Long time behavior of the master equation in mean field game theory. (English) Zbl 1428.35607
Summary: Mean field game (MFG) systems describe equilibrium configurations in games with infinitely many interacting controllers. We are interested in the behavior of this system as the horizon becomes large, or as the discount factor tends to 0. We show that, in these two cases, the asymptotic behavior of the mean field game system is strongly related to the long time behavior of the so-called master equation and to the vanishing discount limit of the discounted master equation, respectively. Both equations are nonlinear transport equations in the space of measures. We prove the existence of a solution to an ergodic master equation, towards which the time-dependent master equation converges as the horizon becomes large, and towards which the discounted master equation converges as the discount factor tends to 0. The whole analysis is based on new estimates for the exponential rates of convergence of the time-dependent and the discounted MFG systems, respectively.
MSC:
35Q89 | PDEs in connection with mean field game theory |
35B40 | Asymptotic behavior of solutions to PDEs |
35F21 | Hamilton-Jacobi equations |
35K51 | Initial-boundary value problems for second-order parabolic systems |
37K55 | Perturbations, KAM theory for infinite-dimensional Hamiltonian and Lagrangian systems |
91A16 | Mean field games (aspects of game theory) |
Keywords:
weak KAM theory; nonlinear transport equations; ergodic master equation; time-dependent master equationReferences:
[1] | 10.1137/S0036141099350869 · Zbl 0960.70015 · doi:10.1137/S0036141099350869 |
[2] | 10.1007/s13235-013-0091-x · Zbl 1314.91043 · doi:10.1007/s13235-013-0091-x |
[3] | 10.1051/cocv/2014044 · Zbl 1319.35273 · doi:10.1051/cocv/2014044 |
[4] | 10.3934/nhm.2012.7.279 · Zbl 1270.35098 · doi:10.3934/nhm.2012.7.279 |
[5] | 10.1137/120904184 · Zbl 1332.35364 · doi:10.1137/120904184 |
[6] | ; Cardaliaguet, The master equation and the convergence problem in mean field games. Annals of Mathematical Studies, 201 (2019) · Zbl 1430.91002 |
[7] | 10.1007/s00222-016-0648-6 · Zbl 1362.35094 · doi:10.1007/s00222-016-0648-6 |
[8] | 10.1016/S0764-4442(97)84777-5 · Zbl 0943.37031 · doi:10.1016/S0764-4442(97)84777-5 |
[9] | 10.1016/S0764-4442(97)87883-4 · Zbl 0885.58022 · doi:10.1016/S0764-4442(97)87883-4 |
[10] | 10.1016/j.jde.2015.08.001 · Zbl 1359.35221 · doi:10.1016/j.jde.2015.08.001 |
[11] | ; Gilbarg, Elliptic partial differential equations of second order. Grundlehren der Mathematischen Wissenschaften, 224 (1977) · Zbl 0361.35003 |
[12] | 10.1515/ACV.2008.012 · Zbl 1181.37090 · doi:10.1515/ACV.2008.012 |
[13] | 10.1016/j.matpur.2009.10.010 · Zbl 1192.91028 · doi:10.1016/j.matpur.2009.10.010 |
[14] | ; Huang, Commun. Inf. Syst., 6, 221 (2006) |
[15] | 10.1016/j.matpur.2016.10.013 · Zbl 1375.35182 · doi:10.1016/j.matpur.2016.10.013 |
[16] | ; Ladyženskaja, Linear and quasilinear equations of parabolic type. Translations of Mathematical Monographs, 23 (1968) · Zbl 0174.15403 |
[17] | 10.1016/j.crma.2006.09.019 · Zbl 1153.91009 · doi:10.1016/j.crma.2006.09.019 |
[18] | 10.1016/j.crma.2006.09.018 · Zbl 1153.91010 · doi:10.1016/j.crma.2006.09.018 |
[19] | 10.1007/s11537-007-0657-8 · Zbl 1156.91321 · doi:10.1007/s11537-007-0657-8 |
[20] | ; Le, Dynamical and geometric aspects of Hamilton-Jacobi and linearized Monge-Ampère equations—VIASM 2016. Lecture Notes in Mathematics, 2183 (2017) · Zbl 1373.35006 |
[21] | 10.1016/j.aim.2016.10.032 · Zbl 1357.35090 · doi:10.1016/j.aim.2016.10.032 |
[22] | 10.1080/03605309908821451 · Zbl 0924.35028 · doi:10.1080/03605309908821451 |
[23] | 10.1137/130907239 · Zbl 1287.49006 · doi:10.1137/130907239 |
[24] | 10.1016/S0764-4442(97)89468-2 · Zbl 0924.70017 · doi:10.1016/S0764-4442(97)89468-2 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.