Interplanetary Mission Design Using Differential Evolution

摘要

Global optimization methods have been increasingly under consideration for preliminary interplanetary mission design. One promising global method, differential evolution, has been identified as being particularly well suited to high-thrust trajectory optimization. Differential evolution is a stochastic direct search optimization method which uses parameter vectors that interact in a manner motivated by the evolution of living species. To improve the performance of differential evolution for this application, the effect of the tuning parameters is investigated over a diverse group of trajectory optimization problems. The quality of solutions obtained with differential evolution is found to be very sensitive to the selection of the routine's tuning parameters. A set of tuning parameter values is found that results in the rapid global optimization of an array of typical ballistic interplanetary missions. The fine-tuned differential evolution routine is implemented in a new tool, the mission-direct trajectory optimization program, and the effectiveness of this tool is demonstrated by the rapid solution of interplanetary trajectory optimization problems that involve complex features such as multiple gravity assists and parking orbit considerations.