A. Wolf, Incremental Adaption of Constraint Handling Rule Derivations, CP'97 Workshop on the Theory and Practice of Dynamic Constraint Satisfaction, Linz, Austria, November 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....to a CSP. Dynamic constraint solvers like DJCHR [21] or the system J.CP we are going to describe in this paper do not require state restoration techniques for search and are applicable to solve DynCSPs. These solvers provide a reverse operation to constraint posting, called constraint retraction [22, 10, 7]. With this operation arbitrary non deterministic search algorithms can be implemented that post and retract instantiation constraints [21, 17] The search algorithm posts instantiations to explore the search space and retracts them, if in inconsistency occurred due to their propagation. The ....

....domains from a history snapshot stored in the constraints that were a ected and by reposting all these constraints. Our algorithm is an adaption of retraction in CLP(FD) 10] where we abandon the use of a centralized history management by storing consequences of constraints locally (similar to [22]) 2.2 Search in ACS In our prototypical ACS implementation J.CP we implemented two generic labeling constraints. The rst one uses Backtracking that allows complete search and is thus used, if all solutions of a CSP are to be found. The other algorithm is a very simple local search that is ....

Armin Wolf, Thomas Gruenhagen, and Ulrich Geske. On incremental adaption of constraint handling rule derivation after constraint deletions. Journal of Applied Arti cial Intelligence, Special Issue on Constraint Handling Rules, 2000.

....(3) The thus reduced variable domains may be used in later constraint executions. If c is retracted, these changes must be reversed in the system. Our constraint retraction algorithm, shown in Figure 4, is a mixture of the algorithms for retraction in clp(fd,S) 6] and deletion in DJCHR [18, 19]. First we nd all constraints that may have been a ected by the retracted constraint from an internal storage of dependencies (like in [19] Second we enlarge the a ected variable domains to the maximum inferrable size (analogously to [6] And nally we repost all relevant constraints such that ....

....the system. Our constraint retraction algorithm, shown in Figure 4, is a mixture of the algorithms for retraction in clp(fd,S) 6] and deletion in DJCHR [18, 19] First we nd all constraints that may have been a ected by the retracted constraint from an internal storage of dependencies (like in [19]) Second we enlarge the a ected variable domains to the maximum inferrable size (analogously to [6] And nally we repost all relevant constraints such that a new reduction of variable domains is achieved (like in [19] and [6] In contrast to the mentioned algorithms we do not make use of any ....

[Article contains additional citation context not shown here]

Armin Wolf, Thomas Gruenhagen, and Ulrich Geske. On incremental adaption of constraint handling rule derivation after constraint deletions. Journal of Applied Arti cial Intelligence, Special Issue on Constraint Handling Rules, 2000.

....or to nd all solutions to a CSP. Dynamic constraint solvers like our ACS implementation J.CP [14] or DJCHR [15] do not require state restoration techniques and are still applicable for dynamic problems. These solvers provide a reverse operation to constraint posting, called constraint retraction [16, 8, 7]. Thus arbitrary non deterministic search algorithms can be implemented that post instantiation constraints and reverse this operation, if an inconsistency has been detected [15] During search, dynamic solvers compute states, that would be the current state, if something would not have been done ....

....domains from a history snapshot stored in the constraints that were a ected and by reposting all these constraints. Our algorithm is an adaption of retraction in CLP(FD) 8] where we abandon the use of a centralized history management by storing consequences of constraints locally (similar to [16]) 2.2 Search in ACS In our prototypical ACS implementation J.CP we implemented two generic labeling constraints. The rst one is Backtracking that allows complete search and is thus used, if all solutions of a CSP are to be found. The other algorithm is a very simple local search that is ....

Armin Wolf, Thomas Gruenhagen, and Ulrich Geske. On incremental adaption of constraint handling rule derivation after constraint deletions. Journal of Applied Arti cial Intelligence, Special Issue on Constraint Handling Rules, 2000.

....variable domains from the history stored in the constraints that were a ected and by reposting all these constraints. Our algorithm is an adaption of retraction in CLP(FD) 8] where we abandon the use of a centralized history management by storing consequences of constraints locally (similar to [18]) 3.2 Search in ACS In ACS non deterministic search is implemented as a global constraint over all variables to be labeled. This labeling constraint posts and retracts variableinstantiation constraints until a solution of the implemented CSP is found. These actions can be taken in any order ....

Armin Wolf, Thomas Gruenhagen, and Ulrich Geske. On incremental adaption of constraint handling rule derivation after constraint deletions. Journal of Applied Arti cial Intelligence, Special Issue on Constraint Handling Rules, 2000.

....CHR have been used successfully in challenging applications, where other existing CLP systems could not be applied with the same results in terms of simplicity, flexibility and efficiency. In most real life applications, soft and dynamic constraints are required. Work that has just been started [Wol97] indicates that CHR are helpful in implementing general schemes to handle such constraints independent of the constraint domain. The topics for research mentioned in the first draft paper on CHR in 1991 were: ffl Correctness w.r.t. specifications ffl Termination and confluence ffl Negation and ....

A. Wolf, Incremental Adaption of Constraint Handling Rule Derivations, CP'97 Workshop on the Theory and Practice of Dynamic Constraint Satisfaction, Linz, Austria, November 1997.

No context found.

A. Wolf, Incremental Adaption of Constraint Handling Rule Derivations, CP'97 Workshop on the Theory and Practice of Dynamic Constraint Satisfaction, Linz, Austria, November 1997.

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