programmer_introduction.tex

由matlab开发的hybrid系统的描述语言

\nonumchapter{Introduction}

The \hysdel{} compiler was originally written in C. With version
1.3, it was ported to C++. The use of the object oriented paradigm
makes it easier to understand the code and to extend it.

Chapter \ref{intern_rep_c} introduces the elementary classes of
the \hysdel{} compiler and states how they store all the
information given in the \hysdel{} source file. Chapter
\ref{sem_check_c} explains what is done to guarantee that the read
file (and its internal representation) is a valid \hysdel{}
description. In chapter \ref{mme_comp_c} we describe how the upper
and lower bounds needed for an AD or DA converter can be computed
automatically. Chapter \ref{unroll_c} introduces the unrolling, a
preprocessing step that simplifies the statements in the \hysdel{}
source such that they can easily be translated into \MLD{}
matrices. This translation is described in chapter
\ref{mld_transl_c}. The \hysdel{} compiler can also produce a
simulator of the hybrid system, a tool that can be used to debug
the \hysdel{} description. It is documented in chapter
\ref{simulator_c}. Chapter \ref{optimization_c} says what is done
to minimize the number of inequalities in the produced \MLD{}
system. Finally, chapter
\ref{alloc_pol_c} and \ref{globals_c} deal with two details of the
implementation, namely the allocation policy and how objects used by
many classes are accessed.

In this document, code and function names are set in
\code{italics}. Code that spans multiple lines is sometimes set in
{\tt Courier}. Class methods are given with opening and closing
braces, e.g., \method{Item}{compute\_mme\_from\_aff}. This does
not mean that the method takes no arguments. If you need the full
function prototype, please refer to the source code.
\nocite{HYSDELU02}