有色金属科学与工程2016，Vol.7Issue(5)：76-80,5.DOI:10.13264/j.cnki.ysjskx.2016.05.014
神经网络与遗传算法预测充填配比的研究
Filling ratio prediction with neural network and genetic algorithm
摘要
Abstract
On the basis of neural network and genetic algorithm, the optimal filling experimental conditions were predicted, which including: gray sand ratio 0.202 4, curing time 5.863 d, solubility 67.8 %, and the maximum compressive strength backfill 0.677 7 MPa. The predicted plan is quite different from the actual optimal matching program: gray sand ratio 1:4, curing time 28 d, solubility 75 %, and the maximum compressive strength 5.48 MPa. It shows that the forecast results are not exact and the prediction method can′t be applied in all conditions. The prediction accuracy of neural networks has effect on the optimization of extreme genetic algorithm. The suggestion of expanding sample size is put forward.关键词
神经网络/充填体/抗压强度/遗传算法/充填配比方案Key words
neural network/filling body/compressive strength/genetic algorithm/filling proportion plan分类
矿业与冶金引用本文复制引用
黄永刚,饶运章,刘剑,张学焱..神经网络与遗传算法预测充填配比的研究[J].有色金属科学与工程,2016,7(5):76-80,5.基金项目
2011年度江西省安全生产重大课题 （）
