基于新鲜度函数的油气产量组合预测方法

Combination forecasting petroleum production based on function of fresh degree

在总结前人产量预测方法的基础上,提出了基于新鲜度函数的多模型组合预测油气产量的方法。该方法是单模型预测技术的自然推广。同时,将新鲜度函数引入组合预测,有效地刻划了时间序列的变化趋势,对传统的优化组合预测和模糊自适应变权重组合预测方法都有不同程度的改进。改进后的组合预测方法不仅综合利用了各种预测方法的有用信息,而且该最佳组合预测的权系数也是时变的,该方法具有预测精度高、计算简便的特点。

The numerous existing methods for prediction of oil and gas production were reviewed, and a multi-model combined fore cast method based on fresh degree function was developed. This method was derived from the single-model prediction technique. The fresh degree function is brought into the multi-model combined forecast method, which can effectively describe the varying trend of the data list and improve the self-adaptive recurrent algorithm and fuzzy variable weight combined forecast method to some degree. So, the improved model can not only adopt useful data of different methods but also make the weight coefficient of combined forecast being time-varying. This method has the features of high precision and high accuracy of prediction, good adaptability and simplicity of calculation.