基于遗传算法的BP神经网络在水库月入库径流量预测中的应用

Application of BP Neural Network in the Prediction of Reservoir's Monthly Runoff based on Genetic Algorithm

水库的入库径流量是水库安全管理和水库群优化调度中的重要依据,由于水文系统是高度非线性的系统,具有模糊性、随机性等多种不确定性,因此水库入库径流量预测的难度较高。正确地建立预测模型可以提高水库运行中的效率以及安全性。本文建立基于遗传算法的反向传播BP神经网络用来进行水库月入库径流量预测,把影响水库月入库径流量的几个主要因素和水库月入库径流量的历史数值作为BP神经网络预测模型的输入变量和期望输出;利用遗传算法特有的全局搜索能力优化BP神经网络的初始权值和阈值;采用优化后的BP神经网络建立水库月入库径流量预测模型。实验结果表明,遗传算法结合BP网络的模型相较于单独的BP神经网络具有预测精度较高、收敛速度较快的优点。

Reservoir runoff is an important basis for reservoir safety management and optimal operation of reservoir group.Due to Hydrological systems are highly nonlinear system,fuzziness,randomness and other uncertainties,so the prediction of runoff is more difficult.The establishment of correct prediction model can improve the efficiency and safety in operation of the reservoir.To establish the BP（Back Propagation）neural network based on genetic algorithm is used for reservoir monthly runoff prediction.The historical data of the monthly runoff of the reservoir is used as the input variables and the expected output of the BP neural network prediction model.Using genetic algorithm to optimize the initial weights and thresholds of BP neural network with the characteristic of global searching ability,The reservoir runoff Prediction model is established by using the optimized BP neural network.Experimental results show that genetic algorithm combined with BP neural network has the advantages of higher accuracy and faster convergence speed compared with BP neural network.