Joint Illumination, Power, and Band Allocation for Multi-Beam LEO Satellites With Beam-Hopping Using Mixed-Integer Linear Programming

The paper presents a novel optimization approach for resource allocation in non-geostationary orbit (NGSO) satellite constellations employing beam-hopping (BH) payloads for large-scale broadband services. Current literature has proposed several optimization strategies to deal with the intricacies of BH applied to large satellite constellations. However, most works limit their study to only sub-portions of the complete problem, often fixing certain variables, such as frequency or power, or disregarding specific objectives, such as user fairness or excess of allocated resources. To address this gap, this work introduces a novel mixed-integer linear programming (MILP) formulation that integrates all available degrees of freedom, namely, power, frequency, and beam illumination, while optimizing for performance, quality of service and fairness through three objectives: unserved capacity (UC), extra-served capacity (EC), and time-to-serve (TTS). To reduce the computational complexity, an alternative time-split MILP is proposed by partitioning the problem into smaller time-based subproblems. The time-split MILP maintains the performance of the full MILP formulation in larger user scenarios and balances execution time with superior computational performance. Simulations demonstrate that the time-split MILP method surpasses other approaches, including deterministic (DB) and meta-heuristic (GA) solutions. It achieves up to a 6% reduction in UC, maintains EC levels below those of DB and GA in low-user scenarios, and nearly halves the TTS, highlighting its efficiency in resource allocation and service optimization.