Multiobjective Global Optimization of Space Mission Design Using Evolutionary Methods and Interval Analysis

In this paper, the possibility of combining evolutionary methods and interval analysis, in order to deal with the multiobjective optimization of the preliminary phase of a space mission design, is investigated. Such combination is expected to give to the resulting optimization algorithm more flexibility in the choice of the optimization level and an effective ability in handling optimization processes in uncertain environments. An algorithm based on the Fast Evolutionary Programming has been implemented. The algorithm has been then adapted in order to allow the innovative application of Fast Evolutionary Programming to multiobjective optimization problems. Moreover, the algorithm structure has been modified in order to describe either the search space or the objective functions parameters by using interval information. Applications of the resulting algorithm to the optimization of both space system design and mission analysis are presented.