Metraipolis › archive › dc_e7c2067bee41
Mitigating Retaliatory Algorithmic Collusion in Repeated Games — arXiv:2609.20548
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.20548
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-10T03:24:21.114Z
- SHA-256 of the archived text
09d56a246057e2053336f2c33804b00f8567dfa0fa941f4095be679be92b0a99- Stable URL of this page
https://metraipolis.com/documento/dc_e7c2067bee41
How to cite
Mitigating Retaliatory Algorithmic Collusion in Repeated Games — arXiv:2609.20548 — arXiv (en), https://arxiv.org/abs/2609.20548 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e7c2067bee41 (Metraipolis archive dc_e7c2067bee41, sha256 09d56a246057e2053336f2c33804b00f8567dfa0fa941f4095be679be92b0a99)
BibTeX
@misc{dce7c2067bee41,
title = {Mitigating Retaliatory Algorithmic Collusion in Repeated Games — arXiv:2609.20548},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.20548},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e7c2067bee41; sha256 09d56a246057e2053336f2c33804b00f8567dfa0fa941f4095be679be92b0a99},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.20548. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reinforcement learning agents trained to maximize their own reward in repeated interactions can converge to supra-competitive outcomes resembling explicit collusion, without communication or shared design. Existing mitigation approaches are largely tied to specific economic settings, like two-sided platforms and auctions, leaving open how to design interventions for general repeated games. We address this gap by formalizing the connection between empirical observations from prior work on Q-learning collusion and classical theory of Simple Penal Codes (SPCs). We show any non-trivial SPC induces a quantifiable conditional dependence in agents' policies, detectable via the total variation distance between an agent's action distributions across cooperation and defection histories. Building on this connection, we propose CURB (Collusion Unwinding via Reward shaping and Belief injection), a reward-shaping framework that penalizes this Total Variation (TV) distance signal during Q-learning and is guaranteed to convert any SPC fixed point of the dynamics into a trivial one, thus precluding collusive equilibria sustained by punishment threats. Empirically, CURB substantially reduces collusion by Q-learning agents in both Bertrand and Cournot Competition Repeated Games. We further demonstrate that CURB extends to deep Q-network agents in Bertrand competition, suggesting the mechanism generalizes beyond tabular Q-learning. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.20548 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-10T03:04:13.970Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).