Analysis of multiple asteroids rendezvous optimization using genetic algorithms

The optimization of a multiple asteroids rendezvous trajectory is a mixed integer nonlinear programming problem, and is hard to solve due to the combination of multiple local minima and its extraordinary sensitivity to discrete variables. This study tries to solve it using a mixed-code genetic algorithm (GA), a variation of that GA with enhancing the continuous variable search for the best solution in each generation, and a two-level GA. These algorithms are tested by solving three cases with four, eight, and sixteen asteroids to visit respectively. The results show that the mixed-code GA with search enhancement presents the best performance and the two-level GA presents the worst performance. The treatment by enhancing the continuous variable search for the best solution in each generation has improved the performance of the algorithm considerably.