改进的QGA-BP模型在弥苴河总氮量预测中的应用

水质预测对水环境规划、评价和管理十分重要。构建一种改进的量子遗传算法（QGA）优化BP神经网络的模型，即在量子遗传算法中引入了旋转角的动态改进策略和遗传算法的交叉变异操作，并以改进的QGA作为进化操作准则优化BP模型的权值和阈值。以弥苴河复杂水环境水质预测为实例，选取一组历史观测数据作为训练样本，对其进行分析。将结果与BP模型、QGA-BP模型仿真结果进行了对比，改进后的QGA-BP模型在进化代数、收敛速度和预测结果的准确率有较大提高。对弥苴河水质的预测结果表明，将改进QGA-BP模型用于水质预测是可行、有效的预测方法。

Application research on total nitrogen content forecast of River Miju with improved QGA-BP model

Water quality prediction is imperative to the planning, assessment, and management of the water environment. A model of a BP neural network optimized by an improved quantum genetic algorithm (QGA) was constructed. Further, a dynamically improving strategy of the rotating angle and a cross variation operation of the genetic algorithm were introduced into the QGA. The improved QGA was adopted as the standard of evolutionary operation to optimize the weight value and the threshold value of the BP model. Taking the water quality prediction of the complicated water environment of River Miju as an example, we analyzed a set of historic observation data selected as the training sample. The results were compared with the simulation results of the BP model and the QGA-BP model, and the improved QGA-BP model showed considerable improvement in terms of the accuracy rates of evolutionary algebra, convergence rate, and prediction result. The water quality prediction results for River Mi Ju suggested that the application of the improved QGA-BP model to water quality prediction is feasible and effective.