How can one succeed among more than 1,400 scientific papers from around the world? Researchers from the Faculty of Information Technology at the Czech Technical University in Prague (FIT CTU), from the GGOAT research group and the Algorithms Laboratory, managed to do so multiple times. Their papers, produced in collaboration with international partners, were accepted into the main program of AAMAS 2026 (The International Conference on Autonomous Agents and Multiagent Systems), the world's leading conference for experts in autonomous agents and multi-agent systems.
The 25th edition of the conference took place from May 25–29, 2026, in Paphos, Cyprus, and attracted record interest. A total of 1,455 full papers were submitted to the main track—approximately 50% more than at any point in the conference’s twenty-five-year history. Of these, only 531 were ultimately accepted. The papers authored by our researchers therefore succeeded in an exceptionally competitive environment.
The first paper, titled “Practical Approach to 2-Euclidean Preferences,” investigates whether voter preferences and candidates can be embedded in a two-dimensional space such that the distances between voters and candidates correspond to individual voter preferences. Such a model enables better analysis and visualization of voting outcomes. The authors Michal Dvořák, Dušan Knop, Jan Pokorný, and Martin Slávik present new approaches based on a novel class of forbidden substructures, integer and quadratically constrained programming, and reduction rules that simplify many real-world instances. Their approach significantly outperforms the previous state-of-the-art algorithm in both the number of solved instances and computational speed.
“Existing methods often struggled with larger or more complex datasets. Our approach is not only faster, but also enables the solution of many cases that were previously practically intractable,” explains Michal Dvořák.
The second paper, “Project Submission Games in Participatory Budgeting,” focuses on participatory budgeting, a process in which citizens propose and select projects funded through public resources. The authors study situations in which project proposers strategically influence outcomes by deciding which projects to submit for voting. The research provides new insights into when such systems can reach stable outcomes and how the behavior of individual proposers can be predicted. The paper also introduces algorithms that enable efficient analysis of these situations. The authors are Piotr Faliszewski, Łukasz Janeczko (AGH University of Krakow, Poland), Grzegorz Lisowski (University of Groningen, the Netherlands), Andrzej Kaczmarczyk (FIT CTU), and Grzegorz Pierczyński (University of Warsaw, Poland).
How can people be divided into groups so that everyone is satisfied with their assignment? This question is addressed in the third paper, “Individual Rationality in Constrained Hedonic Games: Additively Separable and Fractional Preferences.” The authors examine scenarios in which both the number and size of groups are constrained and demonstrate that finding a satisfactory partition can be much more challenging than it may initially appear. The research was conducted by Foivos Fioravantes (FIT CTU), Harmender Gahlawat (Ben-Gurion University of the Negev, Israel), Nikolaos Melissinos (Charles University), and Šimon Schierreich (FIT CTU).
The fourth paper, “Stability in Distance Preservation Games on Graphs,” investigates how to position participants within a network so that they are satisfied with their relative positions. Each participant has preferences regarding how close to—or far from—certain other participants they would like to be. The authors Argyrios Deligkas, Eduard Eiben, and Tiger-Lily Goldsmith (Royal Holloway, University of London), together with Dušan Knop and Šimon Schierreich seek to determine whether it is possible to find an arrangement in which no participant has an incentive to change their position. They also analyze the computational complexity of finding such stable solutions in different types of networks.
“The competition was exceptionally strong this year, so we are delighted that our work stood among the best contributions from around the world. This is recognition not only of the specific results achieved, but also of the efforts of the entire team and our international collaborators,” says Michal Dvořák of FIT CTU.
In addition to full papers, two extended abstracts were also published in the conference proceedings. The first, “Algorithms for Candidate Control in Sequential Participatory Budgeting Rules” (Šimon Schierreich and Krzysztof Sornat), investigates how the inclusion or exclusion of projects can influence participatory budgeting outcomes and how such situations can be analyzed efficiently.
The second abstract, “Detecting Approximate Clones under Approval Voting” (Théo Delemazure, Piotr Faliszewski, Łukasz Janeczko, Dušan Knop, Kristýna Pekárková, Jan Pokorný, Šimon Schierreich, and Ildikó Schlotter), examines how to identify candidates with very similar voter support in approval voting elections. Such candidates can affect election outcomes despite being nearly indistinguishable from the voters’ perspective.
