基于AAWDA-PP回归模型的边坡稳定性评价

Slope Stability Evaluation based on AAWDA-PP Regression Model

在边坡的稳定性评价中保证边坡分类的准确度十分重要.人工非洲野狗群体智能算法,能够通过模拟野狗的觅食行为来对目标函数进行寻优.结合投影寻踪算法以及阻滞增长曲线函数,建立人工非洲野狗优化投影寻踪模型,利用非洲野狗算法AAWDA优化投影指标函数及阻滞增长曲线函数参数,提高了模型求解的准确性,通过对求解后的结果建立回归模型,然后根据分级阈值对边坡等级进行分类,测试结果显示较好的精度.将模型应用于案例边坡的稳定性分析,并同PSO-PP模型,ANN模型所得结果进行比较分析,得出运用AAWDA-PP回归模型预测结果与经验值之间的误差最小,说明模型在研究边坡稳定性评价分级中更加准确有效.

Ensuring the accuracy of classification in slope stability evaluation is very important.Artificial African wild dog,swarm intelligence algorithm can simulate by the dragons of foraging behavior to optimize the objective function.Combining projection pursuit algorithm and logistic curve function,establishing the optimal projection pursuit model artificial African wild dog.We use the African wild dog（AAWDA） algorithm optimize the projection index function and logistic curve function（LCF）,improve the accuracy of the model,through the results of solving after establish regression model,and then classified according to classification threshold on the slope grade,the test results show that the better precision.Will be applied in the case of the slope stability analysis,the model with PSO-PP model,compares and analyses the results of ANN model using AAWDA-PP regression model,minimum error between the predicted results and experience shows that the model is more accurate and effective in the evaluation and grading of slope stability.