改进遗传神经网络在甘蔗产量预测中的应用  

The Application on the Sugar Cane Forecast with the Neural Networks Using Improved Genetic Algorithms

作　　者：徐永春[1,2] 张森文[3] 

机构地区：[1]华南农业大学工程学院,广东广州510642 [2]广东理工职业学院计算机系,广东广州510091 [3]暨南大学应用力学研究所,广东广州510632

出　　处：《华南农业大学学报》2010年第3期102-104,共3页Journal of South China Agricultural University

摘　　要：传统甘蔗产量预测方法对外界多因素关联作用的农作物产量预测的难度大、精度差、准确度低,本文提出改进自适应交叉和变异算子的遗传BP算法,多元逐步回归简化BP网络的输入变量,应用改进的遗传BP算法策略,并以甘蔗产量实例数据进行了验证和分析,结果表明,改进的遗传BP算法总体效果最优.The features of traditional method on sugar cane production forecast were discussed, and an improved intelligence genetic algorthms （GA） strategy and a method of simplified input variables by the multiple stepwise regression are presented, which can avoid the shortage of trained BP and enhance the BP model learning efficiency on the sugarcane data. An improved GABP model was simulated and verified in the forecast example of sugarcane data.

关 键 词：改进遗传算法 BP网络 甘蔗 预测 

分 类 号：TP306.1[自动化与计算机技术—计算机系统结构]