A Genetic Algorithm for Joint Power and Bandwidth Allocation in Multibeam Satellite Systems

• Published in: 2019 IEEE Aerospace Conference pp. 1 - 15
• Main Authors: Paris, Aleix, Del Portillo, Inigo, Cameron, Bruce, Crawley, Edward
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-03-2019
• Subjects: Bandwidth; Channel allocation; Interference; Mathematical model; Radio frequency; Resource management; Satellites
• DOI: 10.1109/AERO.2019.8742238

Communications satellites are becoming more flexible and capable in order to make better use of on-board resources and the available spectrum, and to satisfy the varying demands within the satellite broadband market. New generations of communications satellites will provide hundreds of Gbps of throughput by using advanced digital payloads, which will allow for beam-steering and beam-shaping, in addition to individual allocation of power and bandwidth for each beam. Therefore, dynamic resource management (DRM) techniques for communications satellites will be crucial for operators to fully exploit the capabilities of their satellites. This paper presents a new method for joint power and bandwidth allocation in multibeam satellite systems. To that end, we first develop a multibeam satellite model that accounts for propagation effects, interference among beams, and atmospheric attenuation. Next, we formulate the joint power and bandwidth allocation optimization problem and propose a novel algorithm to solve it. The basis of this algorithm is a genetic algorithm that is combined with repair functions to guarantee the validity of the solutions and speed up convergence. Finally, the usefulness of the algorithm is analyzed through two case studies: a notional case featuring a 37-beam satellite and a realistic case based on Viasat-l. The results obtained show that our joint power and bandwidth allocation algorithm can reduce the unmet system capacity (USC) by up to 40% (compared to just power allocation approaches). Furthermore, our experiments identify the variation of the demand among beams as a parameter that has a large impact on potential improvement: the higher the variation in demand among beams, the more beneficial it is to allow a greater flexibility in the range of bandwidth allocations allowed.