Research On Dam Deformation Forecasting System Based On Gene Expression Programming

For a long time, long-term deformation monitoring of dam and predict the deformation trend,is an important work to ensure the safe operation of the dam and prevent the occurrence of dambreaking. As traditional methods of dam deformation forecasting more or less have some shortages,studying new dam deformation forecasting method is imperative. Gene expression programmingalgorithm is a new generation of genetic algorithm with unique advantages, and studying itscharacteristics and applying it in deformation forecast, has significances to promote thedevelopment of dam deformation.This paper first studied from the basic theory of dam deformation, and states the two ways ofdam deformation forecasting, which is forecast based on impact factor and prediction based ontime series. Then, this paper analyzed the basic principle of classical gene expression programmingalgorithm and its shortages on application, and selected adaptive strategy to improve it. Then, thispaper studied Gene Space Balance Strategy, Dynamic Fitness Strategy and Adaptive Crossoverand Variation Probability Strategy based on cloud mode, and put forward the algorithm toimplement these adaptive strategies. Moreover, in order to make the adaptive gene expressionprogramming algorithm can be applied to deformation forecast better, this paper designed andimplemented the dam deformation prediction system based on the adaptive gene expressionprogramming according to the theory of software development and general process of predictivemodeling. Finally, this paper tested the performance of the adaptive gene expression programmingalgorithm and the function of the system by an application example.Viewed from the results of application example, the prediction accuracy of adaptive geneexpression programming algorithm is almost double that of classical gene expression programmingalgorithm, and the forecasting precision based on the impact factors are under5%. All of thisshowed that the overall performance of the improved gene expression programming algorithm isimproved. In addition, the dam deformation prediction system designed and implemented by thispaper that could predict according to the entire process, also made the process of dam deformationprediction become simpler and more flexible.