In this thesis, a software application for optimizing impulsive interplanetary trajectories is developed and used to design trajectories for hypothetical missions to two near-Earth asteroids. The optimal control problems for impulsive trajectories are...
In this thesis, a software application for optimizing impulsive interplanetary trajectories is developed and used to design trajectories for hypothetical missions to two near-Earth asteroids. The optimal control problems for impulsive trajectories are transcribed into nonlinear programming problems, which can then be solved by a global optimization algorithm. The developed software can optimize four different types of impulsive trajectories, from a simple two-impulse trajectory to difficult trajectories composed of multiple gravity assists and deep space maneuvers. As a global optimization algorithm, monotonic basin hopping in the software is chosen because the algorithm has been claimed to be effective in solving trajectory optimization problems by a few studies. The algorithm is powerful enough in its basic form, but it is further modified mainly in three ways in the hopes of improving its performance. The first modification allows the algorithm to take advantage of parallel computing, resulting in two versions of parallel monotonic basin hopping. The second one is concerned with a novel method of automatically choosing an appropriate perturbation size, which is an important factor affecting the performance of the algorithm. As the final refinement, the algorithm is hybridized with another global optimization algorithm called multi-start. Before using the developed software on designing trajectories for the hypothetical asteroid missions, it is tested on a well-known interplanetary trajectory problem called the reduced Messenger problem to see if the modifications lead to an actual improvement. It is found that parallelizing the algorithm can significantly improve its performance in a multi-core environment and that the other two modifications can help the algorithm perform better in some situations. Finally, one-way and round-trip missions to two near-Earth asteroids are designed using the software. The procedure for finding the optimal or near-optimal trajectory for each mission is discussed in detail and the best trajectory found is presented. During the designing procedure, the optimization performance of the developed algorithm is compared with other well-known metaheuristics on several different problems, which turns out to be better than all other metaheuristic algorithms compared. These exemplary simulations of designing trajectories for fictional missions demonstrate that the developed software can readily be used in preliminary trajectory design of interplanetary missions that employ impulsive maneuvers.