混合遗传神经网络在边坡稳定性评价中应用研究.pdfVIP

中国农村水利水电# 2006 年第7 期 75 : 1007-2284( 2006) 07-0075- 02 1 2 杨 蕾 , 林 红 ( 1. ( ) , 257061; 2. ( ) , 257061) : 将改进的遗传算法与BP 神经网络相结合, 分析了边坡稳定性的影响因 , 建立边坡稳定性评价的神经网 络模型分别用改进的遗传算法和自适应BP 算法对网络权值进行全局优化和局部二次优化选用边坡实例对网络进 行训练检验, 结果表明, 混合遗传神经网络提高了训练速度和泛化能力, 对边坡稳定性分析有较好的适用性, 为边坡稳定 性评价提供了依据 : 混合遗传神经网络; 边坡稳定性; 遗传算法; 自适应BP 算法; 全局优化 :TU432; T P183 :A Study on the Application of Mixed Genetic-Neural Network in SlopeStability Evaluation 1 2 YANGLei , LIN ong ( 1. College of Math a d Computatio Scie ce of Chi a Oil U iversity ( Easter Chi a Bra ch) , Do gyi g 257061, Sha do g Provi ce; 2. College of Mecha ic a d Electric E gi eeri g of Chi a Oil U iversity (Easter Chi a Bra ch) , Do gyi g 257061, Sha do g Provi ce) Abstract: Combi i g modif ied Ge etic Algorithm a d BP Neural Netw ork, the i flue ci g factors for slope stabilit y are a alyzed, a d a eural et work model for the evaluatio of slope stability is established. T he modified ge etic algorithm a d self-adaptive BP algorithm are applied i the global optimizat io a d local seco dary optimizatio . Selectio of slope is used as a example for the trai i g a d verif icatio . The results show that the mixed ge etic- eural etw ork speeds up the process of trai i g a d popularity ca- pacity , suits for the slope stabilit y a alysis, a d provides support for the evaluatio of slope stability . Key words: m ixed ge etic- eural etw ork; slope stability; ge etic algorithm; self-adaptive BP algorithm; global optimizatio [ 2] , 0