Joint UAV Placement, Task Offloading, and Resource Allocation for Low-Altitude Wireless Networks

The rapid development of the low-altitude economy has accelerated the widespread adoption of unmanned aerial vehicle (UAV) systems in various intelligent applications. However, UAV placement is fundamental to low-altitude computing services, while the limited battery and computing resources of UAVs pose a critical challenge for latency-aware offloading and resource allocation. In this paper, we propose a semantic-aware UAV-assisted edge computing architecture that extracts task-related semantic information and offloads it to edge servers, thereby enabling reliable and efficient adaptive adjustment of task offloading communication and computation. Within this network architecture, the UAV placement, task offloading, and computing resource allocation are jointly optimized to minimize the total task execution delay while satisfying the UAV resource constraints. However, the formulated problem is a large-scale mixed-integer nonlinear optimization problem, whose computational complexity increases rapidly with the number of devices. To address this challenge, the original problem is reformulated through successive convex approximation (SCA) and reformulation linearization techniques (RLT). Furthermore, a parallel optimization framework is proposed to decompose the large-scale optimization problem into multiple low-complexity subproblems that can be solved in parallel, thereby accelerating the computation process and reducing the overall computational complexity. Simulation results demonstrate that the proposed method achieves near-optimal performance with significantly lower computational complexity. Compared with existing schemes, the proposed algorithm effectively reduces task execution delay for users and ensures stable and efficient system operation.