基于改进粒子群优化BP网络的城市用水量预测

对城市用水量的科学预测是城市供水管网规划与设计基础,可以给供水系统安排生产与优化调度提供科学依据。由于传统BP神经网络应用于城市用水量预测存在训练收敛速度过慢、预测精度较低等缺陷,本文提出基于改进粒子群优化BP神经网络的城市用水量预测方法。实验结果表明,该方法的训练收敛速度、预测精度明显优于传统BP神经网络、粒子群优化BP网络的方法,可以满足供水系统生产与调度的实际需要。

Urban Water Consumption Forecast Based on Improved Particle Swarm Optimization BP Neural Network

Scientific forecast of urban water consumption is the basis of urban water supply network planning and design, and provides a scientific basis for water supply production and optimal scheduling. Because the convergence of urban water consumption forecast based on BP neural network is slow and low forecast accuracy, the paper proposes a forecast method based on improved particle swarm optimization BP neural network. The experimental results show that both convergence speed and forecast accuracy of the proposed method are better than the method based on BP neural network and particle swarm optimization BP neural net- work, and meet the actual needs of the production and scheduling of water supply system.