Abstract: The Generalized Stochastic Simulation Method (GSSA) is an effective method for dealing with the numerical solution problems of DSGE models, especially when the model size is large, it has obvious advantages compared to other methods. Compared to ordinary random simulation methods, the generalized random simulation method has made several improvements to solve the ill conditioned problem of generating point sets, but in practical applications, the ill conditioned phenomenon is still relatively common and has a significant impact. The innovation of this article lies in proposing a method to introduce an improved EDS set into the generalized stochastic simulation method, which avoids pathological problems by processing the set of generated points, and improves computational efficiency by optimizing the spatial distribution of points, making it more suitable for application in larger scale DSGE models.

Keywords: DSGE; Random simulation; Numerical algorithm;