摘要
采用改进神经网络,利用遗传算法初步修正网络权重,再用LM算法优化权重,加快算法收敛速度,通过对新疆玛纳斯灌区用水量进行预测,结果表明,算法有效,预测精度较高,具有较好的实际应用价值。
Irrigation water due to the influence of various factors in certain uncertainty, the forecast of irrigation water is conducive to the ra- tional use of water resources. With the improved neural network, this paper preliminary network weights by genetic algorithm with LM algo- rithm to optimize weights, to speed up the algorithm convergence speed, through to the Xinjiang manas irrigation area water consumption forecasting, the results show that the algorithm is effective, prediction accuracy is higher, has good practical application value.
出处
《中国农机化学报》
北大核心
2014年第2期75-77,共3页
Journal of Chinese Agricultural Mechanization
基金
新疆兵团青年创新基金项目(2011CB023)
石河子大学自然科学技术项目(ZRKYBLH12)
关键词
灌溉用水
不确定性
改进神经网络
网络权重
irrigation water
uncertainty
improved neural network
the network weights
