点击次数：

论文类型：期刊论文

发表时间：2017-09-01

发表刊物：COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING

收录刊物：SCIE、EI、Scopus

卷号：324

页面范围：689-711

ISSN号：0045-7825

关键字：Non-probabilistic reliability-based design optimization; Efficient adaptive-loop method; Improved single loop method; Chaos control; Complex engineering problem

摘要：For non-probabilistic reliability-based design optimization (NRBDO), existing methods cannot satisfy both efficiency and robustness requirements, especially for complex engineering problems. In this study, an efficient adaptive-loop (EAL) method for NRBDO is proposed based on an improved single loop method and enhanced chaos control method, aiming to enhance the optimization efficiency without sacrifice of numerical stability. The constraint function is reconstructed to improve the efficiency of reliability analysis. Moreover, a criteria condition is defined to guarantee the rationality of rough optimum and eliminate redundant constraints. The proposed EAL method can adaptively switch to a proper method, thus both high efficiency and robustness can be guaranteed. Nonlinear numerical examples and complex engineering problems are utilized to demonstrate the effectiveness of the proposed method by comparison with other existing methods. (C) 2017 Elsevier B.V. All rights reserved.