A Spacecraft Equipment Layout Optimization Method for Diverse and Competitive Design

The spacecraft equipment layout optimization design (SELOD) problems with complicated performance constraints and diversity are studied in this paper. The previous literature uses the gradient-based algorithm to obtain optimized non-overlap layout schemes from randomly initialized cases effectively. However, these local optimal solutions are too difficult to jump out of their current relative geometry relationships, significantly limiting their further improvement in performance indicators. Therefore, considering the geometric diversity of layout schemes is put forward to alleviate this limitation. First, similarity measures, including modified cosine similarity and gaussian kernel function similarity, are introduced into the layout optimization process. Then the optimization produces a set of feasible layout candidates with the most remarkable difference in geometric distribution and the most representative schemes are sampled. Finally, these feasible geometric solutions are used as initial solutions to optimize the physical performance indicators of the spacecraft, and diversified layout schemes of spacecraft equipment are generated for the engineering practice. The validity and effectiveness of the proposed methodology are demonstrated by two SELOD applications.