Reinforcement Learning in Multi-agent Systems for Quantative Finance

PhD project focuses on learning in multi-agent systems within financial contexts. One approach to this problem uses game theory to analyze equilibrium points in multi-agent setups. Another approach involves optimal learning, where much research has focused on stationary learning.

However, real-world interactions between agents often lead to unpredictable dynamics that rarely converge to theoretical stationary points.

My research aims to develop algorithms that learn and perform effectively in non-stationary settings and thus also work well in multi-agent environments.