A niching cross-entropy method for multimodal satellite layout optimization design

Abstract

AbstractSatellite layout optimization design (SLOD) relies on solving a high-dimensional and multimodal optimization problem, in which there exist multiple global optimal solutions. Existing algorithms for SLOD focus on seeking only one approximate global optimum. However, finding multiple solutions simultaneously could provide more design diversity for the designers. To alleviate this problem, multimodal optimization method is studied for SLOD in this paper, and an improved niching-based cross-entropy method (INCE) is proposed. INCE consists of an improved niching strategy, cross-entropy method-based offspring generation and a cross operator. CEC2013 benchmarks and satellite layout optimization design problem are investigated to verify the validity and feasibility of the proposed INCE. Compared with several state-of-the-art algorithms, the proposed algorithm performs better.