Features for Building CSP Portfolio Solvers

Recent research has shown that the performances of a single arbitrarily efficient solver can be significantly outperformed by a portfolio of possibly slower on-average solvers. The solver selection is usually done by using machine learning techniques which exploit features extracted from the problem specification. In this paper we present a tool that is able to extract an extensive set of features from a Constraint Satisfaction Problem (CSP) defined either in the MiniZinc format or in the XCSP format. We also report some empirical results showing that the performances that can be obtained using these features are competitive with state of the art CSP portfolio techniques.