A Study on Automatic Constraint Solver’s Configuration Based on ParamILS

The identification of performance-optimizing parameter settings promotes the development and application of the algorithm, which is one of the important research directions in the area of computer science. This paper is based on the novel and famous ParamILS framework. First it connected the ParamILS framework with three familiar constraint solvers, configured the solvers’ parameters by the framework. Then it carried out some experiments by solving some benchmark instances, and obtained some valuable results. Finally it analyzed the experimental results and compared the performances in the two situations. The experimental results show that the presented work will be useful for the constraint solvers’ optimization.