Adaptive Resource Allocation for IoT With Computing Power Network Based on RIS-UAV-Aided NOMA-THz Communication

The integration of advanced technologies such as sixth-generation mobile communications (6G), artificial intelligence (AI) and blockchain has given new impetus to the development of the Internet of Things (IoT). However, these applications require higher computational power and lower latency, which present challenges to traditional network architectures. To address these issues, this paper proposes a novel computing power network (CPN) architecture based on reconfigurable intelligent surface (RIS)-unmanned aerial vehicle (UAV)-assisted non-orthogonal multiple access (NOMA)-Terahertz (THz) communication, and it aims to meet high computational demands. In the proposed scheme, CPN is introduced to assist IoT devices in executing tasks, thereby enhancing data processing. Concretely, THz communications and NOMA technologies are utilized to increase data rates and spectral efficiency. The combination of RIS and UAV shows promise in overcoming the challenges of high path loss and high sensitivity to blockage in THz communication, thereby improving system performance. To increase the efficiency of the proposed architecture, it is crucial to rationally allocate computational and transmission resources. Therefore, a joint optimization problem is formulated to minimize system consumption, encompassing both time and energy usage. To achieve efficient resource allocation, an adaptive N-Step method based on soft actor-critic (SAC) algorithm is employed. Simulation results demonstrate the superiority of the proposed method over the existing baselines