Dynamic Pricing Through Multi-Agent Systems: Reconstructing Contact for the Legal Characterization of Concerted Practices
DOI:
https://doi.org/10.17398/2695-7728.42.2428Keywords:
algorithmic collusion, competition law, digital markets, intelligent systems, restrictive agreementsAbstract
This article examines dynamic pricing implemented through multi-agent reinforcement learning systems and its implications for the legal characterization of concerted practices under Article 101 TFEU. It analyzes the legal role of contact in determining whether reciprocal interactions between dynamic pricing systems based on multi-agent reinforcement learning fall within the concept of a concerted practice. To this end, it develops a conceptual and evidentiary analysis of Article 101 TFEU and the case law of the Court of Justice of the European Union, considered alongside the technical architecture and available empirical evidence on multi-agent reinforcement learning applied to dynamic pricing. The analysis argues that contact should be functionally reinterpreted where signals processed by MARL systems reduce strategic uncertainty in ways that cannot be explained solely by autonomous market adaptation. On that basis, it proposes a legal framework for characterizing such interactions, built on a rebuttable presumption supported by technical factors and an objective model of corporate attribution.
Downloads
References
BRINNEN, Carl, Culpa in Code: Exploring Autonomous Algorithmic Collusion under EU Competition Law, tesis de máster, Lund University, 2023.
CALVANO, Emilio, CALZOLARI, Giacomo, DENICOLÒ, Vincenzo y PASTORELLO, Sergio, «Artificial Intelligence, Algorithmic Pricing, and Collusion», American Economic Review, vol. 110, núm. 10, 2020, pp. 3267-3297.
CALZOLARI, Luca, «The Misleading Consequences of Comparing Algorithmic and Tacit Collusion: Tackling Algorithmic Concerted Practices Under Art. 101 TFEU», European Papers, vol. 6, núm. 2, 2021, pp. 1193-1228.
EZRACHI, Ariel y STUCKE, Maurice E., Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy, Harvard University Press, Cambridge, MA, 2016.
EZRACHI, Ariel y STUCKE, Maurice E., «Algorithmic Tacit Collusion», en WHELAN, Peter (ed.), Research Handbook on Cartels, Edward Elgar Publishing, Cheltenham, 2023, pp. 187-212.
FRANÇOIS-LAVET, Vincent, HENDERSON, Peter, ISLAM, Riashat, BELLEMARE, Marc G. y PINEAU, Joelle, «An Introduction to Deep Reinforcement Learning», Foundations and Trends in Machine Learning, vol. 11, núm. 3-4, 2018, pp. 219-354.
HU, Junling y WELLMAN, Michael P., «Nash Q-Learning for General-Sum Stochastic Games», Journal of Machine Learning Research, vol. 4, 2003, pp. 1039-1069.
JABLONSKIS, Martynas, «Concerted Practices: Concept and Evolution», International Comparative Jurisprudence, vol. 8, núm. 1, 2022, pp. 13-25.
LITTMAN, Michael L., «Markov games as a framework for multi-agent reinforcement learning», en COHEN, William W. y HIRSH, Haym (eds.), Proceedings of the Eleventh International Conference on Machine Learning (ICML 1994), Morgan Kaufmann, San Francisco, CA, 1994, pp. 157-163.
OLIVEIRA E COSTA, Guilherme, «Algorithmic “Collusion” and the Application of Article 101 TFEU: Can an Agreement Exist without a Concurrence of Wills?», Market and Competition Law Review, vol. 10, núm. 1, 2026, pp. 103-137.
ONG, Burton, «The Applicability of Art. 101 TFEU to Horizontal Algorithmic Pricing Practices: Two Conceptual Frontiers», IIC - International Review of Intellectual Property and Competition Law, vol. 52, núm. 2, 2021, pp. 189-211.
PAUDEL, Diwas y DAS, Tapas K., «Tacit Algorithmic Collusion in Deep Reinforcement Learning Guided Price Competition: A Study Using EV Charge Pricing Game», Engineering Applications of Artificial Intelligence, vol. 166, pt. A, 2026, artículo 113567.
PERALTA FIERRO, Ignacio, «La distinción entre acuerdos y prácticas concertadas (un análisis a partir del “caso supermercados”)», Revista Ius et Praxis, vol. 29, núm. 1, 2023, pp. 145-164.
PROTOPAPA, Aikaterini A., «Beyond Agreement: Algorithmic Pricing and the Structural Limits of Antitrust: Autonomous Learning, Tacit Coordination, and Article 101 TFEU», Working Paper, Université Catholique de Lille y Université Toulouse 1 Capitole, 2025/2026, pp. 1-37.
SCHLECHTINGER, Michael, Investigating, Predicting, and Mitigating Collusive Behavior in Deep Reinforcement Learning-Based Pricing AIs, tesis doctoral, Universität Mannheim, 2024.
SCHLECHTINGER, Michael, KOSACK, Damaris, KRAUSE, Franz y PAULHEIM, Heiko, «By Fair Means or Foul: Quantifying Collusion in a Market Simulation with Deep Reinforcement Learning», en Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence (IJCAI-24), 2024, pp. 485-493.
SHAPLEY, L. S., «Stochastic Games», Proceedings of the National Academy of Sciences of the United States of America, vol. 39, núm. 10, 1953, pp. 1095-1100.
SUTTON, Richard S. y BARTO, Andrew G., Reinforcement Learning: An Introduction, 2ª ed., MIT Press, Cambridge, MA, 2014.
TESAURO, Gerald y KEPHART, Jeffrey O., «Pricing in agent economies using multi-agent Q-learning», en Proceedings of the 1999 Workshop on Agent-Mediated Electronic Commerce (AMEC-99), IBM T. J. Watson Research Center, 1999.
UWANOBA, Fumito, «Implicit Cooperative Learning on Distribution of Received Reward in Multi-Agent System», en Proceedings of the 15th International Conference on Agents and Artificial Intelligence (ICAART 2023), vol. 1, SCITEPRESS, 2023, pp. 147-153.
VAN CLEYNENBREUGEL, Pieter, «Article 101 TFEU’s Association of Undertakings Notion and Its Surprising Potential to Help Distinguish Acceptable from Unacceptable Algorithmic Collusion», The Antitrust Bulletin, vol. 65, núm. 3, 2020, pp. 1-22.
WATKINS, Christopher J. C. H. y DAYAN, Peter, «Q-Learning», Machine Learning, vol. 8, núm. 3-4, 1992, pp. 279-292.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Fernando A. Ramos-Zaga
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Since November 2024 this journal is licensed under CC-BY-NC 4.0
.png){kind=logo}
