华中师范大学学报(自然科学版)
華中師範大學學報(自然科學版)
화중사범대학학보(자연과학판)
JOURNAL OF CENTRAL CHINA NORMAL UNIVERSITY
2007年
1期
51-54
,共4页
遗传算法%BP神经网络%时间序列%用电预测
遺傳算法%BP神經網絡%時間序列%用電預測
유전산법%BP신경망락%시간서렬%용전예측
Genetic algorithm%BP neural network%time series%power consumption prediction
结合遗传算法及神经网络各自的优点,利用改进遗传算法对BP神经网络的连接权进行优化,并提出了一种新的编码方式.通过与时间序列模型对比,基于改进遗传算法的BP模型效果更好.
結閤遺傳算法及神經網絡各自的優點,利用改進遺傳算法對BP神經網絡的連接權進行優化,併提齣瞭一種新的編碼方式.通過與時間序列模型對比,基于改進遺傳算法的BP模型效果更好.
결합유전산법급신경망락각자적우점,이용개진유전산법대BP신경망락적련접권진행우화,병제출료일충신적편마방식.통과여시간서렬모형대비,기우개진유전산법적BP모형효과경호.
This paper mainly studies the optimization of BP neural network by applying modified genetic algorithm, and proposes a combined GA-ANN algorithms by the merging of modified genetic algorithms (GA) and artificial neural network. (ANN). At the same time, a coding method is presented for genetic algorithms. The experimental result shows that the algorithm can improve network generation ability and accuracy for forecasting.