Jean-Noël Monette

Towards solver-independent propagators

We present an extension to indexicals to describe propagators for global constraints. The resulting language is compiled into actual propagators for different solvers, and is solver-independent. In addition, we show how this high-level description eases the proof of propagator properties, such as correctness and monotonicity. Experimental results show that propagators compiled from their indexical descriptions are sometimes not significantly slower than built-in propagators of Gecode. Therefore, our language can be used for the rapid prototyping of new global constraints.

A constraint-based local search backend for MiniZinc

MiniZinc is a modelling language for combinatorial problems, which can then be solved by a solver provided in a backend. There are many backends, based on technologies such as constraint programming, integer programming, or Boolean satisfiability solving. However, to the best of our knowledge, there is currently no constraint-based local search (CBLS) backend. We discuss the challenges to develop such a backend and give an overview of the design of a CBLS backend for MiniZinc. Experimental results show that for some MiniZinc models, our CBLS backend, based on the OscaR/CBLS solver, is able to give good-quality results in competitive time.

A Parametric Propagator for Discretely Convex Pairs of Sum Constraints

We introduce a propagator for abstract pairs of Sum constraints, where the expressions in the sums respect a form of convexity. This propagator is parametric and can be instantiated for various concrete pairs, including DEVIATION, SPREAD, and the conjunction of Sum and COUNT. We show that despite its generality, our propagator is competitive in theory and practice with state-of-the-art propagators. This work is supported by grants 2011-6133 and 2012-4908 of the Swedish Research Council (VR). We thank the reviewers for their constructive comments.

Toward sustainable development in constraint programming

We present a few challenges that we consider important to tackle for the future of constraint programming. The focus is put on simplifying the design and implementation of propagators in solvers.

A Position-Based Propagator for the Open-Shop Problem

The Open-Shop Problem is a hard problem that can be solved using Constraint Programming or Operation Research methods. Existing techniques are efficient at reducing the search tree but they usually do not consider the absolute ordering of the tasks. In this work, we develop a new propagator for the One-Machine Non-Preemptive Problem, the basic constraint for the Open-Shop Problem. This propagator takes this additional information into account allowing, in most cases, a reduction of the search tree. The underlying principle is to use shaving on the positions. Our propagator applies on one machine or one job and its time complexity is in

Auto-tabling for subproblem presolving in MiniZinc

\mathcal{O}(N^2 \log N)

Efficient Filtering for the Unary Resource with Family-based Transition Times

, where Nis either the number of jobs or machines. Experiments on the Open-Shop Problem show that the propagator adds pruning to state-of-the-art constraint satisfaction techniques to solve this problem.

Automated auxiliary variable elimination through on-the-fly propagator generation

A well-known and powerful constraint model reformulation is to compute the solutions to a model part, say a custom constraint predicate, and tabulate them within an extensional constraint that replaces that model part. Despite the possibility of achieving higher solving performance, this tabling reformulation is often not tried, because it is tedious to perform; further, if successful, it obfuscates the original model. In order to encourage modellers to try tabling, we extend the MiniZinc toolchain to perform the automatic tabling of suitably annotated predicate definitions, without requiring any changes to solvers, thereby eliminating both the tedium and the obfuscation. Our experiments show that automated tabling yields the same tables as manual tabling, and that tabling is beneficial for solvers of several solving technologies.

A parametric propagator for pairs of Sum constraints with a discrete convexity property

We recently proposed an extension to Vilim’s propagators for the unary resource constraint in order to deal with sequence-dependent transition times. While it has been shown to be scalable, it suffers from an important limitation: when the transition matrix is sparse, the additional filtering, as compared to the original from Vilim’s algorithm, drops quickly. Sparse transition time matrices occur especially when activities are grouped into families with zero transition times within a family. The present work overcomes this weakness by relying on the transition times between families of activities. The approach is experimentally evaluated on instances of the Job-Shop Problem with Sequence Dependent Transition Times. Our experimental results demonstrate that the approach outperforms existing ones in most cases. Furthermore, the proposed technique scales well to large problem instances with many families and activities.

Relevant subgraph extraction from random walks in a graph

Model flattening often introduces many auxiliary variables. We provide a way to eliminate some of the auxiliary variables occurring in exactly two constraints by replacing those two constraints by a new equivalent constraint for which a propagator is automatically generated on the fly. Experiments show that, despite the overhead of the preprocessing and of using machine-generated propagators, eliminating auxiliary variables often reduces the solving time.