ABIDES-MARL: A Multi-Agent Reinforcement Learning Environment for Endogenous Price Formation and Execution in a Limit Order Book
Patrick Cheridito † {\dagger}, Jean-Loup Dupret † {\dagger} and Zhexin Wu ∗ ‡ {{\ddagger}} · 2025
Read ABIDES-MARL: A Multi-Agent Reinforcement Learning Environment for Endogenous Price Formation and Execution in a Limit Order Book by Patrick Cheridito † {\dagger}, Jean-Loup Dupret † {\dagger} and Zhexin Wu ∗ ‡ {{\ddagger}} (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-02016.txt.
Description
Patrick Cheridito, Jean-Loup Dupret, and Zhexin Wu build a trading laboratory where liquidity is made by the players, not assumed. Classical optimal execution theory tells a trader how to work a large order against a pre-specified price impact. But when market makers adapt by design, that assumption fails structurally: execution becomes a finite-horizon stochastic game among mixed agents.
Their framework, ABIDES-MARL, extends the ABIDES-Gym market simulator to many reinforcement learning agents with synchronized decision periods that preserve proper information flow and market microstructure. The framework is validated in an extended Kyle setting. There, multiple learning agents recover gradual price discovery. The simulations show a sharp contrast. Strategies learned under adaptive liquidity differ at root from fixed-impact benchmarks. Strategies optimized against exogenous impact perform poorly once market makers adapt, especially when information is unbalanced, because the makers exploit predictable execution flow.
Catalog record
| CREATOR | Patrick Cheridito † {\dagger}, Jean-Loup Dupret † {\dagger} and Zhexin Wu ∗ ‡ {{\ddagger}} |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 86 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-02016