The Constrained Niching Differential Evolution Algorithm for Satellite Layout Optimization Design
Abstract: The satellite layout optimization design (SLOD) problem involves various coupling constraints and maintains multiple optimal solutions, which is a typical NP-hard problem. To solve the SLOD problem efficiently and robustly, CNDE-LS-SE, a constrained niching differential evolution algorithm, is proposed with three special efforts. First, based on employing the feasibility rule as a constraint handling technique, an information-guided strengthening evolution (IGSE) mechanism is proposed to further enhance the constraint-handling capability. IGSE is designed to perform multi-generation evolutions on inferior individuals without extra fitness evaluations, which can overcome the severe exploration stagnation issue due to a small portion of discrete feasible regions in the compact layout problem. Second, in order to maximally improve convergence speed and solution quality, an information-guided local search (IGLS) strategy is designed to select promising superior individuals for local optimization by comprehensively taking various information into consideration. Last but not least, to incorporate the multimodal property, an advanced niching method named NBC-minsize, is fulfilled to improve the efficiency of exploration and exploitation by limiting the minimum size of sub-populations and balancing the species. By comparison with several state-of-the-art algorithms, the superiority of our proposed method on efficiency, efficacy, and robustness is demonstrated via two simplified satellite layout design cases.
Loading