC server to execute its computation task. We try to minimize the system delay (i.e., the maximum server delay). The problem is decomposed into task offloading problem and transmit power allocation problem which are solved by matching theory and a heuristic idea, respectively. The experimental results show that the proposed algorithm can not only obtain less delay, but also generate less energy consumption compared with the decomposed computation offloading and resource allocation algorithm, the shortest distance based scheduling algorithm and the random scheduling algorithm, especially when the data amount of tasks is the same but the workload is random.