Dynamic Resource Sharing for Satellite Uplink via Online Convex Optimization

With the escalating demand of users for broadband data and real-time interactive services, both the number of loworbit satellites launched into space and the number of terminals connected to satellite networks have increased significantly. However, due to the inherent time-varying uplink traffic fluctuations of satellite terminals, the traditional pre-allocation mechanism with fixed bandwidth mentioned in DVB-S2X standards for satellite uplink communication causes resource inefficiency and significant spectrum waste. To address this challenge, most existing works have proposed dynamic resource scheduling methods, but these studies ignore the long transmission delay between the Network Operation Control Center (NOCC) and satellite terminals, which implies that the traffic of satellite terminal is regarded as a constant variable rather than a random variable. In this article, a delay scheduling problem with random user traffic is studied, and the online algorithms for dynamic resource sharing mechanism based on Online Convex Optimization (OCO) theory with dynamic regret and constraint violations are proposed. In addition, the impact of user traffic fluctuations on the online algorithms is discussed. Simulation results show that the online algorithms with random user traffic can effectively decrease online decision errors and improve uplink spectrum utilization.