Adaptive kriging-assisted optimization of low-thrust many-revolution transfers to geostationary Earth orbit

To effectively optimize low-thrust many-revolution transfer trajectories to geostationary Earth orbit (GEO), this article proposes a novel metamodel-based low-thrust GEO transfer optimization scheme. A simplified control law is used to convert the optimal low-thrust transfer problem into a parameter optimization problem, where the gains of control law are optimized to determine the time-minimum trajectories. An adaptive kriging-assisted two-stage optimization framework is developed to solve the optimization problem. In the first stage, the kriging metamodels are constructed to replace the expensive transfer model for optimization. The kriging metamodels are gradually refined via a probability of constrained improvement-based infill sampling process to efficiently determine an initial guess of the gains. In the second stage, a sequential quadratic programming-based local search is conducted to precisely compute the gains. Finally, two engineering examples are investigated to demonstrate the effectiveness of the proposed optimization scheme for solving real-world low-thrust GEO transfer optimization problems.