Evolutionary Algorithms in Astronautic Applications

Kurzfassung

Evolutionary algorithms (EA) are a computation tool that utilizes biological principles found in the evolution theory. One major difference to other optimization methods is the fact that a large group of solutions is evaluated, not a single one. Combination of various solutions from such a group, called population, allows improvement of the solutions. Overall several terms in usage in the field of evolutionary algorithms have their origin in genetics or biology, especially the three major function principles of EAs: Selection, recombination and mutation. Intrinsic to evolutionary algorithms is also the fitness function, which provides a numerical quality evaluation of a solution within the population of solutions and thus sets the probability of this solution’s reproduction. Generally a fitness function is a function to be optimized by EAs. One major advantage of EAs in this respect is their ability to shift from one possible optimum to another and thus they are not bound to local optimization but can find global optima. Regarding astronautic applications, evolutionary algorithms have been used for optimization of trajectories of low-thrust engines and impulsive engines. Various other fields apply evolutionary algorithms for optimization, e.g. aerodynamics or warehouse planning. This survey will concentrate on the usage of evolutionary algorithms for space applications, especially trajectory optimization and will try to describe future developments as currently planned and also to determine valuable areas of research.