Efficient Kernelized Learning in Polyhedral Games Beyond Full-Information: From Colonel Blotto to Congestion Games
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909832861188096 |
|---|---|
| author | Kontogiannis, Andreas Pollatos, Vasilis Farina, Gabriele Mertikopoulos, Panayotis Panageas, Ioannis |
| author_facet | Kontogiannis, Andreas Pollatos, Vasilis Farina, Gabriele Mertikopoulos, Panayotis Panageas, Ioannis |
| contents | We examine the problem of efficiently learning coarse correlated equilibria (CCE) in polyhedral games, that is, normal-form games with an exponentially large number of actions per player and an underlying combinatorial structure. Prominent examples of such games are the classical Colonel Blotto and congestion games. To achieve computational efficiency, the learning algorithms must exhibit regret and per-iteration complexity that scale polylogarithmically in the size of the players' action sets. This challenge has recently been addressed in the full-information setting, primarily through the use of kernelization. However, in the case of the realistic, but mathematically challenging, partial-information setting, existing approaches result in suboptimal and impractical runtime complexity to learn CCE. We tackle this limitation by building a framework based on the kernelization paradigm. We apply this framework to prominent examples of polyhedral games -- namely the Colonel Blotto, graphic matroid and network congestion games -- and provide computationally efficient payoff-based learning algorithms, which significantly improve upon prior works in terms of the runtime for learning CCE in these settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_20919 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Efficient Kernelized Learning in Polyhedral Games Beyond Full-Information: From Colonel Blotto to Congestion Games Kontogiannis, Andreas Pollatos, Vasilis Farina, Gabriele Mertikopoulos, Panayotis Panageas, Ioannis Computer Science and Game Theory We examine the problem of efficiently learning coarse correlated equilibria (CCE) in polyhedral games, that is, normal-form games with an exponentially large number of actions per player and an underlying combinatorial structure. Prominent examples of such games are the classical Colonel Blotto and congestion games. To achieve computational efficiency, the learning algorithms must exhibit regret and per-iteration complexity that scale polylogarithmically in the size of the players' action sets. This challenge has recently been addressed in the full-information setting, primarily through the use of kernelization. However, in the case of the realistic, but mathematically challenging, partial-information setting, existing approaches result in suboptimal and impractical runtime complexity to learn CCE. We tackle this limitation by building a framework based on the kernelization paradigm. We apply this framework to prominent examples of polyhedral games -- namely the Colonel Blotto, graphic matroid and network congestion games -- and provide computationally efficient payoff-based learning algorithms, which significantly improve upon prior works in terms of the runtime for learning CCE in these settings. |
| title | Efficient Kernelized Learning in Polyhedral Games Beyond Full-Information: From Colonel Blotto to Congestion Games |
| topic | Computer Science and Game Theory |
| url | https://arxiv.org/abs/2509.20919 |