Cost Minimization in UAV-MEC Systems Through Joint Optimization of Service Placement, Task Assignment, and Power Allocation

In recent years, Unmanned Aerial Vehicles (UAVs) have been widely utilized across many fields, but their limited computing resources and battery capacity make it difficult for them to process computation-intensive tasks locally. The development of Mobile Edge Computing (MEC) enables UAVs connected via cellular networks to offload tasks to ground base stations equipped with MEC servers. To process heterogeneous tasks on the servers, the corresponding services, such as programs, libraries, and databases, should be placed. In this paper, we formulate a Mixed-Integer NonLinear Programming (MINLP) problem in the time-varying multi-UAV multi-MEC server system, focusing on jointly optimizing service placement, task assignment, and transmission power allocation to minimize the system cost (weighted sum of consumed time and energy). To effectively address this optimization problem, we model it as a Markov Decision Process (MDP) and propose a Deep Reinforcement Learning (DRL) based approach for online decision-making. Simulation experiments show that the proposed approach can converge quickly and stably, and more effectively reduce the system cost compared to other baseline schemes.