Olivier Lhomme | IBM - Academia.edu (original) (raw)
Papers by Olivier Lhomme
ion (du monde r#el) [11]. Cette approche permet de proposer une architecture originale dans des m... more ion (du monde r#el) [11]. Cette approche permet de proposer une architecture originale dans des mondes homog#nes o# tout composant peut #tre d#crit par des d#pendances ou en mondes h#t#rog#nes o# l'int#gration d'un gestionnaire de dialogue devient n#cessaire. Ces travaux ont donn# lieu # plusieurs pr#sentations dans le cadre des GDR et # un d#but de collaboration avec une #quipe du CNET Sophia. A Controler P C P C graphical object 1 graphical object 2 constraints rules P P C graphical object 1 graphical object 2 A PAC Hierarchy The FLO Implementation A Controler C C constraints rules C A PAC Hierarchy The FLO Implementation A A application object 1 application object 2 application object 1 application object 2 A A Figure 5: Mod#lisation FLO de PAC 2.3.5 Contraintes r#actives pour les interfaces graphiques L'utilisation de contraintes pour la r#alisation d'interfaces graphiques constitue une extension au mod#le propos# par P. Franchi, dans les ann#es 88-92, pour g...
In this paper, we introduce a new solving algorithm for Constraint Satisfaction Problems (CSP). I... more In this paper, we introduce a new solving algorithm for Constraint Satisfaction Problems (CSP). It performs an overall local search helped with a domain filtering tech-nique to prune the search space. Conflicts detected dur-ing filtering are used to guide the search. First experi-ments with a tabu version of the algorithm have shown good results on hard instances of open shop scheduling problems. It competes well with the best highly special-ized algorithms.
Electronic Proceedings in Theoretical Computer Science, 2011
Modern software systems, like GNU/Linux distributions or Eclipse-based development environment, a... more Modern software systems, like GNU/Linux distributions or Eclipse-based development environment, are often deployed by selecting components out of large component repositories. Maintaining such software systems by performing component upgrades is a complex task, and the users need to have an expressive preferences language at their disposal to specify the kind of upgrades they are interested in. Recent research has shown that it is possible to develop solvers that handle preferences expressed as a combination of a few basic criteria used in the MISC competition, ranging from the number of new components to the freshness of the final configuration. In this work we introduce a set of new criteria that allow the users to specify their preferences for solutions with components aligned to the same upstream sources, provide an efficient encoding and report on the experimental results that prove that optimising these alignment criteria is a tractable problem in practice.
We use a lexicographical preference order on the problem space to combine solution synthesis with... more We use a lexicographical preference order on the problem space to combine solution synthesis with conflict learning. Given two preferred solutions of two subproblems, we can either combine them to a solution of the whole problem or learn a 'fat' conflict which cuts off a whole subtree. The approach makes conflict learning more pervasive for Constraint Programming as it well exploits efficient support finding and compact representations of Craig interpolants.
Interval narrowing techniques are a key issue for handling constraints overreal numbers in the lo... more Interval narrowing techniques are a key issue for handling constraints overreal numbers in the logic programming framework. However, the standardfixed-point algorithm used for interval narrowing may give rise to cyclicphenomena and hence to problems of slow convergence. Analysis of thesecyclic phenomena shows: 1) that a large number of operations carried outduring a cycle are unnecessary; 2) that many others could be removed fromcycles and performed only once when these cycles have been...
The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, alth... more The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, although it may scan an exponential number of supporting tuples. In this paper, we develop a major improvement of this scheme. When searching for a support, our new algorithm is able to skip over a number of tuples exponential in the arity of the constraint by exploiting knowledge about the current domains of the variables. We demonstrate the effectiveness of the method for large table constraints.
European Conference on Artificial Intelligence, 1998
In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. T... more In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. The usual search technique over numeric csps is a dichotomic search interleaved with a consistency ltering, which is called domain splitting. This paper proposes to replace chronological backtracking at the core of domain splitting by a non destructive backtracking technique. ment 25]. In other words, a partial assignment A is a nogood if and only if no solution contains A.
Informs Journal on Computing, 2011
Since their introduction, local search algorithms have consistently represented the state-ofthe-a... more Since their introduction, local search algorithms have consistently represented the state-ofthe-art in solution techniques for the classical job-shop scheduling problem. This is despite the availability of powerful search and inference techniques for scheduling problems developed by the constraint programming community. In this paper, we introduce a simple hybrid algorithm for job-shop scheduling that leverages both the fast, broad search capabilities of modern tabu search algorithms and the scheduling-specific inference capabilities of constraint programming. The hybrid algorithm significantly improves the performance of a state-ofthe-art tabu search algorithm for the job-shop problem, and represents the first instance in which a constraint programming algorithm obtains performance competitive with the best local search algorithms. Further, the variability in solution quality obtained by the hybrid is significantly lower than that of pure local search algorithms. Beyond performance demonstration, we perform a series of experiments that provide insights into the roles of the two component algorithms in the overall performance of the hybrid.
Aaai, 2005
The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, alth... more The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, although it may scan an exponential number of supporting tuples. In this paper, we develop a major improvement of this scheme. When searching for a support, our new algorithm is able to skip over a number of tuples exponential in the arity of the constraint by exploiting knowledge about the current domains of the variables. We demonstrate the effectiveness of the method for large table constraints.
Proceedings of the 13th International Joint Conference on Artifical Intelligence Volume 1, 1993
Many problems can be expressed in terms of a numeric constraint satisfaction problem over finite ... more Many problems can be expressed in terms of a numeric constraint satisfaction problem over finite or continuous domains (numeric CSP). The purpose of this paper is to show that the consistency techniques that have been developed for CSPs can be adapted to numeric CSPs. Since the numeric domains are ordered the underlying idea is to handle domains only by their bounds. The semantics that have been elaborated, plus the complexity analysis and good experimental results, confirm that these techniques can be used in real applications.
Jfplc, 1998
This paper introduces a new way of accelerating the convergence of numeric CSP filtering algorith... more This paper introduces a new way of accelerating the convergence of numeric CSP filtering algorithms, through the use of extrapolation methotis. Extrapolation methods are used in numerical analysis to accelerate the convergence of real number sequences. We will show how to use them for solving numeric csPs, leading to drastic improvement in efficiency.
Dans le domaine militaire, la projection de forces consiste à envoyer le plus rapidement possible... more Dans le domaine militaire, la projection de forces consiste à envoyer le plus rapidement possible, en utilisant les moyens mis à disposition, des matériels en différents théâtres d'opérations tout en respectant un certain nombre de contraintes. Un tel déploiement nécessite une étape consistant à charger les moyens de transport (des avions dans notre cas) avec le matériel concerné. L'objectif du travail présenté ici est de réaliser un outil automatisé qui réalise cette opération en cherchant à minimiser le nombre total de trajets pour acheminer tout le matériel.
Ginsberg and McAllester [5] have shown that systematic and nonsystematic search algorithms can be... more Ginsberg and McAllester [5] have shown that systematic and nonsystematic search algorithms can be quite close. In this paper, we go one step further in the direction of understanding the relationships between systematic and nonsystematic search algorithms by introducing the PLM model. We introduce a generic search algorithm for solving csp, and we show that search algorithms can be decomposed in primitives. A reduced set of primitives is sufficient to express almost all existing search algorithms. The knowledge of these primitives is of great help for comparing search algorithms as well as for configuring new search algorithms.
In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. T... more In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. The usual search technique over numeric csps is a dichotomic search interleaved with a consistency ltering, which is called domain splitting. This paper proposes to replace chronological backtracking at the core of domain splitting by a non destructive backtracking technique. ment 25]. In other words, a partial assignment A is a nogood if and only if no solution contains A.
ion (du monde r#el) [11]. Cette approche permet de proposer une architecture originale dans des m... more ion (du monde r#el) [11]. Cette approche permet de proposer une architecture originale dans des mondes homog#nes o# tout composant peut #tre d#crit par des d#pendances ou en mondes h#t#rog#nes o# l'int#gration d'un gestionnaire de dialogue devient n#cessaire. Ces travaux ont donn# lieu # plusieurs pr#sentations dans le cadre des GDR et # un d#but de collaboration avec une #quipe du CNET Sophia. A Controler P C P C graphical object 1 graphical object 2 constraints rules P P C graphical object 1 graphical object 2 A PAC Hierarchy The FLO Implementation A Controler C C constraints rules C A PAC Hierarchy The FLO Implementation A A application object 1 application object 2 application object 1 application object 2 A A Figure 5: Mod#lisation FLO de PAC 2.3.5 Contraintes r#actives pour les interfaces graphiques L'utilisation de contraintes pour la r#alisation d'interfaces graphiques constitue une extension au mod#le propos# par P. Franchi, dans les ann#es 88-92, pour g...
In this paper, we introduce a new solving algorithm for Constraint Satisfaction Problems (CSP). I... more In this paper, we introduce a new solving algorithm for Constraint Satisfaction Problems (CSP). It performs an overall local search helped with a domain filtering tech-nique to prune the search space. Conflicts detected dur-ing filtering are used to guide the search. First experi-ments with a tabu version of the algorithm have shown good results on hard instances of open shop scheduling problems. It competes well with the best highly special-ized algorithms.
Electronic Proceedings in Theoretical Computer Science, 2011
Modern software systems, like GNU/Linux distributions or Eclipse-based development environment, a... more Modern software systems, like GNU/Linux distributions or Eclipse-based development environment, are often deployed by selecting components out of large component repositories. Maintaining such software systems by performing component upgrades is a complex task, and the users need to have an expressive preferences language at their disposal to specify the kind of upgrades they are interested in. Recent research has shown that it is possible to develop solvers that handle preferences expressed as a combination of a few basic criteria used in the MISC competition, ranging from the number of new components to the freshness of the final configuration. In this work we introduce a set of new criteria that allow the users to specify their preferences for solutions with components aligned to the same upstream sources, provide an efficient encoding and report on the experimental results that prove that optimising these alignment criteria is a tractable problem in practice.
We use a lexicographical preference order on the problem space to combine solution synthesis with... more We use a lexicographical preference order on the problem space to combine solution synthesis with conflict learning. Given two preferred solutions of two subproblems, we can either combine them to a solution of the whole problem or learn a 'fat' conflict which cuts off a whole subtree. The approach makes conflict learning more pervasive for Constraint Programming as it well exploits efficient support finding and compact representations of Craig interpolants.
Interval narrowing techniques are a key issue for handling constraints overreal numbers in the lo... more Interval narrowing techniques are a key issue for handling constraints overreal numbers in the logic programming framework. However, the standardfixed-point algorithm used for interval narrowing may give rise to cyclicphenomena and hence to problems of slow convergence. Analysis of thesecyclic phenomena shows: 1) that a large number of operations carried outduring a cycle are unnecessary; 2) that many others could be removed fromcycles and performed only once when these cycles have been...
The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, alth... more The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, although it may scan an exponential number of supporting tuples. In this paper, we develop a major improvement of this scheme. When searching for a support, our new algorithm is able to skip over a number of tuples exponential in the arity of the constraint by exploiting knowledge about the current domains of the variables. We demonstrate the effectiveness of the method for large table constraints.
European Conference on Artificial Intelligence, 1998
In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. T... more In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. The usual search technique over numeric csps is a dichotomic search interleaved with a consistency ltering, which is called domain splitting. This paper proposes to replace chronological backtracking at the core of domain splitting by a non destructive backtracking technique. ment 25]. In other words, a partial assignment A is a nogood if and only if no solution contains A.
Informs Journal on Computing, 2011
Since their introduction, local search algorithms have consistently represented the state-ofthe-a... more Since their introduction, local search algorithms have consistently represented the state-ofthe-art in solution techniques for the classical job-shop scheduling problem. This is despite the availability of powerful search and inference techniques for scheduling problems developed by the constraint programming community. In this paper, we introduce a simple hybrid algorithm for job-shop scheduling that leverages both the fast, broad search capabilities of modern tabu search algorithms and the scheduling-specific inference capabilities of constraint programming. The hybrid algorithm significantly improves the performance of a state-ofthe-art tabu search algorithm for the job-shop problem, and represents the first instance in which a constraint programming algorithm obtains performance competitive with the best local search algorithms. Further, the variability in solution quality obtained by the hybrid is significantly lower than that of pure local search algorithms. Beyond performance demonstration, we perform a series of experiments that provide insights into the roles of the two component algorithms in the overall performance of the hybrid.
Aaai, 2005
The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, alth... more The GAC-Scheme has become a popular general purpose algorithm for solving n-ary constraints, although it may scan an exponential number of supporting tuples. In this paper, we develop a major improvement of this scheme. When searching for a support, our new algorithm is able to skip over a number of tuples exponential in the arity of the constraint by exploiting knowledge about the current domains of the variables. We demonstrate the effectiveness of the method for large table constraints.
Proceedings of the 13th International Joint Conference on Artifical Intelligence Volume 1, 1993
Many problems can be expressed in terms of a numeric constraint satisfaction problem over finite ... more Many problems can be expressed in terms of a numeric constraint satisfaction problem over finite or continuous domains (numeric CSP). The purpose of this paper is to show that the consistency techniques that have been developed for CSPs can be adapted to numeric CSPs. Since the numeric domains are ordered the underlying idea is to handle domains only by their bounds. The semantics that have been elaborated, plus the complexity analysis and good experimental results, confirm that these techniques can be used in real applications.
Jfplc, 1998
This paper introduces a new way of accelerating the convergence of numeric CSP filtering algorith... more This paper introduces a new way of accelerating the convergence of numeric CSP filtering algorithms, through the use of extrapolation methotis. Extrapolation methods are used in numerical analysis to accelerate the convergence of real number sequences. We will show how to use them for solving numeric csPs, leading to drastic improvement in efficiency.
Dans le domaine militaire, la projection de forces consiste à envoyer le plus rapidement possible... more Dans le domaine militaire, la projection de forces consiste à envoyer le plus rapidement possible, en utilisant les moyens mis à disposition, des matériels en différents théâtres d'opérations tout en respectant un certain nombre de contraintes. Un tel déploiement nécessite une étape consistant à charger les moyens de transport (des avions dans notre cas) avec le matériel concerné. L'objectif du travail présenté ici est de réaliser un outil automatisé qui réalise cette opération en cherchant à minimiser le nombre total de trajets pour acheminer tout le matériel.
Ginsberg and McAllester [5] have shown that systematic and nonsystematic search algorithms can be... more Ginsberg and McAllester [5] have shown that systematic and nonsystematic search algorithms can be quite close. In this paper, we go one step further in the direction of understanding the relationships between systematic and nonsystematic search algorithms by introducing the PLM model. We introduce a generic search algorithm for solving csp, and we show that search algorithms can be decomposed in primitives. A reduced set of primitives is sufficient to express almost all existing search algorithms. The knowledge of these primitives is of great help for comparing search algorithms as well as for configuring new search algorithms.
In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. T... more In this paper, a new search technique over numeric csps is presented: dynamic domain splitting. The usual search technique over numeric csps is a dichotomic search interleaved with a consistency ltering, which is called domain splitting. This paper proposes to replace chronological backtracking at the core of domain splitting by a non destructive backtracking technique. ment 25]. In other words, a partial assignment A is a nogood if and only if no solution contains A.