基于IEA-PNN的边坡岩体稳定性预测研究  

RESEARCH ON FORECASTING OF ROCK SLOPE STABILITY BASED ON IEA-PNN

作　　者：熊建秋[1] 李祚泳[2] 

机构地区：[1]四川大学水电学院,网川成都610065 [2]成都信息工程学院,四川成都610041

出　　处：《岩石力学与工程学报》2005年第A01期4924-4928,共5页Chinese Journal of Rock Mechanics and Engineering

基　　金：国家重点基础研究发展规划（973）项目（2002CB412301）;国家自然科学基金资助项目（40271024）

摘　　要：概率神经网络是一种训练速度快、结构简洁明了、应用广泛的人工神经网络，该方法采用贝叶斯分类决策理论建立系统的数学模型，以高斯函数作为激励函数，具有非线性处理和抗干扰能力强等特点。阐述了概率神经网络的基本结构及其训练算法，提出了基于概率神经网络的边坡岩体稳定性预测方法，并采用一种新的有效随机全局优化技术——免疫进化算法对高斯型函数的标准偏差进行了参数优化。介绍了免疫进化算法的设计思想和特点，并成功地实现了此模型在边坡岩体稳定性预测中的应用，实例预测结果与边坡稳定性实际状态完全一致。理论分析和实例结果验证了基于免疫进化算法的边坡岩体稳定性预测方法切实可行，且具有需要学习样本少、预测精度高、非线性动态数据处理能力强等优点，为边坡稳定性预测提供了一条新的途径。Probabilistic neural network （PNN） model is a kind of artificial neural network, which is simple in structure, easy for training and widely being used. The method uses Bayes classifying and decision-making theory to constitute the mathematic model of system; with Gauss function as activating one, it possesses the characteristics of strong nonlinear processing and anti-interfering ability. The theory and algorithm of PNN are expatiated, and then the application of PNN to rock slope stability forecasting is proposed. Immune evolutionary algorithm （IEA） that is an efficient random global optimization technique is used to optimize the parameter of Gauss function. The design idea and characteristics of IEA-PNN are introduced, and it is successful to apply this model to the rock slope stability forecasting. The results of case study show that the analysis results are completely consistent with the actual situation. It is shown that the IEA-PNN method is feasible in practice, it needs less learning sample, having more prediction-precision, stronger performance of dealing with non-linear dynamic data and better performance of non-linear system modeling than other artificial neural network methods and at the same time it provides a new approach for slope stability forecasting.

关 键 词：岩土力学 边坡岩体稳定性 预测 概率神经网络 免疫进化算法 

分 类 号：TU45[建筑科学—岩土工程]