Reinforcement Learning Optimizes Limit-Order Execution Using Logistic-Normal Allocation

An arXiv preprint (v2 posted Jan 26, 2026) introduces a reinforcement learning framework for optimal trade execution in limit order books, modeling market and limit order allocations with multivariate logistic-normal distributions. The paper trains RL policies to maximize expected revenue and reports outperformance versus traditional benchmark strategies in simulated environments with noise, tactical, and strategic traders.