基于模拟退火优化BP神经网络的变形监测数据预测方法

Deformation Monitoring Data Prediction Method Based on SimulatedAnnealing Optimized BP Neural Network

针对BP神经网络模型用于变形监测数据处理时容易陷入局部极小值并且收敛速度慢的问题,提出一种基于模拟退火(Simulated Annealing,SA)算法优化BP神经网络的建筑物变形预测方法,利用SA的全局寻优能力对BP神经网络的模型参数进行优化,使参数迭代过程始终保持较高的"温度"和"能量",从而确保BP神经网络能够得到全局最优解的同时具备较高的预测精度和收敛速度。采用实际算例对所提SA-BP方法在低信噪比和小样本条件下的预测精度进行验证,结果表明所提方法相对于传统BP神经网络法和小波方法能够获得更高的预测精度,并且在小样本和低信噪比条件下优势更加明显。

When the BP neural network model is used for deformation monitoring,it is easy to fall into the local minimum and the convergence speed is slow.To solve this problem,a method for predicting building deformation based on Simulated Annealing(SA)algorithm to optimize the BP neural network is proposed.The global optimization ability of SA was used to optimize the model parameters of the BP neural network,and the parameter iteration process always kept high"temperature"and"energy",so as to ensure that the BP neural network could obtain the global optimal solution with high prediction accuracy and convergence speed.The actual examples were used to verify the prediction accuracy of the proposed SA-BP method under low signal-to-noise ratio(SNR)and small sample conditions.The results show that the proposed method can obtain higher prediction accuracy than the traditional BP neural network method and wavelet method.And the advantages are more obvious under the conditions of small samples and low SNR.