基于MMAS BP算法的短期风速非线性组合预测模型

为提高风电场短期风速的预测精度，引入一种基于改进蚁群算法优化神经网络的非线性组合预测方法，按误差平方和最小原则对所建灰色GM(1,1)模型、BP网络和RBF网络三种单一预测数据进行非线性组合，并将其结果作为最终预测值。仿真结果表明，该方法的平均绝对误差及均方误差分别为17.76%和3.68%，均小于单一模型、线性组合模型及神经网络组合模型的预测结果，提高了网络的泛化能力，降低了预测风险，为风电场风速预测提供了一种新途径。

Nonlinear Combination Forecast Model for Short-term Wind Speed Based on MMAS BP Algorithm

In order to improve the forecast accuracy of short term wind speed in wind farm, a nonlinear combined forecasting method is proposed based improved ant colony algorithm to optimize neural network. In accordance with the principle of error sum of square minimization, the original forecasting data from grey GM(1,1), BP network and RBF network are combined as final forecasting value. The simulation results show that the mean absolute error and mean squared error of the proposed method are 17.76% and 3.68% respectively, which are less than the errors of single model, linear combination model and neural network combined model. The proposed method improves the generalization capability and reduces the prediction risk, which provides a new idea for wind speed prediction.