瓦斯涌出量是煤矿瓦斯灾害的主要来源,它直接影响煤矿安全生产和经济技术指标.瓦斯涌出量的传统预测方法是将其影响因素线性化后提出的,具有一定的局限性.本文基于群体智能理论,提出了一种基于粒子群算法优化的RBF神经网络瓦斯涌出量预测摸型.研究表明RBF神经网络预测精度与网络权值和RBF参数初始值有很大关系,因此本文采用粒子群算法优化RBF网络权值和其他参数,形成PSO-RBF预测模型.该模型通过计算种群粒子的适应度,确定全局最优值,寻找网络参数的最优值.实验结果表明PSO-RBF优于传统的RBF预测模型,训练速度和预测精度显著提高.%Gas emission was the major source of coal mine disaster, which affects the coal mine safety production and economic technical indicators. Traditional prediction methods had been based on the linear relationship between gas emission and other affect factors, and there were some limitations. Based on theories of swarm intelligence , a model of RBF network for gas emission prediction based on particle swarm optimization was proposed. The prediction accuracy of RBF neural network was concerned with the network weight and initial RBF parameters. So particle swarm optimization(PSO) was investigated for the network weight and initial RBF parameters, then a model of PSO-RBF was formed. The model could determine the global optimal value and find the optimal value of network parameters by calculate the swarm fitness value. The results showed that the model of PSO-RBF was better than traditional RBF network prediction model, the training speed and prediction accuracy increased significantly.
展开▼
