User Satisfaction-Oriented Task Offloading and Resource Allocation in UAV-Assisted Network Systems

As demand grows for computation-intensive applications such as object detection, unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) task offloading systems offer an effective solution due to their flexibility and ability to operate across diverse geographic and functional environments. However, while many studies on UAV-assisted MEC systems focus on optimizing factors such as delay and energy consumption, fewer consider user satisfaction, which is an important determinant of overall system efficiency. This paper aims to address this gap. Specifically, we first develop a novel user satisfaction model by jointly considering processing delay, desired delay, marginal delay, delay sensitivity, and time sensitivity of the tasks. Then, we formulate an optimization problem that considers both computational resource allocation and task offloading decisions to maximize total user satisfaction while minimizing total energy consumption. Finally, a SAC-based US (User Satisfaction)-oriented Task Offloading (SAC-USTO) algorithm is proposed to address the task offloading problem, and a Task Size-based Resource Allocation (TS-RA) algorithm is introduced to optimize resource allocation. Simulation results demonstrate that the proposed algorithms outperform other baseline methods.