GAI-Based Resource Management in RIS-Aided Next-Generation Network and Communication

Abstract

Reconfigurable intelligent surface (RIS) is introduced as a key technology of the sixth generation mobile network (6G) to build RIS-aided next-generation network and communication. In this paper, according to different devices and scenarios, a flexible channel distribution learning (CDL) method is designed to perform efficient base station (BS)-RIS-device cascade channel estimation to adapt to the dynamic and changeable next-generation network environment. According to different service types, generative artificial intelligence (GAI) and distributional reinforcement learning (DBRL) are innovatively combined to propose a method of on-demand network resource allocation in RIS-aided wireless network. The goal is to maximize system utility of joint energy efficiency (EE) and quality of service satisfaction rate (QoSSR), provide higher quality of service (QoS), and achieve efficient resource allocation and management in the next-generation network and communication environment. In addition, The proposed algorithm’s effectiveness is verified through a lot of simulation and numerical analysis. The results show that this algorithm significantly improves system utility, enhancing adaptability and improving QoS.