A Review of Intelligent Computation Offloading in Multiaccess Edge Computing
Hengli Jin, Mark Gregory, Shuo Li
Abstract
Multi-Access Edge Computing (MEC) is a standardized architecture that enables cloud computing capabilities at the edge of heterogeneous networks. The concept is to reduce network congestion by running applications and network services closer to end-users. MEC is designed to be implemented at key locations on the network edge, including co-location with cellular base stations. MEC aims to facilitate computation intensive and delay sensitive applications, such as vehicular networks, face recognition, augmented reality and virtual reality. The service requirements for MEC are stochastic and time varying. Coupled with advances in artificial intelligence, a vast number of computation offloading approaches have been developed based on intelligent algorithms. This article provides a comprehensive review of intelligent computation offloading with critical issues, metrics and future directions.