I. Introduction
With the rapid development of the wireless technology and the Internet, an increasing number of mobile terminals (MTs) have different wireless access requirements for bandwidth and computing, which promotes mobile applications, such as online gaming, virtual reality, intelligent data processing, and other emergent services [1]–[3]. However, these new mobile applications have high energy consumption and high latency, which pose a major challenge for computing and the battery capacities of MTs. Mobile-edge computing (MEC) is a new promotion technology that supports cloud computing capabilities and edge service environments at the edges of cellular networks. The legacy base station (BS) is updated to a MEC-enabled BS (MEC-BS) by being equipped with a computing function (such as with a MEC server), Hence, the MEC-BSs can provide MT capability enhancements, which can reduce MT application execution time and MT energy consumption. The development of these new applications and services is limited by the computing capability and battery of these MTs. If data consumption and computationally intensive tasks need to be offloaded to the cloud for execution, the computing capability of an MT may be low. However, a relatively large delay will be induced when an MT connects to the cloud over a wireless network, which is not suitable for delay-sensitive tasks.