摘要
提出将BP神经网络与遗传算法(GA)相结合,建立了三层GA-BP神经网络模型,模型利用遗传算法来修改网络的连接权值,构筑进化型的神经网络模型,缩短网络学习训练时间,提高模型预测精度。炉温预测主要是高炉铁水硅质量分数的预测,当要求硅质量分数预报的绝对误差为±0.05%时,命中率为90%。结果表明,GA-BP网络模型比传统的BP网络模型能够获得更高的精度。
Three-layer GA-BP model was built by combined GA with BP. The weight value of network was modified by the model that use of GA to build anagenesis model.The time of train and study was shorten.So the forecast precision of model was improved. The forecast temperature of the blast furnace is mainly to forecast the content of Si in the molten iron. When the error of the Si content is 0.05%, the percentage of hits is 90%. The result shows that： GA-BP model can gain higher precision than traditional BP network model.
出处
《宁波职业技术学院学报》
2008年第5期73-77,共5页
Journal of Ningbo Polytechnic
