Mixed Integer Linear Programming for Ground Infrastructure and Inter-Satellite Link Design in Satellite Constellations

With the proliferation of megaconstellations, the design of supporting infrastructure to enable these systems broadband services presents a challenge for satellite operators aiming to optimize their performance. Critical decisions, such as the placement of ground stations, the quantity of gateway antennas, and the configuration of inter-satellite links (ISL), significantly impact both cost and data throughput. Nevertheless, current literature has largely overlooked the potential for optimization. The goal of this paper is to investigate the ground infrastructure and ISL configuration design through the usage of Mixed Integer Linear Programming (MILP). Specifically, our research encompasses two MILP formulations: 1) Determining the global distribution of ground stations, focusing on maximizing data throughput, and 2) Establishing a close-to-optimal mapping between satellites and gateways, by maximizing data throughput and minimizing latency and need for additional gateways. Using simulations and mathematical solvers, our study reveals that satellite operators can reduce by 17% the need for gateways while reducing by 46% the latency, compared to prevailing methods, while maintaining system capacity. Furthermore, our analyses underscore the advantages of constellations with extensive and distant ISL configurations, as they effectively alleviate congestion in densely populated areas.