Optimal execution with an RL approach. Test with China's stock data.

RL agents for minimizing Implementation shortfall. Outperforms Almgren-Chriss by dynamically adapting to market drift, volatility, and microstructure impacts.

RL agents for minimizing Implementation shortfall. Outperforms Almgren-Chriss by dynamically adapting to market drift, volatility, and microstructure impacts.

Deep Reinforcement Learning for Optimal Trade Execution using DQN and Baseline Strategy Comparison

Optimal Trade Execution Research Platform — Almgren-Chriss, Obizhaeva-Wang & PPO Deep RL with live React dashboard. UCL MSc Business Analytics. Calibrated to LOBSTER AAPL order book data.

Hierarchichal RL framework for intraday optimal trade execution. Implements a two-layer architecture: Strategic PPO agent selects execution pace, Tactical DQN agent optimises order slicing and timing All experiments conducted on simulated market data with reproducible seeds for controlled benchmarking. Outperforms traditional algorithms.

Optimal algorithmic trade execution with RL: Almgren–Chriss, TWAP/VWAP, and a PPO agent that learns atop the Almgren–Chriss framework inside a stochastic impact environment with hard liquidation constraints; CVaR/mean costs benchmarked end-to-end, with Optuna tuning plus reproducible MLflow-logged metrics, artifacts, and pipeline.