We introduce groupoids - generalisations of groups in which not all pairs of elements may be multiplied, or, equivalently, categories in which all morphisms are invertible - as the appropriate algebraic structures for dealing with conditional symmetries in Constraint Satisfaction Problems (CSPs). We formally define the Full Conditional Symmetry Groupoid associated with any CSP, giving bounds for the number of elements that this groupoid can contain. We describe conditions under which a Conditional Symmetry sub-Groupoid forms a group, and, for this case, present an algorithm for breaking all conditional symmetries that arise at a search node. Our algorithm is polynomial-time when there is a corresponding algorithm for the type of group involved. We prove that our algorithm is both sound and complete - neither gaining nor losing solutions. © Springer-Verlag Berlin Heidelberg 2007.
CITATION STYLE
Gent, I. P., Kelsey, T., Linton, S. A., Pearson, J., & Roney-Dougal, C. M. (2007). Groupoids and conditional symmetry. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4741 LNCS, pp. 823–830). Springer Verlag. https://doi.org/10.1007/978-3-540-74970-7_60
Mendeley helps you to discover research relevant for your work.