Linear-quadratic extended mean field games with common noises

Tianjiao Hua,Peng Luo

2023-11-07

Abstract:In this paper, we consider a class of linear quadratic extended mean field games (MFGs) with common noises where the state coefficients and the cost functional vary with the mean field term in a nonlinear way. Based on stochastic maximum principle, solving the mean field game is transformed into solving a conditional mean field forward-backward stochastic differential equation (FBSDE). We first establish solvability for a type of (more general) conditional mean field FBSDEs under monotonicity conditions. We further provide some regularity results which lead to classical solutions for the associated master equations. In particular, the linear quadratic extended mean field game is solved and classical solution for (extended mean field game) master equation is obtained.

Optimization and Control,Probability

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to establish a unique equilibrium solution in linear - quadratic Extended Mean Field Games (EMFGs) with common noise. Specifically, the paper focuses on this type of EMFGs: the state coefficients and cost functions therein depend on the mean - field term in a nonlinear manner. By using the stochastic maximum principle and the consistency condition, the author transforms the extended mean - field game problem into solving a Conditional Mean Field Forward - Backward Stochastic Differential Equation (CMF - FBSDE). The main contributions of the paper are as follows:

1. **Prove the existence and uniqueness of CMF - FBSDE (1.1)**: This includes the Hamiltonian system generated by the linear - quadratic extended mean - field games with common noise.
2. **Obtain the global classical solution of the relevant master equation**: These master equations include the master equations of the extended mean - field games with common noise.

To achieve this goal, the author proposes a new monotonicity condition, which not only guarantees the unique equilibrium solution of the extended mean - field games with common noise, but also allows to obtain the global well - posedness of the master equation related to the non - separable Hamiltonian. In addition, the author's method is more general in dealing with the interaction between state and control, removes the restrictions on state dynamics, and does not need to assume the boundedness of the diffusion process.

### Key steps in the paper

1. **Fix the conditional expectation term**: First, fix the conditional expectation term in CMF - FBSDE to obtain the global solvability of a classical FBSDE.
2. **Introduce a mapping**: Then introduce a mapping to transform the fixed - point - seeking problem into solving another classical FBSDE. In these two steps, the author makes full use of the method proposed in reference [25] to ensure the well - posedness of the corresponding FBSDE.

### Novelty

Compared with the existing literature, the main novelties of this paper are as follows:

- **More general form**: Consider more general forms of dynamic and running cost functions, not just the non - degenerate case.
- **Remove boundedness and non - degenerate assumptions**: By adding some monotonicity conditions, remove the boundedness and non - degenerate assumptions on the coefficients.
- **Deal with the interaction of control**: Consider not only the interaction of state, but also the interaction of control.

These innovations make the paper more flexible and powerful in dealing with the extended mean - field game problems with common noise.