期刊名称：Journal of Modelling and Simulation of Systems
印刷版ISSN：1737-9377
电子版ISSN：1737-9385
出版年度：2010
卷号：1
期号：1
出版社：HyperSciences Publisher
摘要：In this study, an efficient Monte Carlo based method utilizing adaptive neuro-fuzzy inference system (ANFIS) is introduced for reliability based optimization of structures. In addition the performance of the particle swarm optimization is investigated. Monte Carlo based methods are powerful tools, simp le to implement and capable of solving a broad range of reliability problems. Ho wever, the amount of computational efforts that may involve is the draw back of such methods. ANFIS is a fuzzy inference system implemented in the framework of adaptive neural networks; therefore, the proposed methodology makes use of the capability of it to approximate structural response for predicting probability of failure, allowing the computation of performance measures at a much lower cost. ANFIS were used in conjunction with PSO and MCS. This methodology made possible to get a solution for the optimum structural design with reliability constraint. A relaxation method for achieving the minimum number oftraining samples and epochs for training ANFIS is proposed. Comparison between back propagation neural network and ANFIS is performed. Also, for the reliability purpose, the performance of the particle swarm optimization method is studied in reliability based optimization of structures in the present work
关键词：Monte Carlo; Reliability; Optimization; Particle swarm; neural networks; Fuzzy systems
