A filtering algorithm for constraints of difference in CSPs
AAAI '94 Proceedings of the twelfth national conference on Artificial intelligence (vol. 1)
New methods to color the vertices of a graph
Communications of the ACM
A machine program for theorem-proving
Communications of the ACM
Relaxations of the Satisfiability Problem Using Semidefinite Programming
Journal of Automated Reasoning
Proving Unsatisfiability of CNFs Locally
Journal of Automated Reasoning
Stochastic Local Search: Foundations & Applications
Stochastic Local Search: Foundations & Applications
Proving Graph Un-Colorability with a Consistency Check of CSP
ICTAI '05 Proceedings of the 17th IEEE International Conference on Tools with Artificial Intelligence
Ten challenges in propositional reasoning and search
IJCAI'97 Proceedings of the 15th international joint conference on Artifical intelligence - Volume 1
GUNSAT: a greedy local search algorithm for unsatisfiability
IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
Detecting unsatisfiable CSPs by coloring the micro-structure
AAAI'97/IAAI'97 Proceedings of the fourteenth national conference on artificial intelligence and ninth conference on Innovative applications of artificial intelligence
Local search for unsatisfiability
SAT'06 Proceedings of the 9th international conference on Theory and Applications of Satisfiability Testing
Improving GASAT by replacing tabu search by DLM and enhancing the best members
Artificial Intelligence Review
Hi-index | 0.00 |
Checking CSP consistency is shown, in theory, to be an NP-complete problem. There is two families of methods for CSP consistency checking. The first family holds the complete methods which make an exhaustive search on the solution space. These methods have the advantage to prove CSP inconsistency, but their complexity grows exponentially when the problem size increases. The second family includes the incomplete methods that make a local search on the solution space. These methods have been efficiently used to find solutions for large size consistent CSPs that complete methods can not solve. One major drawback of the incomplete methods, is their inability to prove CSP inconsistency. One of the challenges that have been put forward by the CP community (Selman et al. 1997) is to provide incomplete methods that can deal with CSP inconsistency efficiently. The work that we present here, is a contribution towards an answer to this hard challenge. We introduce a new incomplete method for CSP inconsistency checking that is based on both a new notion of dominance between CSPs and a coloration of the CSP micro-structure. We experimented the method on randomly generated CSP instances and the results obtained are very promising.