基于遗传算法和BP神经网络的矿井涌水量预测

Prediction of Mine Inflow Based on Genetic Algorithm and BP Neural Network

为避免BP神经网络极易陷入局部解的问题,针对遗传算法具有全局寻优的特点,提出了用遗传算法优化BP神经网络预测方法,并以刘桥二矿为例,对其矿井涌水量进行了预测。首先选取刘桥二矿区的2005年3月至2006年12月的矿井涌水量数据进行分析,然后使用遗传算法优化BP网络的初始权值和阈值,最后使用BP神经网络进行训练。将其成果与纯BP网络算法进行比较,结果表明:遗传算法优化BP神经网络的预测方法的预测精度高于纯BP网络算法,将其应用于矿井涌水量预测是有效可行的。

To avoid the BP neural network common problem of trapped into a local solution and take advantage of the genetic algorithm＇s global optimization, a BP network optimized by genetic algorithm approach has proposed to predict mine inflow. To take the Liuqiao No.2 coalmine as an example, analyzed mine inflow during March 2005 to December 2006. Firstly, the initialized weights and thresholds of BP neural network were optimized with genetic algorithm, and the network was trained by modifying the weights and thresholds. Finally, this method was compared with pure BP network algorithm. The results show that the prediction accuracy of this method is higher than that of pure BP network algorithm and it is feasible and effective to apply BP network optimized by genetic algorithm approach to predict mine inflow.