马尔科夫链优化Logistic曲线的基坑沉降预测模型

Study on the Foundation Pit Settlement Prediction Model of Logistic Curve Optimized by Markov Chain

针对基坑沉降具有非线性及随机性的特点,建立基于马尔科夫链修正的Logistic曲线预测模型进行基坑沉降预测。基坑监测工程实例应用表明:利用Matlab平台的nlinfit函数进行Logistic曲线参数估计是有效的,将Logistic模型拟合值与观测值的残差用于马尔科夫的状态划分,构造状态转移概率矩阵,建立马尔科夫链优化的Logistic模型,预测均方根误差和平均绝对百分误差都比单一Logistic模型小。这表明该方法用于基坑沉降预测是可行的。

Aiming at the nonlinear and stochastic characteristics of the foundation pit settlement,a Logistic curve prediction model based on Markov chain correction is established to improve the prediction accuracy of the model.According to the foundation pit deformation monitoring,it is effective to estimate the Logistic parameters by using the nlinfit function of Matlab platform.The Logistic model fitting value and residuals of observations are used for the state division of Markov process and the construction of the state transition probability matrix,so as to establish the Logistic model optimized by Markov.The root mean square error and the average absolute percentage error of the Logistic model optimized by Markov are smaller than the single Logistic model,which shows that the method can be used for foundation pit settlement prediction and it is feasible.