Edge Computing for Metaverse: Incentive Mechanism versus Semantic Communication

Abstract

We investigate incentive mechanism designs for edge computing trading between virtual service providers (VSPs) and an edge computing provider (ECP). The VSPs deploy unmanned aerial vehicles (UAVs) to collect sensing data from physical objects for updating their digital twins (DTs). In the case with a single computing unit, we design a deep learning (DL)-based auction constructed from the Myerson theorem to maximize the ECP's revenue and guarantee incentive compatibility (IC) and individual rationality (IR). In the case of multiple computing units, a DL-based auction based on an augmented Lagrangian method is proposed that maximizes the ECP's revenue and guarantees IC, IR, and budget (BG) constraints. A semantic communication (SemCom) technique is employed to reduce the collected data and offloading cost for the VSPs. To train the deep learning algorithms, we use valuations of the computing resources to the VSPs, which particularly are a function of the age of DT, semantic symbol size, and communication time of the UAVs. We provide numerical results showing that the proposed auctions outperform the classical auctions in terms of ECP's revenue, IR, IC, BG, and their ability of preventing the false bid submissions. Also, SemCom reduces the offloading cost for the VSPs.