Modeling of Nonlinear Deformation Considering Temperature and Hydrostatic pressure Using Genetic-Neural Networks
Combining genetic algorithms and artificial neural networks, a hybrid genetic-neural method was proposed for modeling the nonlinear dynamic deformation system considering the effect of environmental factors. This method describes the characteristic of nonlinear evolvement of deformation using ANN (the artificial neural network) whose structure (including nodes of input layer and hide layer) is automatically searched by GA (the genetic algorithm). The learning-samples and the testing-samples for training and testing the prediction function of ANN are made up of the input of ANN, which includes the temperature, the hydrostatic pressure and the time-series, while the desired output includes only the deformation. The ANN is trained by the learning-samples and is tested by the testing-samples. The practical example shows that the model constituted by this algorithm has more accurate predicting result and better predicting performance.
Bing-Rui Chen Xia-Ting Feng Cheng-Xiang Yang
Key Laboratory of Rock and Soil Mechanics, Institute of Rock and Soil Mechanics, the Chinese Academy School of Resources and Civil Engineering, Northeastern University, Shenyang 110004, China
国际会议
2009 WASE International Conference on Information Engineering(2009年国际信息工程会议)(ICIE 2009)
太原
英文
607-610
2009-07-10（万方平台首次上网日期，不代表论文的发表时间）