DocumentCode :
624620
Title :
Application of optimization model based on neural network in Softening Slope Stability by Strong Rainfall Infiltration
Author :
Zhigang Yang ; Dong Zhang ; Biao Deng ; Weimin Chen
Author_Institution :
Sch. of Archit. Eng., Nanchang Univ., Nanchang, China
fYear :
2013
fDate :
9-11 June 2013
Firstpage :
289
Lastpage :
292
Abstract :
Through analyzing main affecting factors of the artificial neural network model, the optimization model is established and the optimization parameters is obtain based on the method of momentum and self-adaptive of learn rate, This optimize artificial neural network is not only set up with the limited training samples, but also can improve the operating speed and study efficiency. The optimization mode of the of flood prediction of Slope Stability is build by Bedrock Softening under the Condition of Strong Rainfall Infiltration, it shows that its precision is high, and its computation is simple. The method by optimization neural network, which applied to bedrock strength decreases forecasting under the Condition of Strong Rainfall Infiltration, provides a new attempt for Prediction analysis and prove to be feasible and effective for practical experience in complex system engineering of bedrock Slope Stability.
Keywords :
floods; forecasting theory; geophysics computing; geotechnical engineering; learning (artificial intelligence); mechanical stability; neural nets; optimisation; rain; structural engineering computing; artificial neural network model; bedrock slope stability softening; bedrock strength; complex system engineering; flood prediction analysis; learn rate momentum method; optimization model; optimization parameters; self-adaptive learn rate; strong rainfall infiltration; Analytical models; Artificial neural networks; Biological neural networks; Rocks; Safety; Stability analysis; Training;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Control and Information Processing (ICICIP), 2013 Fourth International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4673-6248-1
Type :
conf
DOI :
10.1109/ICICIP.2013.6568084
Filename :
6568084
Link To Document :
بازگشت