TY - GEN
T1 - Reliability Analysis Using Artificial Neural Network Based Adaptive Parameter Differential Evolution Algorithm
AU - Bui, N. T.
AU - Nguyen, T. T.
AU - Nguyen, V. T.
AU - Tao, N. L.
AU - Hasegawa, H.
N1 - Funding Information:
This work has been supported by the Research Grant Start-up Budget of Shibaura Institute of Technology, Grant No. 216403
Publisher Copyright:
© 2020 ACM.
PY - 2020/6/14
Y1 - 2020/6/14
N2 - Reliability analysis is one of the methods to consider the safety and stability of an engineering system. It is very important to determine whether a system is safe or not. We need to solve the complex nonlinear and implicit the limit state functions to obtain the reliability index. Traditional reliability analysis methods, First-Order Reliability Method (FORM), Second-Order Reliability Method (SORM), and Monte Carlo simulation (MCS), are not effective and have many limitations. In this paper, at the first step, an artificial neural network was used to model the limit state function. After that, the elite opposition-based learning differential evolution algorithm was selected to solve nonlinear equality constrained optimization problem to find the reliability index and the failure probability of problems in terms of random variables. The proposed method and some reference methods were applied to analyze the test problems in the literature to compare their effectiveness.
AB - Reliability analysis is one of the methods to consider the safety and stability of an engineering system. It is very important to determine whether a system is safe or not. We need to solve the complex nonlinear and implicit the limit state functions to obtain the reliability index. Traditional reliability analysis methods, First-Order Reliability Method (FORM), Second-Order Reliability Method (SORM), and Monte Carlo simulation (MCS), are not effective and have many limitations. In this paper, at the first step, an artificial neural network was used to model the limit state function. After that, the elite opposition-based learning differential evolution algorithm was selected to solve nonlinear equality constrained optimization problem to find the reliability index and the failure probability of problems in terms of random variables. The proposed method and some reference methods were applied to analyze the test problems in the literature to compare their effectiveness.
KW - Reliability analysis
KW - differential evolution
KW - global search
KW - neural network
KW - optimization algorithm
UR - http://www.scopus.com/inward/record.url?scp=85089727175&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85089727175&partnerID=8YFLogxK
U2 - 10.1145/3402597.3402614
DO - 10.1145/3402597.3402614
M3 - Conference contribution
AN - SCOPUS:85089727175
T3 - ACM International Conference Proceeding Series
SP - 88
EP - 93
BT - Proceedings of the 2020 3rd International Conference on Robot Systems and Applications, ICRSA 2020
PB - Association for Computing Machinery
T2 - 3rd International Conference on Robot Systems and Applications, ICRSA 2020
Y2 - 14 June 2020 through 16 June 2020
ER -