摘要
针对煤矿回采工作面瓦斯涌出的非线性特征,提出一种基于改进量子粒子群优化BP神经网络(IQPSO-BP)的瓦斯涌出量预测方法。鉴于量子粒子群算法的遍历能力有限,采用混沌序列来初始化量子的初始角位置。同时,采用凸函数调整惯性权重,以平衡算法的全局勘探和局部开发能力。并依此来优化BP神经网络的权值、阈值参数,进而建立了瓦斯涌出量预测模型。试验结果表明,IQPSO-BP算法具有较强的泛化能力及较高的预测精度,可有效用于煤矿瓦斯涌出量的预测。
Aiming at the nonlinear characteristics of gas emission in a coal mine working face,a method based on the improved quantum particle swarm optimized BP neural network（ IQPSO-BP） was proposed. In view of the limited ability to the traverse of the quantum particle swarm,chaotic sequences were used to initialize the initial angle position of particles. At the same time,the convex function was used to adjust the inertia weight and balance the global exploration and local development ability. Based on this,the weight and threshold parameters of BP neural network were optimized,and then the gas emission prediction model was established. The results showed that the IQPSO-BP algorithm had better generalization ability and higher prediction accuracy,and can be effectively used for the prediction of gas emission in a coal mine.
出处
《矿业安全与环保》
北大核心
2016年第4期38-41,共4页
Mining Safety & Environmental Protection
基金
云南省教育厅科学研究基金项目(2012Y450)
关键词
瓦斯涌出量
预测
改进量子粒子群优化算法
BP神经网络
gas emission
prediction
improved quantum particle swarm optimization algorithm（IQPSO）
BP neural network
