基于基因表达式编程的建筑物变形预测模型研究

发布时间：2018-04-02 18:31

本文选题：建筑物　切入点：变形预测　出处：《江西理工大学》2013年硕士论文

【摘要】：近年来，城市化的加快带动了房地产的发展，大量高层建筑随之产生。高层建筑物的安全问题日益受到人们的关注，建筑物变形预测问题就是其中之一。常见的预测方法对建筑物变形预测存在一些不足之处。基因表达式编程是一种基于生物基因结构和功能发明的一种新型自适应演化算法，已在金融、气象水文和灾害预警等预测领域广泛应用。因此，本文将基因表达式编程方法应用于建筑物变形预测中。首先，论文对课题研究的意义以及当前国内外变形预测研究进展进行了介绍，阐述了当前变形预测方法以及基因表达式编程应用的研究现状，确定了本文研究内容和研究路线。其次，根据基因表达式编程原理，对基因表达式编程的基本流程进行了分析，并根据问题的复杂度，，采用多个基因组成的染色体作为研究对象并对其进行了编码；其中个体的选择采用轮盘赌法；变异、插串、重组等算子通过多次实验得出了合理的参数值；同时，为了提高基因表达式编程的效率，针对传统方法的不足，使用了Kquick算法对K-表达式的求解过程进行了改进。然后，考虑到建筑物变形数据的实际情况，将滑动窗口预测法与标准的基因表达式编程算法加以结合提出了改进的基因表达式编程算法；使用Visual C++6.0开发了基于基因表达式编程的建筑物变形预测程序；最后，使用该程序对某公寓的沉降数据进行了预测，通过多次实验，获得最优的染色体，并得到了反映建筑物变形规律的数学模型。将预测结果与灰色系统预测模型、BP神经网络预测模型进行对比分析，结果表明，基于基因表达式编程的预测方法的预测值精度比传统方法提高近一倍，也就是说基于基因表达式编程的预测方法在建筑物变形预测领域具有推广应用价值。
[Abstract]:In recent years, the acceleration of urbanization has led to the development of real estate, and a large number of high-rise buildings have come into being.People pay more and more attention to the safety of high-rise buildings, one of which is building deformation prediction.There are some shortcomings in building deformation prediction by common prediction methods.Gene expression programming is a new adaptive evolutionary algorithm based on biological gene structure and function. It has been widely used in the fields of finance, meteorology and hydrology, disaster warning and so on.Therefore, the genetic expression programming method is applied to the prediction of building deformation.Firstly, this paper introduces the significance of the research and the research progress of deformation prediction at home and abroad, expounds the current research status of deformation prediction methods and gene expression programming, and determines the research content and research route of this paper.Secondly, according to the principle of gene expression programming, the basic process of gene expression programming is analyzed, and according to the complexity of the problem, the chromosome composed of multiple genes is used as the research object and encoded.The selection of individuals is based on roulette method; mutation, string insertion, recombination and other operators obtain reasonable parameter values through many experiments; at the same time, in order to improve the efficiency of gene expression programming, aiming at the shortcomings of traditional methods,The Kquick algorithm is used to improve the solution process of K-expression.Then, considering the actual situation of building deformation data, an improved genetic expression programming algorithm is proposed by combining the sliding window prediction method with the standard genetic expression programming algorithm.The program of building deformation prediction based on genetic expression was developed by using Visual C 6.0. Finally, the settlement data of an apartment was predicted by the program, and the optimal chromosome was obtained through many experiments.A mathematical model reflecting the law of building deformation is obtained.The prediction results are compared with the grey system prediction model and the BP neural network prediction model. The results show that the prediction accuracy of the prediction method based on gene expression programming is nearly twice as high as that of the traditional method.In other words, the prediction method based on genetic expression programming has the value of popularization and application in the field of building deformation prediction.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU196.1

【参考文献】