摘要
阐述了运用粒子群优化人工神经网络建立煤层底板突水预测模型的思路与方法,利用粒子群优化神经网络模型的权值和阈值,克服了神经网络容易收敛到局部最小值,以及收敛速度慢的缺点。实践表明:该方法不仅能更快地收敛于最优解,且预测精度有明显的提高。
Introduces the theory and method of application of artificial neural network trained by particle swarm optimization in establishing forecasting model of the coal floor water bursting. PSO algorithm is applied to optimize weights of BP neural network , it overcomes both slow convergence speed and local optimization of the neural network. Finally an example was used to verify the proposed methodology can speedily converge to the optional solution and had better generalization performance.
出处
《山西焦煤科技》
2009年第1期34-36,42,共4页
Shanxi Coking Coal Science & Technology
关键词
煤层底板突水
粒子群优化算法
神经网络
预测
Coal floor water bursting
Particle Swarm Optimization
Neural Network
Forecast