关键词: 灰色预测模型, DNGM（1, 1）模型, 参数优化

Abstract: The traditional gray GM（1, 1） model is established for the approximation of homogeneous exponent sequence. In order to expand the applied range of gray forecasting models, the DNGM（1, 1） model is present for approximation non-homogeneous exponent sequence. The properties of this model are discussed, and it is proved that the DNGM（1, 1） model not only has coincidence of white non-homogeneous exponent law, but also has uniformity of linear transformation. The forecast accuracy of this model may be affected by the initial value conditions, the improvement DNGM（1, 1） model is obtained by parameters optimal. The data simulation and application example show that this model has higher precision than the traditional GM（1, 1） model.

Key words: gray forecast model, DNGM（1, 1） model, parameters optimal