Caching and 3D Deployment Strategy for Scalable Videos in Cache-Enabled Multi-UAV Networks

Unleashing the potential of unmanned aerial vehicles (UAVs) to play the role of caching devices for furnishing layered video services to ground users (GUs) holds tremendous promise. However, it is crucial to strike a balance between the quality and the diversity of video content when deciding the placement strategies of video caching. In this study, we examine the feasibility of implementing a layered caching scheme utilizing scalable video coding (SVC) in three-dimensional (3D) multi-UAV networks with cache capabilities. The objective is to investigate the minimization of the cumulative video access delay experienced by all GUs through the joint design of layered scalable video caching, user-UAV association, and 3D UAV deployment. This approach enables the network to offer video services across a range of quality levels. We have formulated an intricate mixed-integer non-convex optimization problem which poses a significant challenge to solve directly. Then we decompose the formulated problem into three interrelated sub-problems and put forth specific algorithms designed to address each component, i.e., layered caching strategy and UAV's two-dimensional (2D) deployment optimization with penalty based difference-of-convex (penalty-DC) framework, user-UAV association optimization with bipartite graph based linear programming (LP) relaxation, as well as UAV altitude optimization with successive convex approximation (SCA) method. Then we introduce a highly efficient iterative algorithm that employs an alternate optimization approach for three sub-problems until convergence. Our simulation results demonstrate that our proposed algorithm surpasses other benchmark schemes significantly in terms of mitigating video access delay.