基于GEP的露天矿台阶爆破块度预测模型  

Prediction Model of Bench Blasting-fragmentation by GEP

作　　者：聂军[1] 史秀志[1] 陈新[1] 史采星[1] 

机构地区：[1]中南大学资源与安全工程学院,长沙410083

出　　处：《爆破》2015年第2期82-88,共7页Blasting

基　　金：国家科技支撑计划项目(2013BAB02B05)

摘　　要：针对爆破块度难以预测的问题,选用基因表达式编程(GEP)算法,以MyEclipse为开发工具,建立基于GEP的爆破块度预测模型。选取一组实测数据进行预测,并与库兹列佐夫法(Kuznetsov),多重回归分析法(MVRA),人工神经网络法(ANN)预测的结果进行对比。通过比较,GEP模型预测结果的相关系数最高,平均绝对误差、平方根误差最低,分别是0.957、0.008、0.078。预测结果表明GEP模型在爆破块度预测方面是可行的,为爆破块度预测提供了一种新的方法。In order to predict the fragmentation of blasting, the prediction model of blasting fragmentation was es- tablished based on GEP with Eclipse as a development tool to simulating the date. Compared with other three results by Kuznetsov method, multiple regression analysis method and the artificial neural network method,the predicted re- sults of GEP gave the highest correlation coefficient, and the minimum mean absolute error and the minimum square root error were also obtained. The prediction results show that the GEP model is feasible in blasting fragmentation prediction.

关 键 词：台阶爆破 爆破块度 GEP 预测 

分 类 号：TD854[矿业工程—金属矿开采]