Statistical QoS Provisioning Optimization for RIS Assisted Short-Packet RSMA in xURLLC

This letter investigates the physical security and timeliness based statistical quality-of-service (QoS) provisioning for reconfigurable intelligent surface (RIS) assisted short-packet rate split multiple access (RSMA) in next-generation ultra-reliable low-latency communication (xURLLC) systems. Specifically, closed expressions for the secrecy age outage probability (SAOP) and decoding error probability (DEP) are derived considering imperfect channel state information (CSI) and successive interference cancellation (SIC). Under blocklength and information leakage constraints, we proposes a meta asynchronous advantage actor critic (MA3C) algorithm based on mean-field approximation (MFA) theory (MFA-MA3C) to minimize SAOP and DEP through multi-objective optimization. To enhance stability and accelerate convergence, the proposed algorithm utilizes the mean of parameter gradients to update the global network, and simulation results demonstrate that the MFA-MA3C algorithm effectively reduces SAOP and DEP compared to benchmark schemes.