NordConsNet Workshop 2025

The Workshop

The Nordic Network for researchers and practitioners of Constraint programming (NordConsNet, the network formerly known as SweConsNet) invites you to participate in its next workshop. The purpose is to share information about the research and practice of constraint programming (CP) in the Nordic countries. This is the jubilee 20th edition of this almost annual workshop.

Organisation

The NordConsNet Workshop 2025 is chaired by Pierre Flener and Justin Pearson, both of the Optimisation research group at Uppsala University, Sweden. They founded this network as SweConsNet in 2002, when they knew a bit less of HTML than now.

Dates

The workshop is held from lunch on Wednesday 20 August to lunch on Thursday 21 August 2025.

Location

The workshop takes place at Uppsala University in lecture hall 101195 (Heinz-Otto Kreiss) on floor 1 of house 10 at Ångström Laboratory on Regementsvägen 10 in Uppsala, Sweden. The two lunches are in Restaurang Rullan on floor 0 at the same address. The venue is a pleasant thirty-minute walk from the central railway station, and we have ordered good weather.

Accommodation

We hold 20 rooms for the workshop night at Best Western Hotel Uppsala on Trädgårdsgatan 5A in the centre, for 1,395 SEK (including VAT): write to booking@BestWesternUppsala.se (or call +4618121000) and mention "NordConsNet Workshop 2025": first come, first served.

Submit a Presentation Proposal

We hope for your participation, and encourage you to submit a proposal for a presentation of your ongoing work, recent results, or a relevant discussion topic. There are no abstract and paper submissions, no reviews, and no proceedings, hence recent conference or journal papers may also be presented. Contact Pierre Flener if you would like to present something.

Programme (tentative)

Wednesday 20 August 2025

Thursday 21 August 2025

Registration

Send an email to Justin Pearson. Registration --- including 2 lunches and fika --- must be sent no later than xx August 2025 (to be announced). If you register after that date, then you are still very welcome to attend talks (the lecture hall is large enough), but lunch cannot be guaranteed. There is no registration fee for the workshop.

You may forward this information to anyone who has a legitimate interest in this workshop but is not yet on the NordConsNet mailing list: they can subscribe to it by applying to Justin Pearson.

Abstracts

RAIRE: A Branch and Bound Approach to Auditing Instant-Runoff Voting Elections

Peter J. Stuckey (Monash University, Australia)

Risk-limiting post election audits guarantee a high probability of correcting incorrect election results, independent of why the result was incorrect. Ballot-polling audits select ballots at random and interpret those ballots as evidence for and against the actual recorded result, continuing this process until either they support the recorded result, or they find evidence that it is wrong. Ballot polling for first-past-the-post elections is well understood, and used in some US elections. We define the first approach to ballot-polling risk-limiting audits for Instant Runoff Voting (IRV) elections. We show that for almost all real elections we found, we can perform a risk-limiting audit by looking at only a small fraction of the total ballots. The approach has now been used to audit actual elections in Colorado.

Increasing Fairness in Online Multi-agent Combinatorial Optimisation

Maria Garcia de la Banda (Monash University, Australia)

Production paradigms such as cloud or shared manufacturing must consider the heterogeneous objectives that result from different stakeholders (or agents) sharing resources. While this points to fairness becoming increasingly important in online combinatorial optimisation, the fairness literature often ignores problems of online nature and with underlying NP-hard structures. Solving these multi-agent problems with purely utilitarian objectives, which simply maximise the sum of the agents’ utilities, can be arbitrarily unfair. Thus, ensuring some form of fairness among the agents is imperative. The need is particularly interesting in the online setting, where consecutive combinatorial problems are solved, each taking into account any new requests the agents made in that time period (e.g. a week). The online setting also allows us to consider and evaluate different metrics, as it decouples the overall fairness of the sequence of problems from the solving of each individual problem. We present a problem- and solver-independent priority-balancing framework that can significantly increase the fairness of the overall solution without substantial decrease of utility, and we experimentally explore different fairness measures and different strategies to re-balance the agents’ priority for each individual problem.

The Anatomy of the SICStus CLPFD Solver

Mats Carlsson (Research Institutes of Sweden)

A notable development in the history of CP was the realisation that unification in logic programming is a special case of constraint solving, leading to the constraint logic programming (CLP) framework. This led to the development of innovative CLP solvers as well as to the integration of constraints into classic Prolog systems.

For such an integration, technical solutions must be found for many tasks like domain representation, constraint propagation, search, dedicated filtering algorithms, and peaceful coexistence with the Prolog virtual machine and runtime system.

This talk gives an overview of the CP subsystem of SICStus Prolog: the key extensions to Prolog that were necessary, details of the solver architecture, supporting both integers and reals, and a discussion of design choices.

Automated, Scalable Product Configuration Optimisation

Erik Cervin Edin (Ericsson.com, Sweden)

... abstract coming up ...

Mikael Zayenz Lagerkvist (SambaNova Systems, Sweden)

RosterLogic Variation: Building a Custom CBLS Solver for a Single Problem

In "The Work Task Variation Problem" (CP 2025, Lagerkvist and Rattfeldt) we present a custom CBLS-like solver (built mainly in 2019) for solving the Work Task Variation (WTV) problem, and show that it is competitive for producing reasonable solutions quickly. But why did we make the effort to build our own solver for solving a single problem, and was that a smart choice? In this talk, I will dig deeper into why we built our own solver, including pragmatics, "business decisions", and the evolving landscape of combinatorial optimisation solvers over the last 6 years.

Dependency-Curated Large Neighbourhood Search

Frej Knutar Lewander (Uppsala University)

In large neighbourhood search (LNS), an incumbent initial solution is incrementally improved by selecting a subset of the variables, called the freeze set, and fixing them to their values in the incumbent solution, while a value for each remaining variable is found and assigned via solving (such as constraint programming-style propagation and search). Much research has been performed on finding generic and problem-specific LNS selection heuristics that select freeze sets that lead to high-quality solutions. In constraint-based local search (CBLS), the relations between the variables via the constraints are fundamental and well-studied, as they capture dependencies of the variables. We apply these ideas from CBLS to the LNS context, presenting the novel dependency curation scheme, which exploits them to find a low-cardinality set of variables that the freeze set of any selection heuristic should be a subset of. The scheme often improves the overall performance of generic selection heuristics. Even when the scheme is used with a naïve generic selection heuristic that selects random freeze sets, the performance is competitive with more elaborate generic selection heuristics.

... title ...

... presenter (affiliation) ...

... abstract ...