遗传神经网络模型在地下水质量评价中的应用

设计了用遗传算法训练神经网络权重的新方法,并应用该方法建立了地下水质量评价的GA-BP神经网络模型,对阜新新邱露天煤矿地下水水质进行了评价。研究结果表明遗传算法具有快速学习网络权重和全局搜索的能力,有效地解决了BP算法的局部收敛的问题。应用该方法环境质量评价简便可靠,预测精度高,具有通用性和客观性等优点。

Application of GA-BP Model to Appraise Groundwater Quality

A new method which trained the neural weight with hereditary algorithm has been designed. By this method, GA-BP neural network model of groundwater quality appraise has been set up, and the groundwater quality of Xinqiu open-pit coal mine in Fuxin has been appraised. The research results shown that genetic algorithm can study fast the neural weight,searches the overall situation, and effectively solves the local convergence of BP algorithm. This method applied to appraise groundwater quality is simply, conveniently and reliably, and have advantages such as the high accuracy of prediction and objectivity, etc.