| U. Junker, `Quickxplain: Conflict detection for arbitrary constraint propagation algorithms', in IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, (2001). |
....sure that the search mechanism will avoid (as much as possible) to get back to states that have been explored and proved to be solution less. Learning, in a rough analogy with the brain, is an Computing precise conflicts is studied since works on TMS [4] and ATMS [3] More recent works include [10] and [11] and show that e#cient techniques do exist to compute precise conflicts when using filtering techniques. interaction between two processes, a recording process and a forgetting process. In our model, a learning component consists in three operators: a recording operator record, a ....
Ulrich Junker. QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, USA, August 2001.
....to find : why the problem has no solution; which constraint to relax in order to restore the coherence; etc. These questions yield two di#erent problems: explaining inconsistency and restoring consistency. Several theoretical answers have been provided to address those questions: QuickXPlain [5] computes conflict sets for configuration problems, 6] and [7] introduce tools to dynamically remove constraints, PaLM [8] uses conflict sets to address those issues and defines new search algorithms, 9] introduces constraint specific tools for providing user friendly solutions to constraint ....
Junker, U.: QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In: IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, USA (2001)
....re execution [13] Minimal (w.r.t. inclusion) explanations are the most interesting ones. They allow very precise information on emerging dependencies among variables and constraints, dependencies identified during the search. Unfortunately, computing such explanations can be quite time consuming [9]. A good tradeo# between size and computability is the use of the knowledge that is inside the solver. Indeed, constraint solvers always know (although not often explicitly) why they remove values from the domains of variables . Precise and interesting eliminating explanations can be computed ....
....i ,a) b) 2.4 Related work There are many works on computing conflicts i.e. contradiction explanations. On one hand, intrusive techniques include using (A)TMS systems [4] or adaptations [7] including the PaLM system [11] Non intrusive techniques have recently arisen: for example, QuickXPlain [9] following works from [1] and [5] iteratively tests the consistency of subsets of constraints in order to compute a minimal contradiction explanation for a given inconsistent set of constraints. However, computing inference explanation (i.e. eliminating explanations) has always, as far as we ....
Ulrich Junker. QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, USA, August 2001.
....to find: why the problem has no solution; which constraint to relax in order to restore the coherence; etc. These questions yield two different problems: explaining inconsistency and restoring consistency. Several theoretical answers have been provided to address those questions: QUICKXPLAIN [5] computes conflicts for configuration problems, 6] and [7] introduce tools to dynamically remove constraints, PALM [8] uses conflicts to address those issues and define new search algorithms, etc. User interaction requires user friendly and interactive tools. In this paper, we advocate for the ....
Junker, U.: QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In: IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, USA (2001)
....heuristics and its ability both to perform a local search and to prune the search space. In future work, we will investigate the following issues: In our implementation for open shop scheduling problems, the conflicts we compute are not minimal. A method like the one presented in [25] is able to compute minimal conflict in reasonable time. As more precise conflicts may greatly improve the efficiency of a decision repair implementation, such a conflict detection method deserves experimentation. Also experiments of decision repair over other fields than scheduling problems ....
....(A.1) # # dc 1 ##dc k ) A.1) Notice that, from that definition, if the current decision set CD is inconsistent, CD is a valid conflict. Obviously, a strict subset will be much more interesting. A minimal (w.r.t. the inclusion) conflict would be the best, but could be expensive to compute [25,51]. Our current implementation does not try to find such a minimal conflict. Instead, it tries to compute a good conflict quickly. A conflict must be computed each time the filtering algorithm can prove that no solution exists with the current decision set. This happens when the domain of a ....
U. Junker, QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms, Technical Report, 2001.
....4.2 Computing explanations Minimal (w.r.t. inclusion) explanations are the most interesting. They allow very precise information on emerging dependencies among variables and constraints, dependencies identified during the search. Unfortunately, computing such explanations is time consuming [Junker, 2001] . A good compromise between size and computability is the use of the knowledge that is inside the solver. Indeed, constraint solvers always know (although not often explicitly) why they remove values from the domains of variables. Precise and interesting eliminating explanations can be computed ....
Ulrich Junker. QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. Technical report, Ilog SA, 2001.
No context found.
U. Junker, `Quickxplain: Conflict detection for arbitrary constraint propagation algorithms', in IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, (2001).
No context found.
U. Junker, QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms, in: Proc. IJCAI-01, Workshop on Modelling and Solving Problems with Constraints, Seattle, WA, 2001.
No context found.
U. Junker. QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In IJCAI'01 Workshop on Modelling and Solving problems with constraints, 2001.
No context found.
U. Junker. QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, USA, Aug. 2001. 45
No context found.
U. Junker. Quickxplain: Conflict detection for arbitrary constraint propagation algorithms. In IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, 2001.
No context found.
Junker U, 2001, "Quickxplain: Conflict detection for arbitrary constraint propagation algorithms", IJCAI-2001 Workshop on Modeling and Solving Problems with Constraints.
No context found.
Ulrich Junker, `Quickxplain: Conflict detection for arbitrary constraint propagation algorithms', in IJCA'01 workshop on Modelling and Solving Problems with Constraints, pp. 81--88, (2001).
No context found.
Ulrich Junker. QUICKXPLAIN: Conflict detection for arbitrary constraint propagation algorithms. In IJCAI'01 Workshop on Modelling and Solving problems with constraints, Seattle, WA, USA, August 2001. 14
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC