改进的QGA-BP模型在弥苴河总氮量预测中的应用

发布时间：2018-03-23 15:30

  本文选题：量子遗传算法　切入点：BP神经网络　出处：《环境工程学报》2016年11期

【摘要】：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。
[Abstract]:Water quality prediction is very important for water environment planning, evaluation and management. An improved Quantum genetic algorithm (QGA) is proposed to optimize BP neural network. That is, the dynamic improvement strategy of rotation angle and the crossover mutation operation of genetic algorithm are introduced in quantum genetic algorithm. The weight and threshold of BP model are optimized by using the improved QGA as the evolutionary operating criterion, and a set of historical observation data is selected as the training sample, taking the water quality prediction of the complex water environment of the Miju River as an example. The results are compared with the simulation results of BP model and QGA-BP model. The improved QGA-BP model is improved greatly in evolutionary algebra, convergence rate and accuracy of prediction results. It is feasible and effective to apply the improved QGA-BP model to water quality prediction.
【作者单位】： 昆明理工大学信息工程与自动化学院;大理州洱海流域保护局;
【基金】：云南省科技厅科技惠民项目(2014RA051);云南省科技厅面上项目(2013FZ010)
【分类号】：X832;X52
，

本文编号：1654062