摘要
基于BP神经网络,定义了衡量网络输入对输出作用大小的相对作用强度RSE(RelativeStrengthofEffect),并结合实际的岩石工程实例数据用RSE分析了其各个作用参数对工程稳定性的影响的相对大小与作用方式。实际分析的结果表明,所提出的方法能够较全面地反映岩石工程现场的复杂实际情况,具有易于处理不确定性、动态与非线性问题等优点,是对岩石工程进行参数分析的有效工具。
The RSE (Relative Strength of Effect) is defined with backpropagation neuralnetwork. With the RSE, the various roles of relevant factors are analysed according topractical rock ensineering data, and the result has proved the efficiency of RSE. The newmethod is suitable for non-linear, dynamic, and uncertain problem, and is a powerful toolfor the factor analysis of stability of underground opening.
出处
《岩石力学与工程学报》
EI
CAS
CSCD
北大核心
1998年第3期336-340,共5页
Chinese Journal of Rock Mechanics and Engineering
关键词
岩石工程
人工神经网络
稳定性
控制参数
rock engineering, artificial neural networks, effect factors