Dekker1/dekker-phd-thesis
Archived
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Full draft of the abstract

Browse Source
This commit is contained in:
Jip J. Dekker 2021-07-23 11:00:54 +10:00
parent 4cf69764f7
commit 2c3bd8bef9
No known key found for this signature in database
GPG Key ID: 517DF4A00618C9C3
1 changed files with 16 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
chapters/0_abstract.tex
View File

@ -2,15 +2,17 @@
\chapter{Abstract}\label{ch:abstract}
%************************************************
\noindent{}\Cmls{}, such as \glsxtrshort{ampl}, have long provided a \ldots{} way to model and solve real world problems.
In the field of Operations Research, their use is \ldots{} and used in areas such as scheduling, supply chain management, and transportation.
\vspace{-5em}
\noindent{}\Cmls{} are a prominent way to model and solve real world problems.
Their use extends to areas such as scheduling, supply chain management, and transportation.
In the past, these languages served mainly as a standardized interface between different \solvers{}.
The \gls{rewriting} required to translate an \instance{} of a \cmodel{} into a model to use as solver input was minimal.
The \gls{rewriting} required to translate an \instance{} of a \cmodel{} into a \gls{slv-mod} was negligible.
However, \cmls{} have evolved to include functionality that is no longer directly supported by the target \solver{}.
As such, the \gls{rewriting} process has become more important and complex.
\minizinc{}, one such language, was originally designed to target constraint programming \solvers{} where the result is a small number of highly complex \constraints{}.
The same \minizinc{} models now target mixed integer programming and Boolean satisfiability \solvers{}, resulting is a large number of very simple \constraints{}.
\minizinc{}, one such language, was originally designed for constraint programming \solvers{}, whose \glspl{slv-mod} contain small number of highly complex \constraints{}.
The same \minizinc{} models can now target mixed integer programming and Boolean satisfiability \solvers{}, resulting is a large number of very simple \constraints{}.
Distinctively, the \minizinc{}'s \gls{rewriting} process is founded on its functional language.
It generates \glspl{slv-mod} through the application of increasingly complex \minizinc{} functions from \solver{}-specific libraries.
Consequently, the efficiency of the functional evaluation of the language can be a limiting factor.
@ -18,22 +20,15 @@ For many applications, the current \minizinc{} implementation now requires a sig
This problem is exacerbated by the emerging use of \gls{meta-optimisation} algorithms, which require solving a sequence of closely related \instances{}.
In this thesis we revisit the \gls{rewriting} of functional \cmls{} into \glspl{slv-mod}.
We design and evaluate an architecture for \cmls{} that can accommodate the modern uses of these languages.
At the core of the this architecture lies a formal execution model that allows us rewrite \cmodels{} efficiently and actively manage the \gls{slv-mod}.
We show how our architecture can better detect and eliminate parts of the model that have become unused.
We design and evaluate an architecture for \cmls{} that can accommodate its the modern uses.
At its core lies a formal execution model that allows us rewrite \cmodels{} efficiently and actively manage the \gls{slv-mod}.
We show how it can better detect and eliminate parts of the model that have become unused.
The architecture is extended using a range of well-known simplification techniques to unsure the quality of the produced \glspl{slv-mod}.
In addition, we incorporate new analysis techniques to avoid the use of \glspl{reif} or replace them with \glspl{half-reif}, where possible.
Crucially, the architecture is designed to incorporate incremental \constraint{} modelling in two ways.
Primarily, the \gls{rewriting} process is fully incremental: changes made to the \instance{} through a provided interface require minimal addition \gls{rewriting} effort.
Moreover, we introduce \gls{rbmo}, a way to specify \gls{meta-optimisation} algorithms directly in \minizinc{}.
These specification are executed by a normal \minizinc{} \solver{}, requiring only a slight extension of its capabilities.
Crucially, this architecture should allow us to:
\begin{itemize}
\item easily integrate a range of well-known and new \textbf{optimisation and simplification} techniques,
\item effectively manage the \gls{slv-mod} and \textbf{detect and eliminate} parts of the model that have become unused,
\item formally \textbf{reason about \gls{reif}} to avoid it or use \gls{half-reif} where possible, and
\item support \textbf{incremental usage} of the \cml{}.
\end{itemize}
Together, the functionality of this architecture helps make \cmls{} a more powerful and attractive approach to solve real world problems.