Demand-Aware User Association and Beamforming Design in Integrated Satellite-Terrestrial Networks

In this paper, we investigate the demand-aware user association and beamforming issue in integrated satellite-terrestrial networks (ISTNs), where terrestrial base stations (BSs) and satellites collaborate to match offered capacity with ground users' (GUs) demands. Our objective is to maximize this matching by jointly optimizing beamforming vectors of satellite and BS, and the user association between GU and both BS and satellite, while ensuring load balancing constraints and resource allocation constraints. The offered capacity is influenced by binary user association variables and beamforming vectors, posing a mixed-integer non-convex optimization challenge. To tackle this, we decouple the problem into a beamforming subproblem and a user association subproblem, which are solved in an alternating iterative manner. Specifically, we address the beamforming subproblem using successive convex approximation techniques, proposing a low-complexity closed-form iterative algorithm to obtain beamforming vectors for satellite and BS. Additionally, the user association subproblem is reformulated as a many-to-one matching problem with load balancing constraints, which is then effectively solved using a swap matching-based algorithm. Extensive simulation results show that the proposed algorithm better matches offered and requested capacities and significantly reduces unmet demands compared to benchmarks.