Insertion sequence variables have recently been introduced as a computational domain for modeling routing and sequencing problems in constraint programming. Typically, search heuristics guide the insertion process of new nodes into a partial growing path, while constraints eliminate infeasible insertions. This paper investigates filtering for the (minimum) distance constraint over insertion sequence variables. This global constraint links a sequence to a distance variable based on a given distance matrix. So far, only a simple filtering algorithm has been proposed, which considers the partial path but ignores mandatory nodes. Our contribution is to introduce stronger lower bounds that also take mandatory nodes into account. These bounds further enable the derivation of additional filtering rules for node insertions. An experimental evaluation on the TourMustSee problem shows that the proposed filtering rules significantly reduce the search space compared to the existing filtering approach.

The maximal clique enumeration (MCE) problem consists of computing and listing all maximal cliques in a finite graph. The problem is NP-complete since it subsumes the classic version of the NP-complete clique decision problem. The most well-known algorithm to solve this problem has been proposed by Bron & Kerbosch (BK). Generalizing this algorithm to hypergraphs presents significant challenges and admits multiple solutions. The primary difficulty stems from the fundamental difference in structure: while in standard graphs a vertex has a bounded number of incident edges, in hypergraphs a vertex can participate in an exponential number of hyperedges. Specifically, in a hypergraph with$n$vertices where hyperedges have arity$r$, a single vertex may belong to up to$\binom{n-1}{r-1}$hyperedges. This structural complexity renders the direct generalization of BK computationally inefficient. In this article, we propose three new approaches to solve the MHE problem. First, a relaxation method based on the relaxation of a hypergraph in a graph. Then, a method based on an efficiently computation of the hypercliques containing a set of vertices. Finally, a method combining the previous ones. We conducted an experimental study on a set of benchmarks, showing one or two orders of magnitude performance improvements of our approach with regards to the direct generalization of BK to hypergraphs.

Large Language Models (LLMs) excel at generating fluent text but struggle to enforce external constraints because they generate tokens sequentially without explicit control mechanisms. GenCP addresses this limitation by combining LLM predictions with Constraint Programming (CP) reasoning, formulating text generation as a Constraint Satisfaction Problem (CSP). In this paper, we improve GenCP by integrating Masked Language Models (MLMs) for domain generation, which allows bidirectional constraint propagation that leverages both past and future tokens. This integration bridges the gap between token-level prediction and structured constraint enforcement, leading to more reliable and constraint-aware text generation. Our evaluation on COLLIE benchmarks demonstrates that incorporating domain preview via MLM calls significantly improves GenCP's performance. Although this approach incurs additional MLM calls and, in some cases, increased backtracking, the overall effect is a more efficient use of LLM inferences and an enhanced ability to generate feasible and meaningful solutions, particularly in tasks with strict content constraints.

The success of Constraint Programming relies partly on the global constraints and implementation of the associated filtering algorithms. Recently, new ideas emerged to improve these implementations in practice, especially regarding the all different constraint. In this paper, we consider the cardinality constraint with costs. The cardinality constraint is a generalization of the all different constraint that specifies the number of times each value must be taken by a given set of variables in a solution. The version with costs introduces an assignment cost and bounds the total sum of assignment costs. The arc consistency filtering algorithm of this constraint is difficult to use in practice, as it systematically searches for many shortest paths. We propose a new approach that works with upper bounds on shortest paths based on landmarks. This approach can be seen as a preprocessing. It is fast and avoids, in practice, a large number of explicit computations of shortest paths.

In robust optimization, finding a solution that solely respects the constraints is not enough. Usually, the uncertainty and unknown parameters of the model are represented by random variables. In such conditions, a good solution is a solution robust to most-likely assignments of these random variables. Recently, the Confidence constraint has been introduced by Mercier-Aubin et al. in order to enforce this type of robustness in constraint programming. Unfortunately, it is restricted to a conjunction of binary inequalities In this paper, we generalize the Confidence constraint to any constraint and propose an implementation based on Multi-valued Decision Diagrams (MDDs). The Confidence constraint is defined over a vector of random variables. For a given constraint C, and given a threshold, the Confidence constraint ensures that the probability for C to be satisfied by a sample of the random variables is greater than the threshold. We propose to use MDDs to represent the constraints on the random variables. MDDs are an efficient tool for representing combinatorial constraints, thanks to their exponential compression power. Here, both random and decision variables are stored in the MDD, and propagation rules are proposed for removing values of decision variables that cannot lead to robust solutions. Furthermore, for several constraints, we show that decision variables can be omitted from the MDD because lighter filtering algorithms are sufficient. This leads to gain an exponential factor in the MDD size. The experimental results obtained on a chemical deliveries problem in factories – where the chemicals consumption are uncertain – shows the efficiency of the proposed approach.

Multi-objective problems are frequent in the real world. In general they involve several incomparable objectives and the goal is to find a set of Pareto optimal solutions, i.e. solutions that are incomparable two by two. In order to better deal with these problems in CP the global constraint Pareto was developed by Schaus and Hartert to handle the relations between the objective variables and the current set of Pareto optimal solutions, called the archive. This constraint handles three operations: adding a new solution to the archive, removing solutions from the archive that are dominated by a new solution, and reducing the bounds of the objective variables. The complexity of these operations depends on the size of the archive. In this paper, we propose to use a multi-valued Decision Diagram (MDD) to represent the archive of Pareto optimal solutions. MDDs are a compressed representation of solution sets, which allows us to obtain a compressed and therefore smaller archive. We introduce several algorithms to implement the above operations on compressed archives with a complexity depending on the size of the archive. We show experimentally on bin packing and multi-knapsack problems the validity of our approach.

This paper introduces a new approach to generating strongly constrained texts. We consider standardized sentence generation for the typical application of vision screening. To solve this problem, we formalize it as a discrete combinatorial optimization problem and utilize multivalued decision diagrams (MDD), a well-known data structure to deal with constraints. In our context, one key strength of MDD is to compute an exhaustive set of solutions without performing any search. Once the sentences are obtained, we apply a language model (GPT-2) to keep the best ones. We detail this for English and also for French where the agreement and conjugation rules are known to be more complex. Finally, with the help of GPT-2, we get hundreds of bona-fide candidate sentences. When compared with the few dozen sentences usually available in the well-known vision screening test (MNREAD), this brings a major breakthrough in the field of standardized sentence generation. Also, as it can be easily adapted for other languages, it has the potential to make the MNREAD test even more valuable and usable. More generally, this paper highlights MDD as a convincing alternative for constrained text generation, especially when the constraints are hard to satisfy, but also for many other prospects.

International audience

In CP, the most efficient model solving the TSP is the Weighted Circuit Constraint (WCC) combined with the k-cutset constraint. The WCC is mainly based on the edges cost of a given graph whereas the k-cutset constraint is a structural constraint. Specifically, for each cutset in a graph, the k-cutset constraint imposes that the size of the cutset is greater than or equal to two. In addition, any solution contains an even number of elements from this cutset. Isoart and Régin introduced an algorithm for this constraint. Unfortunately, their approach leads to a time complexity growing with the size of the considered cutsets, i.e. with k. Thus, they introduced an algorithm with a quadratic complexity dealing with k lower or equal to three. In this paper, we introduce a linear time algorithm for any k based on a DFS checking the consistency of this constraint and performing its filtering. Experimental results show that the size of most of the k-cutsets is lower or equal than 3. In addition, since the time complexity is improved, our algorithm also improves the solving times.

The LKH algorithm based on k-opt is an extremely efficient algorithm solving the TSP. Given a non-optimal tour in a graph, the idea of k-opt is to iteratively swap k edges of this tour in order to find a shorter tour. However, the optimality of a tour cannot be proved with this method. In that case, exact solving methods such as CP can be used. The CP model is based on a graph variable with mandatory and optional edges. Through branch-and-bound and filtering algorithms, the set of mandatory edges will be modified. In this paper, we introduce a new constraint to the CP model named mandatory Hamiltonian path constraint searching for k-opt in the mandatory Hamiltonian paths. Experiments have shown that the mandatory Hamiltonian path constraint allows us to gain on average a factor of 3 on the solving time. In addition, we have been able to solve some instances that remain unsolved with the state of the art CP solver with a 1 week time out.

One of the main problems for a realistic journey planning in public transit is the need to give the user multiple qualitative choices. Usually, public transit journeys involve 4 main criteria: the departure time, the arrival time, the number of transfers and the walking distance. The problem of computing Pareto sets with these criteria is called the Pareto range query problem. This problem is complex and difficult to solve within the constraints of the industrial world of smartphone applications, like a response time of the order of a second. In this paper, we present the Goal Directed Connection Scan Algorithm (GDCSA), an algorithm that allows, for the first time, to solve this problem with run times of less than 0.5 seconds on most European city or country-wide networks, like Berlin or Switzerland. In addition, GDCSA satisfies other industrial needs: it is conceptually simple and easy to implement. It partitions the graph in geographically small areas and precomputes some lower bounds on the duration of a trip in order to select for each itinerary a subset of these areas to decrease the number of scanned connections. Combining this subset and a journey planning using 4 criteria, we scan up to 17 times fewer connections than the best algorithms (CSA and RAPTOR), the number of nodes opened during the search is lowered by a factor of up to 2.9 and the query times are lowered by a factor of up to 9 on metropolitan networks. The integration of GDCSA in a smartphone app backend server led to an improvement in results by a factor of 5.

Multi-valued Decision Diagrams (MDDs) have been extensively studied in the last ten years. Recently, efficient algorithms implementing operators such as reduction, union, intersection, difference, etc., have been designed. They directly deal with the graph structure of the MDD and a time reduction of several orders of magnitude in comparison to other existing algorithms have been observed. These operators have permitted a new look at MDDs, because extremely large MDDs can finally be manipulated as shown by the models used to solve complex application in music generation. However, MDDs become so large (50GB) that minutes are sometimes required to perform some operations. In order to accelerate the manipulation of MDDs, parallel algorithms are required. In this paper, we introduce such algorithms. We carefully design them in order to overcome inherent difficulties of the parallelization of sequential algorithms such as data dependencies, software lock-out, false sharing, or load balancing. As a result, we observe a speed-up , i.e. ratio between parallel and sequential runtimes, growing linearly with the number of cores.

Recent developments of efficient propagators, operations and creation methods for MDDs allow us to directly build efficient MDD-based models, without the need for intermediate data structures. In this paper, we take another step in this direction by improving the propagators of cost MDDs. In addition, we introduce a soft MDD propagator in order to deal with unsatisfiable problems. This directly offers cost and soft versions for table constraints and any constraints which can be represented by an MDD (regular, slide, knapsack...).

We introduce an Embarrassingly Parallel Search (EPS) method for solving constraint problems in parallel, and we show that this method matches or even outperforms state-of-the-art algorithms on a number of problems using various computing infrastructures. EPS is a simple method in which a master decomposes the problem into many disjoint subproblems which are then solved independently by workers. Our approach has three advantages: it is an efficient method; it involves almost no communication or synchronization between workers; and its implementation is made easy because the master and the workers rely on an underlying constraint solver, but does not require to modify it. This paper describes the method, and its applications to various constraint problems (satisfaction, enumeration, optimization). We show that our method can be adapted to different underlying solvers (Gecode, Choco2, OR-tools) on different computing infrastructures (multi-core, data centers, cloud computing). The experiments cover unsatisfiable, enumeration and optimization problems, but do not cover first solution search because it makes the results hard to analyze. The same variability can be observed for optimization problems, but at a lesser extent because the optimality proof is required. EPS offers good average performance, and matches or outperforms other available parallel implementations of Gecode as well as some solvers portfolios. Moreover, we perform an in-depth analysis of the various factors that make this approach efficient as well as the anomalies that can occur. Last, we show that the decomposition is a key component for efficiency and load balancing.

International audience

In this paper we introduce a new cardinality constraint: Ordered Distribute. Given a set of variables, this constraint limits for each value v the number of times v or any value greater than v is taken. It extends the global cardinality constraint, that constrains only the number of times a value v is taken by a set of variables and does not consider at the same time the occurrences of all the values greater than v. We design an algorithm for achieving generalized arc-consistency on ORDERED DISTRIBUTE, with a time complexity linear in the sum of the number of variables and the number of values in the union of their domains. In addition, we give some experiments showing the advantage of this new constraint for problems where values represent levels whose overrunning has to be under control.

Constraint programming techniques are widely used to solve real-world problems. It often happens that such problems are over-constrained and do not have any solution. In such a case, the goal is to find a good compromise. A simple theoretical framework is the Max-CSP, where the goal is to minimize the number of constraint violations. However, in real-life problems, complex rules are generally imposed with respect to violations. Solutions which do not satisfy these rules have no practical interest. Therefore, many frameworks derived from the Max-CSP have been introduced. In this paper, we classify the most usual types of rules, and we show that some of them are not expressible in existing frameworks. We introduce a new paradigm in which all these rules can be encoded, through meta-constraints. Moreover, we show that most of existing frameworks can be included in our model.

C. Bessiere and M.O. Cordier (1994) said that the AC-6 arc consistency algorithm is optimal in time on constraint networks where nothing is known about the constraint semantics. However, in constraint networks, it is always assumed that constraints are bidirectional. None of the previous algorithms achieving arc-consistency (AC-3, AC-4, AC-6) use constraint bidirectionality. We propose here an improved version of AC-6 which uses this property. Then, we claim that our new algorithm is optimal in the number of constraint checks performed (i.e. given a variable, value, and arc ordering, it performs the minimum possible number of constraint checks according to these orders).>