遗传算法在中长期电量最优组合预测中的应用

Application of Genetic Algorithm in Optimal Combination Forecast of Medium-Long Term Load

为综合考虑多种电量影响因素,以提高中长期电量预测的精度,引入了最优组合模型,在多种预测模型组合中,以各预测模型的组合权重作为优化变量,由于优化组合模型约束条件所构成的解的空间具有非凸的特点,导致常规线性优化算法无法直接求解最优的组合权重,为此,采用了遗传算法计算各预测模型的组合权重,预测了某地区2010年各月的负荷电量。

To improve the accuracy of middle/long-term load forecast,the optimal combination model is introduced in this paper by taking various load influencing factors into account.This method takes the weights of the combined forecast models as optimal variable.The genetic algorithm has been used to calculate the combination weights,because the solution space formed by the constraints of optimal combination model is non-convex,and the general linear optimization algorithm fails to calculate the optimal combination weights.To validate the effect of the method,the monthly load in the year 2010 has been forecasted in one area.