25 lines
2.4 KiB
TeX
25 lines
2.4 KiB
TeX
\noindent{}In previous chapters, we explored \gls{rewriting} as a definitive linear process, where an \instance{} of a \cmodel{} is translated into a \gls{slv-mod}, for which a \solver{} produces \glspl{sol}.
|
|
However, to solve large-scale real-world problems, we often need to use \gls{meta-optimization} algorithms that solve very similar problems multiple times.
|
|
While improvements of the \gls{rewriting} process, such as the ones discussed in previous chapters, can increase the performance of these approaches, the overhead of completely\gls{rewriting} an almost identical \instances{} time and again may still prove prohibitive.
|
|
\Gls{meta-optimization} warrants direct support from the \cml{} architecture.
|
|
In this chapter, we introduce the following two methods to provide this support.
|
|
|
|
\begin{itemize}
|
|
\item We introduce the notion of \gls{rbmo} algorithms.
|
|
Using a minimal extension to a \cml{} and its target \solver{}, we can model \gls{meta-optimization} algorithms and rewrite them into efficient \glspl{slv-mod}.
|
|
The \solver{} will then incrementally execute the methods through the use of \solver{} \glspl{restart}.
|
|
This method completely avoids the need for repeated \gls{rewriting} of the \instance{}.
|
|
\item Alternatively, we extend our architecture with an incremental interface for adding and removing \constraints{}.
|
|
Although this method does not avoid repeatedly \gls{rewriting} the \instance{}, it uses \gls{incremental-rewriting} to reduce the \gls{rewriting} to the changes to the \instance{}.
|
|
This approach can be used when an incremental method cannot be described using \gls{rbmo} or when the required extensions are not available for the target \solver{}.
|
|
\end{itemize}
|
|
|
|
This chapter is organized as follows.
|
|
We first introduce examples of \gls{meta-optimization} algorithms and how they can be used from a \cml{}.
|
|
\Cref{sec:inc-modelling} discusses the declarative modelling of \gls{rbmo} algorithms in a \cml{}.
|
|
In \Cref{sec:inc-solver-extension}, we introduce the method to rewrite these \gls{meta-optimization} definitions into efficient \glspl{slv-mod} and the minimal extension required from the target \gls{solver}.
|
|
Then, \Cref{sec:inc-incremental-compilation} introduces the alternative method.
|
|
It extends the architecture presented in the previous chapters with an incremental \constraint{} modelling interface.
|
|
Finally, \Cref{sec:inc-experiments} reports on the experimental results of both approaches.
|
|
|