基于加速混合遗传算法搜索复杂边坡的最危险滑动面

Searching for the most dangerous failure surface of complex slope based on accelerating hybrid genetic algorithm

基于简化Bishop条分法分析复杂边界边坡的稳定性,以圆弧滑裂面与边坡面的左右交点的两个X坐标——XL和XR以及滑裂面圆心坐标的Y坐标——Y0作为设计变量,提出一种加速混合遗传算法(AHGA)对设计变量进行优化。该方法是通过把无约束优化的Powell算法作为一个与遗传算法的选择、交叉和变异平行的算子,嵌入到改进的实数编码遗传算法中而得到的,它同时具有两个方法的优点,有较强的自适应能力,使得收敛速度加快,还有可能搜索到最优化问题的全局最优解。通过一高速公路边坡工程实例验证,加速混合遗传算法(AHGA)搜索到的圆弧滑动面与现场勘察结果很吻合,优于简单遗传算法(SGA)。

Based on Bishop Slices Method to analyze the stability of complex slope, X coordinates of two points where the circular failure surface intersected the slope and the Y coordinate of the centre of circular were chosen to be the variables. One of accelerating hybrid genetic algorithm （AHGA） was presented to optimize the variables. AHGA was established by setting Powell Method in real-code GA, which was taken as a genetic operator that parallels to the operators of selection, crossover and mutation. The advantages of Powell Method and GA were considered in the new method, such as better adaptive ability, rather high convergence speed. Moreover, this kind of method can get the more optimal and precise solution than SGA.