摘要
分析遗传算法 (GA)及BP神经网络 (NN)的理论基础 ,提出了GA与NN结合的主要方式和步骤。讨论了基于遗传神经网络的建模思路 ,主要针对矿石可选性预测、选矿生产指标预报问题 ,建立相应的遗传神经网络模型。论述了选矿数据预处理的方法和GA -BP神经网络的设计。通过实例验证 ,模型的预测精度达到 90
Theoretical basis and integration methods of genetic algorithm (GA) and BP neural network (NN) are analyzed. Based on genetic neural network, the idea of building model is discussed. The genetic neural network models for prediction of ore beneficiability, flotation production targets are built. The methods of original data pre processing, the design of BP neural network, and some parameters selection are studied. Through verification of production examples, the prediction accuracy has achieved over 90%.
出处
《有色金属》
CSCD
2000年第4期27-32,共6页
Nonferrous Metals
基金
国家自然科学基金!(5 95 740 34)
湖南省自然科学基金
关键词
选矿
遗传算法
神经网络
预测模型
mineral processing
genetic algorithm
neural network
prediction model
