The Research On Dam Deformation Forecasting Model Based On Gene Expression Programming

In order to guarantee the safe operation of the dam, need basing on existing observationdata to predict deformation of dam in the future. But there are a lot of methods for predictionof dam deformation now. The traditional methods to predict exists problems of complexprocess for prediction, slow speed, and low prediction precision. Gene expressionprogramming is a new evolutionary algorithm is developed basing on GA and GP. It widelyused in field of disaster warning. So it is necessary to carry out the research on damdeformation prediction based on GEP.This paper expounds the dam deformation forecasting and gene expression programmingresearch status at home and abroad firstly, and introduces dam deformation monitoring theoryand prediction model; secondly, according to the principle of gene expression programmingand algorithm process, to determine the model building including the function set andterminators, population initialization, chromosome decoding, fitness evaluation, geneticoperation, such as process, under Visual Studio platform,using C#programming language tocomplete the building of the model; thirdly, design program of Fibonacci weighted slidingwindow method to preprocess dam monitoring data by using C#language; finally,establishing the grey GM (1,1) model and BP neural network model by using MATLABsoftware, using the prediction model based on gene expression programming, gray GM (1,1)forecasting model and BP neural network model make a prediction for a dam deformationrespectively, and analysis the prediction results of three kinds prediction models.Through calculation, the three models in the horizontal displacement of the averagerelative error is1.43%,3.85%and3.08%respectively; in the vertical displacement of theaverage relative error was1.87%,5.54%and4.83%respectively. The experimental resultsshow that the gene expression programming of dam deformation forecasting model accuracyis relatively high, it is better than grey GM (1,1) model and the prediction precision of BPneural network model. It can be seen that the dam deformation forecasting model based ongene expression programming provides a new method for dam deformation prediction.