This work is concerned with large scaled nonlinear systems of equation
s resulting from discretization of problems in plasticity and viscopla
sticity in the context of the finite element method. The main purpose
is to show, how standard linear multigrid methods can be applied for s
olving the associated linear systems of equations in the frame of the
Newton-algorithm. To this end, a so-called Galerkin-approach is used f
or construction of coarse grid matrices by transformation of fine grid
matrices. It will be shown, how this transformation can be performed
very efficiently element-by-element wise. Stopping criteria for the in
ner iteration are based on theories for so-called inexact Newton metho
ds, where the linear systems are only solved approximately, however wh
ich preserve the rapid local convergence of Newtons method. In the num
erical examples it is demonstrated, how the proposed strategy reduces
the CPU-time for large scaled problems compared to solution techniques
, where the associated systems of linear equations are solved directly
.