基于基因表达式编程的大坝变形预测系统的研究

发布时间：2018-04-23 00:37

本文选题：大坝变形 + 基因表达式编程　；参考：《江西理工大学》2014年硕士论文

【摘要】：一直以来，对大坝进行长期的变形监测并预测其变形趋势，是一项确保大坝安全运行，防止溃坝发生的重要工作。传统的大坝变形预测方法或多或少都存在着一些不足，所以研究新的大坝变形预测方法十分必要。基因表达编程算法是新一代独具优势的遗传算法，，研究其特性并应用于大坝变形预测中，对促进大坝变形预测的发展具有重要意义。本文首先从大坝变形基本理论出发，阐述了大坝变形预测的两种途径，即基于影响因子的预测和基于时间序列的预测。其次，本文分析了经典基因表达式编程算法的基本原理和其在实际应用中的不足，并选择从自适应的角度对其进行改进。再次，本文分别研究了基于基因均匀分布的初始种群策略、动态适应度策略和云模型调整的变异交叉概率策略，并给出了这些改进策略的算法实现。然后，为使改进的基因表达式编程算法更好地应用于大坝变形预测，本文依据软件开发理论和预测建模的一般流程，从系统的分析设计和系统的开发实现两方面，建立了基于基因表达式编程的大坝变形预测系统。最后，本文通过一个应用实例对改进的基因表达式编程算法的性能和系统的功能进行了测试。从应用实例的结果看，改进的基因表达式编程算法的预测精度比经典基因表达式编程算法高出一倍左右，且其基于影响因子的预测精度均在5%以下，这说明了改进后的基因表达式编程算法的整体性能得到了提高。此外，使用本文实现的系统进行大坝的变形预测，可以依流程执行变形预测的整个过程，也使大坝变形预测的过程变得更加简单灵活。
[Abstract]:For a long time, it is an important task to monitor the dam deformation and forecast its deformation trend to ensure the safe operation of the dam and prevent the dam break. Traditional dam deformation prediction methods have some shortcomings, so it is necessary to study new dam deformation prediction methods. Gene expression programming algorithm is a new generation of genetic algorithms with unique advantages. Studying its characteristics and applying it to dam deformation prediction is of great significance to promote the development of dam deformation prediction. In this paper, based on the basic theory of dam deformation, two kinds of methods of dam deformation prediction, namely, the prediction based on influence factors and the prediction based on time series, are expounded in this paper. Secondly, this paper analyzes the basic principle of classical gene expression programming algorithm and its shortcomings in practical application, and chooses to improve it from the point of view of adaptation. Thirdly, the initial population strategy, dynamic fitness strategy and mutation crossover probability strategy based on uniform gene distribution are studied, and the algorithm implementation of these improved strategies is given. Then, in order to better apply the improved genetic expression programming algorithm to dam deformation prediction, according to the software development theory and the general process of prediction modeling, this paper analyzes and implements the system from two aspects: the analysis and design of the system and the development and implementation of the system. A dam deformation prediction system based on gene expression programming is established. Finally, the performance of the improved genetic expression programming algorithm and the function of the system are tested by an application example. The prediction accuracy of the improved genetic expression programming algorithm is about twice as high as that of the classical gene expression programming algorithm, and the prediction accuracy based on the influence factor is less than 5%. This shows that the overall performance of the improved genetic expression programming algorithm has been improved. In addition, the whole process of dam deformation prediction can be carried out by using the system realized in this paper, and the process of dam deformation prediction becomes more simple and flexible.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV698.11;TP18

【参考文献】