A Genetic Algorithm with Observation Window Refinement for Earth Observation Scheduling

Agile earth observation satellites (AEOS) possess flexible three-axis attitude maneuvering capabilities, which significantly enlarge the visible time windows (VTW) of ground point targets and consequently increase the complexity of observation mission scheduling. To address this challenge, this paper proposes a genetic algorithm based on observation window refinement (OWR-GA), aiming to maximize the total observation profit while satisfying both time window constraints and attitude maneuvering constraints. The proposed algorithm integrates a greedy initialization strategy and an observation window refinement mechanism, moreover, to prevent the loss of high quality solutions, a Gaussian mutation with variance annealing is adopted. Simulation results demonstrate that the proposed method achieves a higher total observation profit compared with other three algorithms.