Joint optimization of UAV-WPT and mixed task offloading strategies with shared mode in SAG-PIoT: A MAD4PG approach

Abstract

The Space-Air-Ground Power Internet of Things (SAG-PIoT) is a promising paradigm for the development of emerging smart grid systems, and Unmanned Aerial Vehicle (UAV) wireless power transmission (UAV-WPT) is also a promising technology. Utilizing UAV and satellites as aerial base stations to assist in PIoT communication and task offloading can solve problems such as limited SAG-PIoT network coverage, limited computing capabilities, and difficulties in energy acquisition, ensuring the safe and stable operation of SAG-PIoT. In this paper, we propose a mixed offloading strategy coexisting with partial offloading and binary offloading under the UAV-WPT assisted power supply and shared offloading mode, and integrate a multi-agent distributed deep deterministic policy gradient (MAD4PG) framework. The UAV-WPT and task offloading problem is formulated as a joint optimization problem, and the MAD4PG algorithm is utilized to find the optimal strategy in task offloading for resource allocation and energy supplementation. Finally, a simulation model is constructed based on the actual data from the Belgian power grid. Compared to the state-of-the-art, a large number of experimental results show that our proposed method has significant advantages. It reduces task completion latency by 18.93 % and lowers energy consumption by 28.14 %.