"Bridging the gap of reinforcement learning for Mean Field Games and Mean Field control problems" by Andrea Angiuli (UCSB)

Event Date: 

Monday, November 9, 2020 - 3:30pm to 4:30pm

A unified reinforcement learning approach is proposed to solve non-cooperative (MFG) and cooperative games (MFC) with large populations of players. The numerical scheme is designed as a two time scale stochastic approximation. We show how its calibration depends strictly on the nature of the problem. Convergence analysis and numerical examples are discussed.