【摘要】 设计了用遗传算法训练神经网络权重的新方法 ,实验结果显示了遗传算法快速学习网络权重和全局搜索的能力 ,有效地解决了BP算法的局部收敛问题。误差反向传播的遗传———神经网络 (GA—BP)模型用于大气环境质量综合评价 ,具有简便、准确、客观和适应性强等优点更多还原

【Abstract】 A new method is designed for training connection weight of NN by using Genetics Algorithm The experiment indicates that Genetic Algorithm has the capability o f fast learning the weight of NN and escaping local optima T he GA-BP Model which based on error back propagation has advantages of simple , precise, objective and practicable for the assessment of atmospheric enviro nmental quality更多还原