DQNAMM

This repository contains experimental work in which deep Q-learning is used to teach an agent to play deterministic market making game. The algorithm seeks to capture bid-ask spread while keeping inventory close to zero.