Improved heuristic auction-initialized large neighborhood search for online scheduling of Earth observation satellites in large-scale emergency mission

Abstract

Real-time and reliable online scheduling for large-scale Earth observation mission using multiple agile satellites remains challenging, due to stringent response-time and complex orbit-attitude coupling dynamics during observation maneuvers. To address this problem, this paper proposes a novel improved heuristic auction-initialized large neighborhood search (IHALNS) algorithm framework. First, a precise attitude-pointing maneuver model based on iterative correction strategy is developed to ensure the reliability of planned schedules. Then, a mission planning model is established to maximize observation profit while minimizing maneuver energy consumption. Subsequently, the IHALNS algorithm framework is proposed for real-time online scheduling. Within IHALNS, an improved heuristic auction (IHA) algorithm module with multiple heuristic rules and a self-decision mechanism is developed to generate a high-quality initial scheduling scheme in real time. Furthermore, a large neighborhood search incorporating density-guided destroy and spatiotemporal coupling repair operators is designed to rapidly achieve global optimization based on the initial scheme. Finally, numerical simulations demonstrate that the proposed IHALNS algorithm framework outperforms conventional methods in terms of computational efficiency, solution quality, scalability and stability, showing significant advantages in real-time online scheduling.