D. Iteration Schemes in

The need for iteration schemes arises from the fact that, when solving very
complex coupled equation systems, the solution can very often not be obtained
from the available initial-guess as the region of attraction for the Newton
scheme would be to small. Hence, the problem can be split into different
levels of complexity with each of them using the previous level as an
initial-guess to further refine the solution by applying more complicated
models. This procedure will be called *iteration scheme* in the
following. Typical iteration schemes are:

- For hydrodynamic simulations it is beneficial to use a drift-diffusion simulation as an initial-guess.
- For unipolar devices like MOSFETs it is sometimes useful to neglect the continuity equation for the minority carriers and to use a constant quasi-Fermi level approximation to calculate an initial-guess for the fully coupled system.
- Several relaxation schemes are known which solve subsets of the equation system alternatingly.

These iteration schemes are normally hard-coded in the simulator and can only
be marginally adjusted by the developer. This makes it very cumbersome to
implement new schemes or experiment with various different configurations.
This is especially important when new or different equations are added since
their influence on the coupled system can be more easily detected then. Hence
it was decided to provide *MINIMOS-NT* with an interface so that iteration schemes
can be arbitrarily programmed with several additional options making use of
the features provided by the *IPL*.

An iteration scheme consists of arbitrarily nested iteration blocks. Each block can have subblocks which will be evaluated recursively. The following gives a short overview of the major features available for defining iteration blocks.

1999-05-31