摘要
在分析城市用水特点、筛选相关影响因素的基础上建立城市生活需水量预测模型,并研究了模型求解过程中智能算法的应用。采用改进的粒子群优化(PSO)算法对反向传播(BP)神经网络的初始设置进行智能优化,避免了传统BP神经网络模型在训练过程中容易陷入局部极小值的缺点。应用该粒子群优化神经网络(PSO-BP)算法求解需水量预测模型,其实例结果表明,该算法提高了神经网络的训练效率,基于该算法的预测模型具有较理想的可靠性和精度。
According to the characteristics of urban comprehensive domestic water demand, a water demand prediction model was established based on the analysis of related factors. The application of intelligent algorithm was studied to solve the prediction model. Used individually, back propagation （BP） neural network algorithm needs random setting of the initial weight and threshold. Falling into a local minimum point is easy. To overcome this defect, the modified particle swarm optimization （PSO） algorithm was used to optimize the initial setting of the BP neural network algorithm. The PSO-BP algorithm was utilized to solve the water demand prediction model. In the case study, the PSO-BP algorithm improved the training efficiency of the neural network, and the water demand prediction model based on the algorithm showed reliability and accuracy.
出处
《中国给水排水》
CAS
CSCD
北大核心
2012年第21期66-68,共3页
China Water & Wastewater
基金
国家自然科学基金资助项目(71203158)
教育部人文社科基金资助项目(12YJC630248)
关键词
综合生活需水量
智能算法
粒子群优化
神经网络
comprehensive domestic water demand
intelligent algorithm
particle swarm optimization
neural network
