A main concern in Constraint Programming (CP) is to determine good variable and value order heuristics. However, this is known to be quite difficult as the effects on the solving process are rarely predictable. A novel solution to handle this concern is called Autonomous Search (AS), which is a special feature allowing an automatic reconfiguration of the solving process when a poor performance is detected. In this paper, we present a preliminary architecture for performing AS in CP. The idea is to perform an "on the fly" replacement of bad-performing heuristics by more promising ones. Another interesting feature of this architecture is its extensibility. It is possible to easily upgrade their components in order to improve the AS mechanism. © 2011 Springer-Verlag.
CITATION STYLE
Crawford, B., Soto, R., Castro, C., & Monfroy, E. (2011). Extensible CP-based autonomous search. In Communications in Computer and Information Science (Vol. 173 CCIS, pp. 561–565). https://doi.org/10.1007/978-3-642-22098-2_112
Mendeley helps you to discover research relevant for your work.