Collaborative Computation Offloading for Multi-access Edge Computing

Computation offloading is a proven successful paradigm for enabling resource-intensive applications on the mobile devices in multi-access edge computing (MEC) network. Moreover, in view of emerging mobile collaborative application (MCA), the offloaded tasks can be duplicated when multiple users are in the same proximity. This motivates us to design novel collaborative offloading schemes. In this context, we first consider the MEC offloading scenario, where multiple mobile users offload duplicated computation tasks to the network edge servers, and share the computation results among them. Our goal is to develop the optimal fine-grained collaborative offloading strategies with data caching enhancements to minimize the overall execution delay at the mobile terminal side. Next, we extend the MEC offloading to hybrid offloading (i.e., joint MEC and Device-to-Device (D2D) offloading) with social relationship consideration, and propose a hybrid D2D multicast-based task execution framework to achieve an energy-efficient task assignment policy for mobile users. To overcome the great complexity for the deployment problems, we then formulate the offloading decision problem as a multi-label classification problem and develop the Deep Supervised Learning (DSL) method to achieve a rapid offloading decisions making, as well as minimize the computation and offloading overhead. Last but not least, we evaluate their performance through extensive numerical study, which shows the superior performance of the proposed scheme.