基于遗传神经网络模型的水质综合评价

建立了用于水质综合评价的遗传神经网络模型.该模型运用遗传算法优化改进型BP神经网络的初始权值和阈值,具有快速学习网络权重和全局搜索的能力,有效解决了BP神经网络容易陷入局部极小点和训练结果不稳定的问题.采用苏帕河梯级电站的水质监测数据对该模型进行了测试,并与其他方法进行了比较.结果表明,该方法用于水质综合评价客观、合理、准确,有其独特的优越性.

Comprehensive Assessment of Water Quality Based on Genetic Neural Network Model

A model based on genetic neural network was set up for comprehensive assessment of water quality. Genetic algorithms were used to optimize the initial weights and threshold of the improved BP neural network in simulation. The simulated results indicate that genetic algorithms has the capability of fast learning the weight of network and globally searching, and the training speed of the improved BP network is greatly raised. As a case, the model was applied to analyze the water quality for cascade hydropower stations in Supa River Basin, and the calculated values were compared with other methods. The results show that the model is objective, reasonable and accurate in comprehensive assessment of water quality with its unique advantages.