Data Mining of Pareto-Optimal Solutions of a Solar Observatory Trajectory Design Problem
Akira Oyama, Yasuhiro Kawakatsu, Kazuko Hagiwara · Infotech@Aerospace 2012 · 2012
Multiobjective design exploration (MODE) framework is used to obtain useful knowledge from the Pareto-optimal solutions of a Solar observatory trajectory design problem. The trajectory design optimization problem is defined as a multiobjective design optimization problem and solved with a multiobjective evolutionary computation. Many designs that are better than the nominal design in all design objectives are found. The obtained Pareto-optimal solutions visualize tradeoff between the design objectives. Some data mining approaches are applied to the database of the obtained Pareto-optimal designs and show useful design information such as sensitivity of the design parameters to the design objectives. The present results shows MODE approach is useful for trajectory design optimization problems.