A Framework for Autonomous Search in the Eclipse Solver
Autonomous Search (AS) is a special feature allowing systems to improve their performance by self-adaptation. This approach has been recently adapted to Constraint Programming (CP) reporting promising results. However, as the research lies in a preliminary stage there is a lack of implementation frameworks and architectures. This hinders the research progress, which in particular, requires a considerable work in terms of experimentation. In this paper, we propose a new framework for implementing AS in CP. It allows a dynamic self-adaptation of the classic CP solving process and an easy update of its components, allowing developers to define their own AS-CP approaches. We believe this will help researchers to perform new AS experiments, and as a consequence to improve the current preliminary results.
Unable to display preview. Download preview PDF.
- 1.Crawford, B., Montecinos, M., Castro, C., Monfroy, E.: A hyperheuristic approach to select enumeration strategies in constraint programming. In: Proceedings of ACT, pp. 265–267. IEEE Computer Society, Los Alamitos (2009)Google Scholar
- 2.Hamadi, Y., Monfroy, E., Saubion, F.: Special issue on autonomous search. Contraint Programming Letters 4 (2008)Google Scholar
- 3.Hamadi, Y., Monfroy, E., Saubion, F.: What is autonomous search? Technical Report MSR-TR-2008-80, Microsoft Research (2008)Google Scholar
- 7.Soubeiga, E.: Development and Application of Hyperheuristics to Personnel Scheduling. PhD thesis, University of Nottingham School of Computer Science (2009)Google Scholar