Modeling abstractions is the subject of
. It introduces TCAD models which are
abstract representation of arbitrary TCAD simulation tools. We will
see how *SIESTA* encapsulates these simulation tools, and how
elaborate simulation models can be created.

introduces
*SIESTA*'s TCAD experiments. Those TCAD experiments allow users
to perform thorough investigations on simulation models. It explains
how optimization, calibration, design of experiments (DOE), or
statistical analysis can be carried out.

goes into more detail about how
*SIESTA*'s TCAD models can be utilized to build simulation
models. It gives examples of the encapsulation of simulation tools by
**models**. Furthermore, possibilities for a structured and hierarchical
description of simulation problems will be discussed. This will give
the reader an idea of the strengths which are intrinsic to TCAD
**models**. is concluded with insight
into evaluation mechanisms of TCAD **models**.

discusses the job farming capabilities of
*SIESTA*. The implications of parallel computation on a cluster of
heterogenous workstations are outlined and a load-balancing mechanism
for the efficient distribution of simulation workload is explained.

*SIESTA*'s implementation and its core components will be
outlined in . It highlights features which contributed
to the implementation of a modular simulation environment. Some key
components are explained in detail to give the reader an idea of
*SIESTA*'s internal structure.

The application of *SIESTA*'s facilities to the optimization of IC
fabrication technology is demonstrated in . A
comprehensive simulation model of a fabrication technology is
explained and an optimization is performed based on that model. The
optimization procedure, in particular, how *SIESTA* is dealing with
it, concludes .

shows how inverse modeling can be performed
with *SIESTA*. It describes how calibrations can be carried out with
comprehensive sets of measurements. Furthermore, an example
demonstrates how *SIESTA* can be used for inverse modeling of doping
profiles on the basis of electrical measurements.

1999-05-27