Speeding up constraint propagation by redundant modeling

Cheng, Bing-shu; Lee, Jimmy; Wu, Jianhuang

doi:10.1007/3-540-61551-2_68

Cited by 50 publications

(66 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The columns corresponding to the tuples (1,2,3), (1,2,4), (1,2,5),...,(1,2,k ) were labelled, in that order; clearly, the matrix is then completely specified. We tried labelling sets of non-overlapping columns (as far as possible given the value of k), for instance, those corresponding to the tuples (1,2,3) and (4,5,6) when fk ¼ 6: this was inspired by the labelling of non-overlapping Fsupercell_ variables in [30], which in that case was a successful labelling strategy. In this case, however, it performed very poorly.…”

Section: Methodsmentioning

confidence: 99%

“…Note that we can still extract a CSP solution from any SAT solution: by clauses (5) in any SAT solution each combination of symbols occurs in at least one row of A for each combination of t columns; by clauses (6) each such occurrence induces the corresponding entries in M; and by clauses (2) no more than one value is possible in each M position. In fact the omitted clauses (1,3,4) are implied by clauses (2,5,6), and experiments suggest that omitting them makes little difference to the search effort. It reduces the size of the encoding but not its space complexity, which is dominated by the channelling constraints and is Oð k t À Á btg t Þ literals.…”

Section: A Model For Local Searchmentioning

confidence: 99%

“…For instance, y r;ð1;2;3Þ ¼ 4 represents x r1 ¼ 1, x r2 ¼ 0, x r3 ¼ 0 while y r;ð1;2;4Þ ¼ 5 represents x r1 ¼ 1, x r2 ¼ 0, x r4 ¼ 1, so that the pair of values (4,5) should be allowed by the intersection constraint.…”

Section: Coverage Constraintsmentioning

confidence: 99%

“…However, we still have a large number of intersection constraints, even after redundant constraints are removed. We propose to integrate the two models by using the variables of both, linked by channelling constraints, in the manner proposed by Cheng et al [5]. The dis- Fig.…”

Section: An Integrated Modelmentioning

confidence: 99%

See 3 more Smart Citations

Constraint Models for the Covering Test Problem

Hnich

Prestwich

Selensky³

et al. 2006

Constraints

106

View full text Add to dashboard Cite

Covering arrays can be applied to the testing of software, hardware and advanced materials, and to the effects of hormone interaction on gene expression. In this paper we develop constraint programming models of the problem of finding an optimal covering array. Our models exploit global constraints, multiple viewpoints and symmetry-breaking constraints. We show that compound variables, representing tuples of variables in our original model, allow the constraints of this problem to be represented more easily and hence propagate better. With our best integrated model, we are able to either prove the optimality of existing bounds or find new optimal solutions, for arrays of moderate size. Local search on a SAT-encoding of the model is able to find improved solutions and bounds for larger problems.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: A Model For Local Searchmentioning

confidence: 99%

Section: Coverage Constraintsmentioning

confidence: 99%

Section: An Integrated Modelmentioning

confidence: 99%

See 2 more Smart Citations

Constraint Models for the Covering Test Problem

Hnich

Prestwich

Selensky³

et al. 2006

Constraints

106

View full text Add to dashboard Cite

show abstract

“…Good modellers know that combining alternative representations in the same model can improve propagation, among other benefits. To maintain the consistency between these simultaneous alternatives we need to add channelling constraints [3] to the model. Section 4 discuses channelling constraints (also called channels) between alternative representations.…”

Section: Introductionmentioning

confidence: 99%

Towards the Systematic Generation of Channelling Constraints

Martínez‐Hernández

Frisch

2005

Principles and Practice of Constraint Programming - CP 2005

View full text Add to dashboard Cite

Abstract. The automatic modelling tool Conjure generates CSP models from problem specifications. The generated models may contain multiple alternative (redundant) representations of the same variable. However, the combined models generated by Conjure do not include the channelling constraints needed to maintain the consistency between the alternative representations. In this paper we show that by using only the tools already provided by the Conjure system we can generate correct channelling constraints.

show abstract

The Automatic Generation of Redundant Representations and Channeling Constraints

Martínez‐Hernández¹,

Frisch²

2007

Trends in Constraint Programming

View full text Add to dashboard Cite

Constraint modelling is the process of encoding a given problem to be solved by constraint satisfaction technology. Automatic modelling systems aim to reduce the number of decisions human modellers must take. To do so, these systems implement common modelling guidelines and techniques. In this paper we focus on the automatic addition of redundant information and most importantly the corresponding channelling constraints to synchronise it. We discuss and formalise a systematic method of generation of both elements; redundancy and channelling constraints. We provide in this paper a new insight on this formalisation that aims to clarify and increase previous work on the subject [1].

show abstract

Speeding up constraint propagation by redundant modeling

Cited by 50 publications

References 10 publications

Constraint Models for the Covering Test Problem

Constraint Models for the Covering Test Problem

Towards the Systematic Generation of Channelling Constraints

The Automatic Generation of Redundant Representations and Channeling Constraints

Contact Info

Product

Resources

About