一种基于区间灰数几何特征的灰数预测模型

Prediction model of interval grey number based on its geometrical characteristics

灰色系统预测模型的已有研究主要建立在实数的基础之上,没有考虑灰数信息对预测结果的影响,根据区间灰数序列在二维坐标平面上的几何特征,通过"面积转换"与"坐标转换",将区间灰数序列转换成等信息量的实数序列,然后推导并建立区间灰数的预测模型,通过建立多组随机区间灰数序列的预测模型,与其他灰数预测模型的模拟精度进行了比较.模型适用于具有不确定信息特征的区间灰数序列的模拟及预测.

The existent researches of grey system prediction model were mainly based on real numbers, which did not consider the effect of grey number information on prediction results. According to the geometric characteristics of interval grey number sequence on the two-dimensional coordinate plane, an interval grey number sequence is converted into real number sequences without any information loss through the conversion of area and ＂conversion of coordinates＂. And then, a simulation-prediction formula of grey number lower/upper limit is established. Through building prediction models of multi-group stochastic interval grey number sequences, this paper compared simulation errors obtained using the present model and others. The novelpresent model is suitable for the simulation and the prediction of the sequence of interval grey number with the characteristic of uncertain information.