Assessing the performance of a heuristic simulated annealing algorithm for the design of distributed satellite systems

A heuristic multidisciplinary design optimization technique has been investigated for applicability to the conceptual design of distributed satellite systems. Different permutations of a simulated algorithm are applied to the problem of developing a systems architecture that minimizes a multicriterion metric (cost per image) for the NASA Origins Terrestrial Planet Finder mission. It was found that the efficiency of the simulated annealing algorithm could be improved by modifying the degrees of freedom within the algorithm. Further, a methodology has been developed for finding the Pareto-optimal set of design solutions within the system trade space when the distributed satellite system is modeled as a multi-objective design problem. These advances combine to yield new tools for systems engineers exploring the trade space of distributed satellite systems during the conceptual design stage.