多策略灰狼算法优化SVM的尾矿坝地下水位预测

为实现尾矿坝地下水位预测及预报预警,以尾矿坝地下水位为研究对象,将库水位、最小干滩长度等5个变量作为影响参数,尾矿坝地下水位为输出变量,建立支持向量机(SVM)预测模型.引入多策略改进的灰狼优化算法(MGWO)确定SVM模型参数,并以工程实例进行仿真预测.结果表明,MGWO?SVM模型平均相对误差0.24％,平均绝对误...

Prediction of Groundwater Level of Tailings Dam Based on MGWO-SVM

In order to achieve timely prediction and early warning of the groundwater level of the tailings dam, a support vector machine (SVM) prediction model was established for a tailings dam with five variables as the influencing parameters, including the reservoir water level, the minimum length of dry beach, among others, and the groundwater level of the tailings dam as the output variable. The multi-strategy gray wolf optimizer (MGWO) was introduced to determine the parameters of SVM model, which was then used for simulative prediction in the practical engineering. The results show that the MGWO-SVM model has an average relative error of 0.24%, and an average absolute error of 0.02 m. Compared with the traditional model, the MGWO-SVM model is characterized by higher accuracy and superior performance. It shows better applicability and reliability when adopted in the prediction of the groundwater level of the tailings dam.