Efficient Methods for Agile Earth Observation Satellite Scheduling

摘要

Agile Earth observation satellite (AEOS) scheduling is a complex optimization problem with longer visible time and Time-dependent transition time constraint. We address one method to transform the AEOS scheduling problem to Maximum weight independent set (MWIS) problem to reduce the difficulty of modeling and solving. We also propose reduction and decomposition strategy to reduce the scale of the problem. Experiments proved IP model established by these methods can get optimal solution of the problem, whose size is larger than previous research. Additional, iterated local search hybrid with Variable neighborhood search (VNS-ILS) is designed to solve MWIS problem from AEOS scheduling. The proposed algorithm almost increases all solution quality than ALNS and ALNS-TS adopting in previous research for AEOS scheduling.