RAVEN: Reverse Auction-Based Resource Allocation for UAV-Enabled Edge Computing

Abstract

In this work, we explore the problem of optimal resource allocation in an Unmanned Aerial Vehicle (UAV)-enabled edge computing platform. In the existing literature, researchers have focused on developing edge platforms in the presence of UAVs and ground nodes. However, in real-world scenarios requiring temporary computing, such as augmented reality, industrial IoT, and situational events, deploying ground stations to enable edge computing is difficult. To address this issue, we propose a reverse-auction-based resource-allocation scheme, RAVEN, for UAV-enabled edge networks. The proposed architecture operates without a ground node, distributing all computations exclusively among the UAVs. In RAVEN, we aim to maximize overall system utility by employing a unique bidding strategy that ensures user satisfaction and solves the winner-determination problem for efficient resource allocation. We develop an algorithm that approximates the winner determination problem in polynomial time, facilitating practical implementation. We also evaluated the performance of RAVEN through simulations and theoretical analysis, demonstrating its feasibility and effectiveness.