A Control Theoretical Approach to Mean Field Games. Part II: Global Well-Posedness of Master Equations.

Saved in:
Bibliographic Details
Title: A Control Theoretical Approach to Mean Field Games. Part II: Global Well-Posedness of Master Equations.
Authors: Bensoussan, Alain1 (AUTHOR) axb046100@utdallas.edu, Tai, Ho Man2 (AUTHOR) taihoman@link.cuhk.edu.hk, Wong, Tak Kwong3 (AUTHOR) takkwong@szu.edu.cn, Yam, Sheung Chi Phillip4 (AUTHOR) scpyam@sta.cuhk.edu.hk
Source: Applied Mathematics & Optimization. Apr2026, Vol. 93 Issue 2, p1-56. 56p.
Abstract: In the second part of our work, we aim to establish the global-in-time well-posedness of classical solution of the master equations associated with general mean field games studied in Part I, which is beyond the specific linear-quadratic setting, provided the mean field sensitivity effect is not too large. We characterize the gradient of the value function by the backward process of the forward-backward stochastic differential equations (FBSDEs) introduced in Part I. Then we study the higher regularity of Jacobian flows of the FBSDEs in the state and measure variables so as to establish classical well-posedness of the master equation on. As far as we know, it is the first work to investigate the master equations, with general cost functions having quadratic growth and allowing non-convexity in the state variable, under the small mean field effect. Our current approach directly imposes the structural assumptions (most notably, the small mean field sensitivity effect) on the cost functions, which provides the following advantages: (i) the structural conditions imposed in this work are easily verified and less demanding on the assumptions of the cost functions; (ii) we illustrate how the displacement monotonicity should be formulated when the assumptions are imposed on the cost functions instead of the Hamiltonian; and (iii) we provide an accurate lifespan, which may not be that small in many circumstances, for the local-in-time existence when the mean field sensitivity effect is relatively large, the cost functions are not convex in the state variable or we do not have the monotonicity of cost functions. [ABSTRACT FROM AUTHOR]
Copyright of Applied Mathematics & Optimization is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:In the second part of our work, we aim to establish the global-in-time well-posedness of classical solution of the master equations associated with general mean field games studied in Part I, which is beyond the specific linear-quadratic setting, provided the mean field sensitivity effect is not too large. We characterize the gradient of the value function by the backward process of the forward-backward stochastic differential equations (FBSDEs) introduced in Part I. Then we study the higher regularity of Jacobian flows of the FBSDEs in the state and measure variables so as to establish classical well-posedness of the master equation on. As far as we know, it is the first work to investigate the master equations, with general cost functions having quadratic growth and allowing non-convexity in the state variable, under the small mean field effect. Our current approach directly imposes the structural assumptions (most notably, the small mean field sensitivity effect) on the cost functions, which provides the following advantages: (i) the structural conditions imposed in this work are easily verified and less demanding on the assumptions of the cost functions; (ii) we illustrate how the displacement monotonicity should be formulated when the assumptions are imposed on the cost functions instead of the Hamiltonian; and (iii) we provide an accurate lifespan, which may not be that small in many circumstances, for the local-in-time existence when the mean field sensitivity effect is relatively large, the cost functions are not convex in the state variable or we do not have the monotonicity of cost functions. [ABSTRACT FROM AUTHOR]
ISSN:00954616
DOI:10.1007/s00245-025-10319-6