MPE model: a method for the ideal parameters selection based on multiple regression in YHGSM

In the Yin-he Global Spectral Model (YHGSM) developed by the National University of Defense Technology, an issue to be studied is that how to determine the parameters and achieve the ideal execution time in the model. Based on the multiple regression analysis, we firstly construct a quantitative relationship between the parameters and the execution time in the model by giving a set of random parameters and its execution time in advance. Then we establish the multiple regression equation between them. In this way, we finally establish the model parameter evaluation (MPE) model. In the new model, by giving more random parameters, we can predict their corresponding execution time. On this basis, given a critical value at which the execution time may be accepted, when the execution time is less than the critical value and the model can be operated normally, the determined parameters at this time are the ideal parameters that we recognize.