论文部分内容阅读
针对需水量预测的非线性、随机性和模糊性等特点,引入模糊系统理论,建立了基于T-S模型的模糊神经网络需水量预测模型。该模型将模糊系统良好的模糊知识表达能力与神经网络强大的自学习和自适应能力有机结合,具有逼近最优、收敛速度快、训练时间短等优点。应用该模型预测了天津市2015年的需水量。结果表明,采用基于T-S模型的模糊神经网络方法进行需水量预测的拟合与预测平均相对误差分别为3.39%和2.67%。将该模型与BP神经网络和非线性回归方法的预测结果进行对比分析,该模型的拟合与预测精度最高。
According to the nonlinearity, randomness and fuzziness of water demand prediction, the fuzzy system theory is introduced to establish the water demand prediction model of fuzzy neural network based on T-S model. This model combines the good fuzzy knowledge expression ability of fuzzy system with the powerful self-learning and adaptive ability of neural network. It has the advantages of approximation, convergence speed and short training time. The model is used to predict the water demand in Tianjin in 2015. The results show that the average relative error of fitting and forecast of water demand prediction based on the fuzzy neural network based on T-S model are 3.39% and 2.67% respectively. By comparing the model with the prediction results of BP neural network and nonlinear regression, the fitting and prediction accuracy of the model is the highest.