Hyper-heuristics for the dynamic variable ordering in constraint satisfaction problems

摘要: The idea behind hyper-heuristics is to discover some combination of straightforward heuristics solve a wide range problems. To be worthwhile, such should outperform the single heuristics. This paper presents GA-based method that produces general for dynamic variable ordering within Constraint Satisfaction Problems. GA uses variable-length representation, which evolves combinations condition-action rules producing after going through learning process includes training and testing phases. Such hyper-heuristics, when tested with large set benchmark problems, produce encouraging results most cases. testebed composed problems randomly generated using an algorithm proposed by Prosser.