A Unified Resource Allocation Framework and Impact Evaluation for NGSO Satellite Constellations

Abstract

The new era of satellite communications will rely on thousands of highly flexible spacecraft capable of autonomously managing constellation resources, such as power or frequency. Previous work has focused on the automation of the individual tasks that compose the resource allocation problem (RAP). However, two aspects remain unaddressed: (1) A unified method that autonomously solves the RAP under nongeosynchronous conditions is still to be developed, and (2) the cost–benefit of using optimization methods remains to be studied. Note that these studies are critical for satellite operators to take appropriate decisions concerning the automation of communications constellations operations. To close this gap, this work proposes an adaptive framework to solve the RAP for high-dimensional nongeosynchronous satellite constellations. The proposed framework uses a divide-and-conquer approach that solves each step of the RAP, leveraging different optimization algorithms at the subproblem level to produce a valid and efficient allocation of resources over long time horizons. When comparing the proposed method against scalable greedy solutions, the former achieves up to four times more constellation capacity and reduces the overall consumed power by up to a factor of 3. The cost–benefit analysis reveals which RAP subproblems should be prioritized depending on the operator's objectives. Studying diverse operational conditions, we find that optimization methods enhance capacity consistently yet might raise power consumption due to trade-offs in the routing algorithms.