Optimal Portfolio Liquidation

Yuheng Cai, Taichen FAN

Keywords: Almgren and Chriss model, Optimal Liquidation Problem, Deep Reinforcement Learning

Abstract: Optimal Portfolio Liquidation is one of the common challenges in stock trading whose aim is to minimize a combination of risks and transactions cost arising from permanent and temporary market impact when selling a large number of stocks within a given time frame. Our attempt in this project is to turn this challenge into a reinforcement learning problem and compare the research result with a known baseline method achieved with Almgren and Chriss model. A quick introduction video is here https://youtu.be/dBjzqcKjlBQ