Source language: Engels
Although it is a hybrid method, GSQN is closer to baseline Broyden than to
baseline Gauss–Seidel. There is another alternative, a reduced GSQN method,
R-GSQN, that is closer to baseline Gauss–Seidel. Rather than updating a full Jacobi
matrix, R-GSQN only updates the elements of W and works with this lowdimensional
object throughout.
All of the conventional methods discussed so far have in common that they solve
the model for each time element of all aggregate variables.