Based Gene Expression Programming Building Deformation Prediction Model

In recent years, the acceleration of urbanization led to the development of real estate, theresulting large number of high-rise buildings. The high-rise building security has attractedincreasing attention, building deformation prediction is one of them. There are someshortcomings of common forecasting method of building deformation prediction. GeneExpression Programming is based on biological gene structure and function of the invention, anovel adaptive evolutionary algorithm, which widely used in the prediction field of financial,meteorological, hydrological and disaster warning. Therefore, this method is applied to geneexpression programming building deformation prediction.First of all, the thesis summarizes the research significance of subject, as well as the currentdomestic and international deformation prediction research progress. Describes the currentresearch situation of deformation prediction method and gene expression programmingapplication. Determine the research content and the research route of this paper.Secondly, based on gene expression programming principle, analysed the basic flow of GEP,according to the complexity of the problem, used multiple genes of the chromosome as theresearch object and encoding it; And individual choice takes roulette method; Variation, insertstring, re-operator through several experiments obtained reasonable parameter values; Meanwhile,in order to improve the efficiency of GEP, use the Kquick algorithm to improve solution processof K-expression.Then, taking into account the actual situation of the building deformation data, combine thesliding window prediction method with standard GEP algorithm to present novel GEP algorithm;developed building deformation prediction based on gene expression programming proceduresusing the VC6.0; Finally, use the program on an apartment settlement data were predicted bymany experiments, the optimal chromosome, and has been a mathematical model reflects thebuilding deformation law. The prediction results were compared with prediction model of greysystem forecasting model and BP neural network, the results show that the predictive value ofaccuracy of the prediction method based on gene expression programming to nearly double thantraditional methods, That is to say prediction method based on gene expression programmingdeformation prediction has popularization and application value in the building.