基于蚁群神经网络的煤矿瓦斯含量预测模型  

Predicting Model of Gas Content Based on Ant Colony Neural Network

作　　者：王鹤[1] 邵良杉[2] 邱云飞[3] 

机构地区：[1]辽宁工程技术大学基础教学部,辽宁葫芦岛125105 [2]辽宁工程技术大学系统工程研究所,辽宁葫芦岛125105 [3]辽宁工程技术大学软件学院,辽宁葫芦岛125105

出　　处：《微计算机信息》2011年第5期197-198,226,共3页Control & Automation

摘　　要：为了准确而快速预测煤矿煤层的瓦斯含量,采用了蚁群神经网络算法进行预测,建立了针对预测瓦斯含量的蚁群神经网络模型。该模型是将蚁群算法引入到神经网络的优化训练中,以避免神经网络算法容易陷入局部最优的不足,从而获得稳定的网络结构,再用训练好的神经网络对对象进行预测。以辽宁三家子矿区煤层样本为例,应用蚁群神经网络算法进行仿真实验,实验结果表明,该算法与传统算法相比,具有较高的预测精度和较快的运算速度,是一种十分有效的瓦斯含量预测方法。In order to accurately and quickly predict coal mine gas content,a neural network model based on ant colony algorithm is proposed.In this model an ant colony algorithm is introduced into neural network optimization training,which could avoid falling into local optimum.After that,a stable network structure is got.Then the trained neural network is used to predict the test samples.Based on t he primary mineable coal bed in the San Jia Zi of Liao Ning the network model is used to predict coal mine gas content,by experimental analysis,the results show that the algorithm is a more effective and accurate compared with the traditional algorithms.

关 键 词：瓦斯含量 蚁群优化算法 BP神经网络 预测 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]