阶梯状黄土边坡稳定性分析的改进遗传算法及应用  

An Improved Genetic Algorithm and Application of the Stability Analysis for Ladder-Like Loess Slope

作　　者：龙熙华[1] 杨玲[1] 叶万军[2] 

机构地区：[1]西安科技大学计算机科学与技术学院,陕西西安710054 [2]西安科技大学建筑与土木工程学院,陕西西安710054

出　　处：《数学的实践与认识》2014年第12期143-150,共8页Mathematics in Practice and Theory

基　　金：荷载-湿度-温度耦合效应下黄土崩塌灾害形成的机理研究(41172262);切坡卸荷下黄土崩塌响应的机理研究(KJXX-37);教育部新世纪优秀人才支持计划项目(NCET2012-1044)

摘　　要：阶梯状黄土边坡稳定性分析的关键是估算其稳定系数的最小值.稳定系数的求解涉及诸多因素且计算过程繁杂,传统优化算法往往不能有效地搜索到其全局最小解.为此,提出一种改进的自适应遗传算法.算法对基因变量空间进行网格状划分,采用迭代选优法建立均匀分布的初始种群,运用优质个体保留遗传策略,并按照特定的准则自适应地调整交叉概率和变异概率,提高算法的全局搜索能力和收敛速度.实例应用表明算法能够快速有效地收敛于土坡稳定系数的全局最小解,且计算结果与实际情况更加吻合.The key of analyzing the ladder-like loess slope stability is to calculating its mini- mum stability coefficient. The stability coefficient involves many factors, and its computation process is more complicated, so classical optimization algorithms will not achieve the global minimum solution effectively. In view of this, we propose an improved method of adaptive Genetic Algorithm （GA）. This algorithm makes the gene variable space generate the grid, and uses the iterative selection methods to build the initial population with uniform distribution. Moreover, the algorithm employs the elitism strategy and adaptively adjusts the crossover and mutation probability according to the given rules, as well as improves the global searching ca- pability and convergence rate of GA.The application instance indicates that the algorithm can converge to the global minimum solution of ladder-like slope stability coefficient more quickly and efficiently, and its computation results are exactly consistent with actual facts.

关 键 词：阶梯状黄土边坡 稳定性分析 遗传算法 初始种群 全局收敛 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]