摘要
考虑水质、经济和生态环境影响等因素,建立佛山水道的引水规划优化模型。利用河网水环境数学模型模拟多组引水冲污方案的水质,将输入输出数据作为样本用于人工训练神经网络;将训练好的网络嵌入遗传算法,形成混合智能算法,求解引水规划优化模型。结果表明,混合智能算法能够自动求出不同引水流量下的最优方案,精度较高,无需人工试算,运算速度快,不必对遗传算法与河网模型进行接口处理,具有普遍适用性,为求解耦合复杂模拟模型的优化问题提供了一种理想的工具。
Taking into account factors including water quality, economy and enviromnental impact, an optimal water diversion programming model was developed for the Foshan Channel. A river network water environment simulation model was utilized to predict the water quality of several water diversion schemes, and the input/output data were used as samples for training an artificial neural network (ANN). A hybrid intelligent algorithm (HIA) coupling a genetic algorithm (GA) with the trained ANN was employed to solve the optimal water diversion p^g model. The results showed that the HIA can automatically fred the optimal water diversion scheme for different quantifies of diverted water with high accuracy and without trial periods. Furthermore, the HIA ran quickly and did not need to couple the GA with the river network simulation model, which is a universal algorithm and provides an ideal tool for solving optimal problems linked with a complicated numerical simulation model.
出处
《水资源保护》
CAS
2009年第4期31-36,共6页
Water Resources Protection
基金
广东省“科技计划百项工程”资助项目(4202112)
关键词
人工神经网络
遗传算法
引水冲污
优化
感潮河网
artificial neural network
genetic algorithm
diverting water to flush out pollutants
optimization
tidal fiver networks