Abstract: A reconstruction method based on an improved multi-objective particle swarm optimization algorithm was proposed. This method reconstructed damaged constellations by combining multi-satellite launch and on-orbit satellite phase maneuver. Firstly, reconstruction indexes were selected, including coverage, reconstruction cost, reconstruction time and elasticity. Secondly, the multi-satellite launching process and on-orbit satellite phase maneuvering were analyzed, and a uniform phase strategy was adopted for remaining normal satellites in damaged constellations. In order to restore original constellation performance, considering the maximum revisit time, elasticity, reconstruction cost and time, optimization models for reconstruction time and reconstruction cost were established. Finally, a MOPSO algorithm was improved, and a population update strategy based on learning mechanism was proposed. Discrete variables were transformed into continuous variables through variable transformation, which solved mixed variable optimization problems in reconstruction optimization models. Simulation for a damaged constellation shows that the time-optimal reconstruction scheme is launching 6 new satellites combined with the uniform phase of on-orbit satellites, and the cost-optimal reconstruction scheme is launching 4 new satellites combined with the uniform phase of on-orbit satellites. The case shows that the proposed reconstruction method is effective and can provide a reference for construction of reconnaissance constellations.

Key words: multi-satellite launch, phase maneuver, constellation reconstruction, elastic constellation, multi-objective particle swarm optimization