温特斯模型参数并行求解方法与应用

针对能源生产中压缩空气日产量具有明显周期性和规律变化的特性,提出了一种基于霍尔特-温特斯模型的参数并行计算预测方法。从时间序列分析、指数平滑以及最优平滑系数的并行求解入手,建立了压缩空气日产量预测模型。实验结果表明该预测模型在提高运算速度的同时具有较高预测精度。

Research of Winters Prediction Model Based on Parallel Computing

According to the obvious periodical and characteristics of the rule changes of energy production in the compressed air daily output, this paper gives out a predicting method for holt-winters model based on parameters solve using parallel computing, building the compressed air daily output prediction model from the time series analysis, exponential smoothing and optimal smooth coefficient of parallel solution. The result shows that the speed of computing and the accuracy of prediction was greatly improved by the use of winters prediction based on parameters solve using parallel computing.