摘要
以大隆矿矿井为例,针对BP算法存在极易陷入局部最优与过拟合导致网络泛化能力不足问题,采用灰色神经网络理论与遗传算法相结合,在构建GNNM（1,1）基础上引入GA加以优化,预测分析了2001~2003年煤矿涌水量,构建了煤矿涌水量预测模型,并进行了Matlab仿真。
BP algorithm has shortcomings of trapping into local optima easily and over-fitting leading to poor generalization capacity of network.Taking Dalong Mine for an example,combination of gray neural network and genetic algorithm（GA） is applied to establish GNNM（1,1） model with GA for parameter optimization.And then the GNNM（1,1） model is used to predict water yield of coal mine from 2001 to 2003 years.Finally,the results are simulated with Matlab software.
出处
《水电能源科学》
北大核心
2011年第2期25-27,共3页
Water Resources and Power
基金
黑龙江省教育厅振兴老工业基地基金资助项目(1152gzd08)
关键词
GNNM(1
1)
GA
优化
煤矿
涌水量
预测
gray neural network model
genetic algorithm
optimization
coal mine
water discharge
prediction
