Programmable Restoration Granularity in Constraint Programming

TL;DR

This paper proposes a method to dynamically adapt restoration granularity in constraint programming systems, enhancing flexibility and efficiency by using aspect-oriented programming concepts, demonstrated through a Gecode implementation.

Contribution

It introduces a novel approach for programmable restoration granularity in constraint programming, enabling dynamic adaptation during search.

Findings

Dynamic adaptation improves search efficiency.

Aspect-oriented programming effectively models restoration.

Gecode implementation shows promising results.

Abstract

In most constraint programming systems, a limited number of search engines is offered while the programming of user-customized search algorithms requires low-level efforts, which complicates the deployment of such algorithms. To alleviate this limitation, concepts such as computation spaces have been developed. Computation spaces provide a coarse-grained restoration mechanism, because they store all information contained in a search tree node. Other granularities are possible, and in this paper we make the case for dynamically adapting the restoration granularity during search. In order to elucidate programmable restoration granularity, we present restoration as an aspect of a constraint programming system, using the model of aspect-oriented programming. A proof-of-concept implementation using Gecode shows promising results.