Diaz D. and Codognet P., A minimal extension of the WAM for clp(FD). In Proc. of the 10th International Conference on Logic Programming (ICLP'93), pp:774790, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and interpreted code. CLAIRE was inspired from the work of LAURE. It attempted to contain many of the design features of LAURE with added simplicity and C compliance. CLAIRE has had great success with scheduling problems such as the Job Shop Problem [CL1,96] 27 2.3. 5 clp(FD) The clp(FD) DC,93] system consists of a constraint logic programming language (over finite domains) and the wamcc [Diaz,91] Prolog compiler, which translates Prolog to C via the Warren Abstract Machine [Warren,83] It was one of the first systems to provide methods for defining high level constraints. This was ....

Diaz, D.; Codognet, P.: A Minimal Extension of the WAM for clp(FD), Proceedings of the Tenth International Conference on Logic Programming, pages 774-790, Budapest, Hungary, 1993.

....to support more exible experimentation with constraint solvers. First generation CLP languages, such as CLP(R) 15] provided almost no support. They had a xed underlying solver for each constraint domain which was viewed as a closed black box. Second generation CLP languages, such as clp(fd) [6], provided more support by viewing the solver as a glass box which could be extended to support problemspeci c complex constraints. However, CLP programmers want more than this: they want to be able to develop new problem speci c constraint solvers, for example by using hybrid methods that ....

D. Diaz and P. Codognet. A minimal extension of the WAM for clp(fd). In Procs. of ICLP93, pages 774-790, 1993.

....based on a single generic constraint that allows the user to define and control the constraint propagation. These constraints, often referred to as indexicals, are very efficient [20] since the implementation uses a simple interval narrowing technique which can be smoothly integrated into the WAM [1, 17]. As a result, clp(FD) is now part of mainstream CLP systems such as SICStus Prolog [10] and IF Prolog [27] On the other hand, the CHR (now included as a library in SICStus Prolog [10] enable the creation of new user defined domains and their solvers and allowany interaction between them. ....

....new system of approximation for infinite domains that ensures that the constraint solver only has to deal with computable values. Indexical constraints The indexical approach was first implemented in [14] where finite interval constraints of the form x in r were solved via an extension of the WAM [17]. Well known CLP systems such as SICStus [31] and IF Prolog [27] nowintegrate the xinrconstraint to provide a glass box solver for FD, the finite domain of integers. In [12] the idea was extended to the clp(FD B) system. This system integrates a Boolean solver with the existing FD solver. A ....

D. Diaz and P. Codognet. A minimal extension of the WAM for clp(FD). In D.S. Warren, editor, 10th International ConferenceonLogic Programming (ICLP'93), pages 774--790, Budapest, Hungary, 1993. The MIT Press.

....other, that is, the constraint is not available first class (see also [28] Thus, algorithmic techniques employing global reasoning on all arguments of the constraint cannot be incorporated in this setting. The first implementation of an indexical based finite domain solver was in clp(FD) see [10]) The language is Prolog based and it compiles to C code. It provides for a couple of symbolic constraints (e.g. element) as well as boolean constraints. Due to an indexical based solver, it does not support 20 global constraints. AKL(FD) 4] implements indexicals in a concurrent constraint ....

D. Diaz and P. Codognet. A minimal extension of the WAM for clp(FD). In Proceedings of the International Conference on Logic Programming, pages 774--790, Budapest, Hungary, 1993. MIT Press.

....that integrate powerful inferences on global constraints into backtracking algorithms [Nadel, 1989] Because of the well founded semantics of hard constraints [Mackworth, 1992] PROLOG like languages denote constraints by built in predicates. Constraint propagation is used to improve unification [Diaz and Codognet, 1993] of these predicates. When using soft constraints, the situation is a bit more complex. Fages presented a characterization of optimization within constraint logic programming [Fages et al. 1996] However, this characterization makes a distinction between objective functions and ....

D. Diaz and P. Codognet. A minimal extension of the wam for clp(fd). In Proceedings of the International Conference on Logic Programming,pages 774--790,Budapest, Hungary, 1993.

....the benchmark sets used in the evaluation, and finally in Section 3.4 we discuss the results. 1 Chapter 2 The theory of NP(FD) 2. 1 Introduction Propagation algorithms have turned out to be quite powerful in proving theorems and solving combinatorial search problems [Hen89, HSD92b, JM94, DC93b, Car95] In particular, algorithms specialized to discrete constraint satisfaction problems, so called finite domain problems, have been carefully studied [HDT92] Similarly, propagation techniques tailored for propositional logic have been designed which compare favorably with other approaches ....

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In Proceedings of the International Conference on Logic Programming. MIT Press, 1993.

....use some of the techniques developed in the LOGFLOW system [21, 14] for which there already exists an animation tool [20] that can be used to tune load balancing strategies. Another foreseen extension is to combine constraint propagation techniques, already available in the inference engine wamcc [11] with ORparallelism. Such a combination was already proven to be very effective [22] Acknowledgements All Apache project research reports are available by ftp at ftp.imag.fr in the directory pub APACHE RAPPORTS. Compulog Net Meeting on Parallelism and Implementation Technology 11 ....

D. Diaz and P. Codognet. A minimal extension of the wam for clp(fd). In Proc. of ICLP'93, 10th International Conference on Logic Programming, Budapest, Hungary, 1993. The MIT Press.

....by the Swedish National Board for Technical and Industrial Development (NUTEK) under the auspices of Advanced Software Technology (ASTEC) Center of Competence at Uppsala University. We include a collection of examples, some of which have been distributed with the INRIA implementation of clp(FD) Diaz Codognet 93] 33.2 Solver Interface The solver is available as a library module and can be loaded with a query : use module(library(clpfd) The solver contains predicates for checking the consistency and entailment of finite domain constraints, as well as solving for solution values for your problem ....

....is associated with its particular constraint solving algorithm. The solver maintains two scheduling queues, giving priority to the queue of indexicals. The feasibility of integrating the indexical approach with a Prolog based on the WAM was clearly demonstrated by Diaz s clp(FD) implementation [Diaz Codognet 93] one of the fastest finite domains solvers around. 33.2.1 Posting Constraints A constraint is called as any other Prolog predicate. When called, the constraint is posted to the store. For example: X in 1. 5, Y in 2. 8, X Y #= T. X in 1. 5, Y in 2. 8, T in 3. 13 yes X in 1. 5, ....

D. Diaz and P. Codognet, A Minimal Extension of the WAM for clp(FD), Proceedings of the International Conference on Logic Programming, MIT Press, 1993.

....the benchmark sets used in the evaluation, and finally in Section 3.4 we discuss the results. 1 Chapter 2 The theory of NP(FD) 2. 1 Introduction Propagation algorithms have turned out to be quite powerful in proving theorems and solving combinatorial search problems [Hen89, HSD92b, JM94, DC93b, Car95] In particular, algorithms specialized to discrete constraint satisfaction problems, so called finite domain problems, have been carefully studied [HDT92] Similarly, propagation techniques tailored for propositional logic have been designed which compare favorably with other approaches ....

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In Proceedings of the International Conference on Logic Programming. MIT Press, 1993.

....and pruning. 23 It is interesting as well to observe that early versions of the paper also started the so called glass box approach to constraint programming. This had the sidee ect of producing new results in the implementation of constraint programming over nite domains. The CLP(FD) system [14] is a good example of this approach and it is signi cantly more ecient than cc(FD) 7.2. The Failure Towards Lower Level Constraint Languages The underlying dream behind the paper was the declarative speci cation of userde ned constraints. The cc(FD) combinators all have a simple logical ....

D. Diaz and P. Codognet. A minimal extension of the wam for clp(fd). In Proceedings of the Tenth International Conference on Logic Programming (ICLP-93), pages 774-792, Budapest (Hungary), June 1993.

.... ; Ai) P gen : generate(M; M = generate(M; M = h ; JijM 0 ] indomain(J) generate(M 0 ; 4 Conclusion At least one order of magnitude is gained in efficiency by switching from an ordinary symbolic language to a constraint one, and our automatically synthesisable CLP(FD) [2] programs are only 3 to 5 times slower than carefully hand crafted ones [2] which is encouraging since none of the obvious problem specific optimising transformations have been performed yet on our programs. Since our synthesis is fully automatable, starting from short and elegant formal ....

.... 0 ] indomain(J) generate(M 0 ; 4 Conclusion At least one order of magnitude is gained in efficiency by switching from an ordinary symbolic language to a constraint one, and our automatically synthesisable CLP(FD) 2] programs are only 3 to 5 times slower than carefully hand crafted ones [2], which is encouraging since none of the obvious problem specific optimising transformations have been performed yet on our programs. Since our synthesis is fully automatable, starting from short and elegant formal specifications (which can even be generated from some form of controlled English ....

D. Diaz and Ph. Codognet. A minimal extension of the WAM for clp(FD). In: D.S. Warren (ed), Proc. of ICLP'93, pp. 774--790. The MIT Press, 1993.

....a single generic constraint This work was partly supported by EPSRC grants GR L19515 and GR M05645 and by CICYT grant TIC98 0445 C03 03. often referred to as an indexical. The implementation of indexicals uses a simple interval narrowing technique which can be smoothly integrated into the WAM [2, 6]. This approach has been shown to be adaptable and very efficient and now integrated into mainstream CLP systems such as SICStus Prolog. This paper has two contributions. First, we provide a theoretical framework for the indexical approach to constraint solvers. This is formulated for any ordered ....

Diaz D. and Codognet P., A minimal extension of the WAM for clp(FD). In Proc. of the 10th International Conference on Logic Programming (ICLP'93), pp:774790, 1993.

No context found.

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In 10th International Conference on Logic Programming, Budapest, Hungary, The MIT Press 1993.

....boolean solver can be derived from a nite domain constraint solver for free. It was therefore quite natural to investigate such a possiblity with our CLP system clp(FD) which handles nite domain constraints similar to that of CHIP, but being nevertheless more than twice as fast on average [6]. clp(FD) is based on the so called glass box approach proposed by [23] as opposed to the black box approach of the CHIP solver for instance. The basic idea is to have a single constraint X in r, where r is a range (e.g. t 1 : t 2 ) More complex constraints such as linear equations and ....

....in CLP languages, which o er much more exibility by providing a full logic language to state the problem and generate the boolean formulas. PrologIII, CHIP and clp(B FD) only fall in the latter category. 3 clp(FD) in a nutshell As introduced in Logic Programming by the CHIP language, clp(FD) [6] is a constraint logic language based on nite domains, where constraint solving is done by propagation and consistency techniques originating from Constraint Satisfaction Problems [25] 11] 14] Very close to those methods are the interval arithmetic constraints of BNR Prolog [1] The novelty of ....

[Article contains additional citation context not shown here]

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In proc. ICLP'93, 10th International Conference on Logic Programming, Budapest, Hungary, MIT Press 1993.

....solver. Furthermore, the implementation of a glass box system lifts the complexity from the emulator level to the compiler level, as well as making available traditional compiler optimization techniques. The net result being that the implementations of glass box systems can be made highly ecient [3, 13, 1]. In this paper we study how to use the glass box system FD [12] to de ne 1 non trivial nite domain constraints and to check their entailment. We show that by using conditional reasoning, based on entailment, complex symbolic constraints can be de ned in FD. Entailment has previously been ....

....for describing node and arc consistency propagation. The unary constraints of FD can be used as the target language for compilers of arbitrary nite domain constraints [2] and in fact FD subsumes basically all existing nite domain constraint systems with preserved and sometimes improved eciency [1, 3, 4]. 2.1 The theory FD is based on domain constraints X 2 I, where I is a set of integers described by a nite union of intervals. The set I is the set of possible values of X, and X is said to be constrained to I. X 2 I is satis ed by assigning a value in I to X. A set S of domain constraints, ....

[Article contains additional citation context not shown here]

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In Proceedings of the 10th International Conference on Logic Programming, 1993.

....implicit Prolog C type conversion, transparent I O argument handling, non deterministic C code, ISO error support, etc. This is a key point to allow users to write their own extensions. Finally, GNU Prolog includes a very ecient constraint solver over nite domains inspired from clp(FD) [5, 6] containing many prede ned constraints: arithmetic constraints, boolean constraints, symbolic constraints, rei ed constraints; there are more than 50 FD built in constraints predicates, and several prede ned labeling heuristics. Moreover the solver is extensible, and new highlevel constraints ....

P. Codognet and D. Diaz. A Minimal Extension of the WAM for clp(FD). In Proc. ICLP'93, 10th International Conference on Logic Programming. Budapest, Hungary, MIT Press, 1993.

....from 0am) The output data is a set of satisfactory departure times fSET k;1 g of the ights F k such that all capacity constraints are satis ed and the average delay undergone by a ight is minimized. 3 clp(FD) in a Nutshell As introduced in Logic Programming by the CHIP language, clp(FD) [6] is a constraint logic language based on nite domains, where constraint solving is done by propagation and consistency techniques originating from Constraint Satisfaction Problems [15, 17, 10] The novelty of clp(FD) is the use of a unique primitive constraint which allows users to de ne their ....

....and particular optimizations hidden inside the black box. When a constraint X in r has been reexecuted, if D X = DX it was useless to reexecute it (i.e. it has neither failed nor reduced the domain of X) Hence, we have designed three simple but powerful optimizations for the X in r constraint [6, 2] which encompass many previous particular optimizations for FD constraints: some constraints are equivalent so only the execution of one of them is needed. In the previous example, when C 2 is called in the store fX in 0: 4; Y in 0: 1g Y is set to 0: 4. Since the domain of Y has been updated, ....

[Article contains additional citation context not shown here]

D. Diaz and P. Codognet. A minimal extension of the WAM for clp(FD). In 10th International Conference on Logic Programming, Budapest, Hungary, 1993. MIT Press.

....Higher Order, constraints, parallel or concurrent LP. However the design of simple and yet ecient Prolog systems is still an active topic. We started to develop our own logical engine in 1991, in order to provide a sound basis for various extensions and in particular Constraint Logic Programming [6]. The requirements for this new system were as follows: extensibility: the system should be used as an experimental platform. This implies simplicity in the design of the Prolog engine without complex optimizations that would excessively enlarge the size of the system. portability: in order to ....

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In 10th International Conference of Logic Programming, Budapest, Hungary, MIT Press 1993.

No context found.

P. Codognet and D. Diaz. A Minimal Extension of the WAM for clp(FD). In 10th International Conference on Logic Programming, Budapest, Hungary, MIT Press, 1993.

.... C type conversion, transparent I O argument handling, non deterministic C code, ISO error support, etc. Such an interface allows users to easily write their own extensions. GNU Prolog also includes an ecient constraint solver over nite domains (FD) similar to that of the clp(FD) described in [5, 6]. The key feature of such a solver is the use of a simple (low level) language that embeds the core propagation mechanism in order to de ne all (high level) FD constraints. There are many advantages to this approach: constraints can be compiled, the user can de ne his own constraints (thanks to ....

....Programming, which has had a signi cant impact on a variety of industrial applications, see [9] It is thus natural to include a constraint solving extension to any modern Prolog based system. GNU Prolog compiles nite domain constraints in the same way as its predecessor clp(FD) described in [5, 6]. It is based on the so called RISC approach which consists in translating at compile time all complex user constraints (e.g. disequations, linear equations or inequations) into simple, primitive constraints (the FD constraint system) at a lower level which really embeds the propagation ....

[Article contains additional citation context not shown here]

P. Codognet and D. Diaz. A Minimal Extension of the WAM for clp(FD). In Proc. ICLP'93, 10th International Conference on Logic Programming. Budapest, Hungary, MIT Press, 1993.

No context found.

Diaz D. and Codognet P., A minimal extension of the WAM for clp(FD). In Proc. of the 10th International Conference on Logic Programming (ICLP'93), pp:774790, 1993.

No context found.

D. Diaz and P. Codognet. A minimal extension of the WAM for clp(FD). In D.S. Warren, editor, 10th International ConferenceonLogic Programming (ICLP'93), pages 774--790, Budapest, Hungary, 1993. The MIT Press.

No context found.

Codognet, P. & Diaz, D. (1993). A minimal extension of the WAM for clp(FD). Proc. of the 10th International Conference on Logic Programming, MIT Press, pp. 774--790.

No context found.

D. Diaz and P. Codognet. A Minimal Extension of the WAM for clp(FD). In David S. Warren, editor, Proceedings of the Tenth International Conference on Logic Programming, pages 774--790, Budapest, Hungary, 1993. The MIT Press.

No context found.

Daniel Diaz and Phillipe Codognet. A minimal extension of the WAM for clp(FD). In David S. Warren, editor, Proceedings of the 10th International Conference on Logic Programming, pages 774-790. MIT Press, 1993.

First 50 documents  Next 50

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