基于蚁群优化算法的瓦斯预测模型研究  

Study on gas forecast model based on ant colony optimization algorithm

作　　者：杨桢[1] 李鑫[1] 付华[1] 王涛[1] 

机构地区：[1]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105

出　　处：《传感器与微系统》2009年第10期61-63,共3页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(50874059);辽宁省高校创新团队项基金资助项目(2007T071);辽宁省优秀人才基金资助项目(2007R24)

摘　　要：针对瓦斯煤尘爆炸和煤与瓦斯突出给煤炭矿山企业带来的危害极大的问题,将蚁群优化算法和BP神经网络技术结合应用到瓦斯涌出量预测,建立比较准确的预测模型。重点研究了BP网络模型的选择与优化训练,通过蚁群算法优化解决了BP神经网络易陷入局部收敛的问题。仿真与实际数据验证表明:改进的神经网络算法对瓦斯涌出量预测能达到良好的效果。In view of the harmful effects of gas and coal dust explosion and gas outburst in coat mine,the ant colony optimization algorithm and BP neural network technology are applied to the prediction of gas emission to establish more accurate prediction models. Focus on the BP network model selection and optimization of training, the ant colony optimization algorithm is used to solve the problem that the BP neural networks is easy to fall into local convergence. Simulation and actual data show that the improved neural network algorithm for the prediction of gas emission can achieve good results.

关 键 词：瓦斯预测 BP神经网络 蚁群优化 

分 类 号：TD713.2[矿业工程—矿井通风与安全]