Abstract

Mean field games (MFGs) offer a macroscopic and mathematical tractable framework in studying optimal decision problems involving a large number of agents. In this talk, two modelling applications of MFGs will be presented. In the first part, we consider a linear-quadratic Stackelberg game between two leaders and a large group of followers. The two leaders can either cooperate or compete, which are respectively formulated as a Pareto and a Nash game. The relative merits of the two regimes for the followers will be discussed. In the second part, we consider the problem of optimal insurance strategies of policyholders in a mutual insurance company, which is formulated as an extended MFG. We will look at the well-posedness of the resulting forward-backward stochastic differential equation. We will also briefly discuss the implementation of optimal insurance strategies using a deep neural network approach and examine how different parameters influence these strategies.

 

WebEx Link: https://nationaltaiwanuniversity-ksz.my.webex.com/nationaltaiwanuniversity-ksz.my/j.php?MTID=ma9f9b4f2a4b9c0c333d74770631ccbe9

Meeting number (access code): 2515 947 1472

Meeting password: 3mQKZVjj4g3

 

Organizers: Feng-Bin Wang (CGU), Chang-Hong Wu (NYCU) , Chang-Yuan Cheng (NKNU)