一种改进的灰色关联模型验证方法研究

Study on Validation of Simulation Models Based on Improved Gray Relational Analysis

在灰色关联算法优化的研究中,为了降低常规灰色关联模型验证方法存在的风险,提出了一种改进的灰色关联模型验证方法。在常规灰色关联分析的基础上,改进方法考虑了曲线相似性和数值相近性两个属性。使用差因子综合了相似性属性,同时通过辨识升降两种变化趋势提高了灰色关联分析的精度;使用商因子进行参考序列和行为序列的数值比较,综合了数值相近性的属性,克服了常规灰色关联分析主要依据序列曲线几何形态的相似程度进行关联判断的缺陷,提高了灰色关联分析的可靠性。给出了改进灰色关联模型验证方法的步骤。最后,通过实例分析与比较验证了改进灰色关联度模型的有效性和实用性。

In order to reduce the risk in application of GRA to validation of simulation models,a new validation method of simulation models based on gray relational analysis is proposed,which is in view of the similarity by subtraction and the nearness by division between the simulation and flight-test time series,so it has a high accuracy and reliability. The steps of the new validation method are listed in detail. Finally,a case study is given to show the reasonability and validity of the improved model.