摘要
为避免原始人工蜂群算法(原始ABC算法)搜索时陷入局部最优解,提出一种改进的人工蜂群算法(MABC算法),该方法先将原始蜜源的适应度进行排序,找出适应度最高的蜜源,再在其周围搜索更优解,并采用MABC算法对支持向量回归(SVR)模型参数进行优化,实现对边坡安全系数的回归分析与预测。通过对两种算法进行函数测试,结果表明:MABC算法较原始ABC算法收敛速度快、全局性好。选取实例边坡数据构造训练集和测试集,采用MABC-SVR方法基于建立的边坡安全系数预测模型进行预测,结果表明:均方根误差为0.004 6,最大相对误差为7.62%,回归系数为0.967 2。可见,建立的边坡安全系数预测模型准确度较高,可推广使用。
In order to avoid the original artificial bee colony algorithm (original ABC algorithm) searching for local optimal solution,this paper proposes a new modified artificial bee colony algorithm (MABC algorithm) which first sorts the fitness of the original honey source and finds out the most suitable honey source,then searches for better solutions around it,and uses the MABC algorithm to optimize the parameters of the support vector machine to achieve regression analysis and prediction of the slope safety factor.By testing the functions of the two algorithms,the results show that the MABC algorithm has faster convergence and better overall performance than the original ABC algorithm.The paper selects the instance slope data construction training set and test set,and uses the MABC-SVR method to make prediction based on the established slope safety coefficient prediction model.The results show that the root mean square error is 0.004 6,maximum relative error is 7.62% and regression coefficient is 0.967 2.It can be seen that the slope safety factor prediction model established in the paper has high accuracy,and the model has a good generalization.
作者
王芬
刘阳
郝建斌
魏兴梅
WANG Fen;LIU Yang;HAO Jianbin;WEI Xingmei(School of Geology Engineering and Geomatics,Chang’an University,Xi’an 710054,China)
出处
《安全与环境工程》
CAS
北大核心
2019年第2期178-183,189,共7页
Safety and Environmental Engineering
基金
国家自然科学基金项目(41472266)
中央高校基本科研业务费专项资金项目(310826172007
300102268209)
关键词
改进的人工蜂群算法
支持向量回归
边坡稳定性
安全系数预测
modified artificial bee colony algorithm (MABC algorithm)
Support Vector Regression(SVR)
slope stability
safety factor prediction
