Optimal Auction for Effective Energy Management in UAV-Assisted Vehicular Metaverse Synchronization Systems
Nguyen Cong Luong, Le Khac Chau, Nguyen Do Duy Anh, Nguyen Huu Sang, Shaohan Feng, Van‐Dinh Nguyen, Dusit Niyato, Dong In Kim
Abstract
In this article, we investigate an effective energy management scheme in a unmanned aerial vehicle (UAV)-assisted vehicular Metaverse synchronization system. UAVs purchase energy resources from an energy service provider (ESP) and collect data for a virtual service provider (VSP) to perform synchronization between physical objects and digital twins (DTs). The key issue is to motivate both ESP and UAVs to participate in the energy trading market. To doing so, we design an incentive mechanism that maximizes the revenue of the ESP while guaranteeing desired economic properties, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> individual rationality (IR) and incentive compatibility (IC). In particular, we first consider a single energy unit market, where a deep learning (DL)-based auction scheme is developed to construct neural networks from the analytical results of Myerson auction. The proposed DL-based auction is guaranteed to fulfill the optimal auction. We then consider a general scenario in which ESP has multiple energy units available to UAVs. A novel DL-based auction with feed-forward neural networks (FNNs) is proposed to jointly optimize the energy unit allocation and payment rules. We provide numerical results to demonstrate the performance improvement of the DL-based auction schemes compared to the classical auctions in terms of revenue, IC and IR. In particular, for the single energy unit market, the proposed DL-based auction scheme significantly improves the revenue compared with the classical auction and more interestingly, is able to avoid prevent UAVs from submitting their false values.