Optimization of regional coverage reconnaissance satellite constellation by NSGA-II algorithm

This paper aims at the optimizing design of regional coverage reconnaissance satellite constellation. Firstly, model of regional coverage satellite constellation is set up before the design is formulated into a multiple objectives optimization problem. Secondly, a nondominated sorting genetic algorithm (NSGA) and its improved version, elitist and fast nondominated genetic algorithm (NSGA-II), are introduced. With respect to characteristic of optimization of satellite constellation, an improvement is made to NSGA-II. Combined with dictionary ordered method, improved NSGA-II is applied to solve the satellite constellation optimization. The new algorithm is implemented by STK and visual C++. Simulation results show that the new algorithm has some advantage over the conventional genetic algorithm, being more flexible and avoiding selecting weight of multiple objectives.