An investigation of the applicability of genetic algorithms to spacecraft trajectory optimization

This dissertation deals with the use of genetic algorithms to determine optimal spacecraft trajectories. Genetic algorithms (GAs) are optimization algorithms that are based on principles of biological evolution and adaptation. The goal of this study is to determine what role genetic algorithms can play in the field of spacecraft trajectory optimization and to explore and develop ways to customize genetic algorithms to that task. The theory behind genetic algorithms is presented followed by a discussion of two spacecraft trajectory optimization problems solved using genetic algorithms. This study represents the first time genetic algorithms have been applied to these problems.; A genetic algorithm begins with a population of randomly generated candidate solutions to an optimization problem and attempts to improve these solutions through the use of operators that imitate biological processes. Unlike traditional calculus-based optimization methods, genetic algorithms are able to efficiently search through many local optima to find the global optimum. There are some drawbacks to the use of GAs, the primary one being the number of performance index evaluations required during a typical GA run. However, the ability to sort through many possible solutions and localize a search to the region near the global optimum offsets most drawbacks.; The first trajectory optimization problem studied is that of launching a spacecraft from the lunar surface to a specified altitude/velocity combination in minimum time. This problem is first solved using a calculus-based optimization algorithm and the results are compared to results from a genetic algorithm. A new hybrid scheme that uses the strengths of the GA and the calculus-based optimization algorithm is shown to be the best approach to this problem.; The second trajectory optimization problem involves finding initial conditions for Earth-Moon transfer trajectories generated using a three-body model that incorporates Birkhoff's global regularization. This problem is very sensitive to the initial conditions used to generate the trajectories. A genetic algorithm is developed that successfully handles the sensitivity problem and is used to generate a trajectory that is comparable to the Apollo Earth-Moon transfer trajectories.