摘要
针对目前应用人工神经网络(ANN)方法预测地下洞室围岩变形时间序列的缺陷,提出一种将神经元网络和模糊逻辑有机结合的新型模糊推理系统——ANFIS(adaptive-network-based fuzzy inference systems),该系统采用反向传播算法和最小二乘法的混合算法分别调整前提参数和结论参数,充分地利用了神经网络的学习能力和模糊逻辑的表达能力,实现了回归模型的自适应调整.通过龙滩电站的实例应用可以发现,ANFIS预测系统较传统ANN方法具有简单、快速以及预测精度高等特点.
In allusion to the insufficiency of the prediction accuracy of artificial neural network （ANN） algorithm for underground cavern rock surrounding stability, the method of ANFIS （adaptive-network-based fuzzy inference systems） machines is applied to researching into evolution law for a nonlinear deformation time series of rock surrounding. ANFIS is a new fuzzy inference system which combines neural network and fuzzy logic organically. The hybrid algorithm of back propagation algorithm and least square method is adopted to adjust the premise parameter and the consequent parameter respectively, which makes full use of the learning ability of neural network and the expression ability of fuzzy logic to realize the adaptive adjustment of the regression model. Compared with the conventional ANN method, Longtan Hydropower Station can be found simple, the application of ANFIS forecasting system in fast and accurate.
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
2009年第4期576-579,共4页
Journal of Dalian University of Technology
基金
岩土力学与工程国家重点实验室资助项目(Z110801)
