Sequential Propagation of Chaos for Mean-Field BSDE Systems

Abstract

A new class of backward particle systems with sequential interaction is proposed to approximate the mean-field backward stochastic differential equations. It is proven that the weighted empirical measure of this particle system converges to the law of the McKean-Vlasov system as the number of particles grows. Based on the Wasserstein metric, quantitative propagation of chaos results are obtained for both linear and quadratic growth conditions. Finally, numerical experiments are conducted to validate our theoretical results.

Keywords

Backward propagation of chaos / Particle system / Sequential interaction / McKean-Vlasov BSDE / Convergence rate

References

[1] McKean, H. P., Propagation of chaos for a class of non-linear parabolic equations, Stochastic Differential Equations (Lecture Series in Differential Equations, Session 7, Catholic Univ., 1967), 1967, 41–57.

[2] Bossy M, Talay D. A stochastic particle method for the McKean–Vlasov and the Burgers equation. Mathematics of Computation, 1997, 66(217): 157-192

[3] Bossy M, Talay D. Convergence rate for the approximation of the limit law of weakly interacting particles: Application to the burgers equation. Annals of Applied Probability, 1996, 6(3): 818-861

[4] Malrieu F. Convergence to equilibrium for granular media equations and their Euler schemes. Annals of Applied Probability, 2003, 13: 540-560

[5] Carmona R, Delarue F. Probabilistic Theory of Mean Field Games with Applications I–II Probab, 2018, Cham: Springer-Verlag 83-84

[6] Chaintron, L.-P. and Diez, A., Propagation of chaos: A review of models, methods and applications, 2021, arXiv:2106.14812.

[7] Kac M. Foundations of kinetic theory. Proceedings of the Third Berkeley Symposium, on Mathematical Statistics and Probability, 1956, 3: 171-197

[8] Lacker D. On a strong form of propagation of chaos for McKean-Vlasov equations. Electronic Communications in Probability, 2018, 23: 1-11

[9] Sznitman A-S. Topics in propagation of chaos, Ecole d’été de probabilités de Saint-Flour XIX-1989, 1991, Berlin: Springer-Verlag 165-251

[10] Jabin P E, Wang Z F. Quantitative estimates of propagation of chaos for stochastic systems with w _1,∞ kernels. Inventiones Mathematicae, 2018, 214(1): 523-591

[11] Chaintron L-P, Diez A. Propagation of chaos: A review of models, methods and applications, I. Models and methods. Kinetic and Related Models, 2022, 15(6): 895-1015

[12] Proceedings of the Royal Society A, 2021, 477 2245 27pp

[13] dos Reis G, Engelhardt S, Smith G. Simulation of McKean–Vlasov SDEs with super-linear growth. IMA Journal of Numerical Analysis, 2022, 42(1): 874-922

[14] Pardoux E, Peng S G. Adapted solution of a backward stochastic differential equation. Systems & Control Letters, 1990, 14(1): 55-61

[15] El Karoui N, Peng S, Quenez M C. Backward stochastic differential equations in finance. Mathematical Finance, 1997, 7(1): 1-71

[16] Peng S G. Backward stochastic differential equations and applications to optimal control. Applied Mathematics and Optimization, 1993, 27: 125-144

[17] Pardoux É, Tang S J. Forward-backward stochastic differential equations and quasilinear parabolic pdes. Probability Theory and Related Fields, 1999, 114: 123-150

[18] Buckdahn R, Djehiche B, Li J A, Peng S G. Mean-field backward stochastic differential equations: A limit approach. The Annals of Probability, 2009, 37(4): 1524-1565

[19] Buckdahn R, Li J A, Peng S G, Rainer C. Mean-field stochastic differential equations and associated PDEs. The Annals of Probability, 2017, 45(2): 824-878

[20] Laurière M, Tangpi L. Backward propagation of chaos. Electronic Journal of Probability, 2022, 27: 1-30

[21] Ma J, Protter P, Martín J S, Torres S. Numerical method for backward stochastic differential equations. Annals of Applied Probability, 2002, 12(1): 302-316

[22] Zhang J F. A numerical scheme for BSDEs. The Annals of Applied Probability, 2004, 14(1): 459-488

[23] Gobet E, Lemor J-P, Warin X. A regression-based monte carlo method to solve backward stochastic differential equations. Annals of Applied Probability, 2005, 15(3): 2172-2202

[24] Zhao W D, Chen L F, Peng S G. A new kind of accurate numerical method for backward stochastic differential equations. SIAM Journal on Scientific Computing, 2006, 28(4): 1563-1581

[25] Chassagneux J-F, Richou A. Numerical simulation of quadratic BSDEs. Annals of Applied Probability, 2013, 26: 262-304

[26] Chessari, J., Kawai, R., Shinozaki, Y. and Yamada, T., Numerical methods for backward stochastic differential equations: A survey, 2021, arXiv:2101.08936.

[27] Sun D Q, Liang G C, Tang S J. Quantitative stability and numerical analysis of markovian quadratic BSDEs with reflection. Probab Uncertain. Quant Rist, 2022, 7(1): 13-30

[28] Angiuli A, Graves C V, Li H Z Cemracs 2017: umerical probabilistic approach to mfg. ESAIM: Proceedings and Surveys, 2019, 65: 84-113

[29] Chassagneux J-F, Crisan D, Delarue F. Numerical method for FBSDEs of McKean Vlasov type. The Annals of Applied Probability, 2019, 29(3): 1640-1684

[30] Achdou Y, Cardaliaguet P, Delarue F Mean field games and applications: Numerical aspects, 2019, Cetraro, Italy: Mean Field Games 249-307 2020

[31] Lauriere M. Numerical methods for mean field games and mean field type control. Mean field Games, 2021, 78: 221

[32] Ee W N, Han J Q, Jentzen A. Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations. Communications in Mathematics and Statistics, 2017, 5(4): 349-380

[33] Han J Q, Jentzen A, E W N. Solving high-dimensional partial differential equations using deep learning. Proceedings of the National Academy of Sciences, 2018, 115(34): 8505-8510

[34] Fouque, J.-P. and Zhang, Z. Q., Deep learning methods for mean field control problems with delay, Frontiers in Applied Mathematics and Statistics, 2019.

[35] Campbell, S., Chen, Y. C., Shrivats, A. and Jaimungal, S., Deep learning for principal-agent mean field games, 2021, arXiv:2110.01127.

[36] Carmona R, Laurière M. Convergence analysis of machine learning algorithms for the numerical solution of mean-field control and games I: The finite horizon case. The Annals of Applied Probability, 2022, 32(6): 4065-4105

[37] Jiang Y F, Li J F. Convergence of the deep bsde method for fbsdes with non-lipschitz coefficients. Probability, Uncertainty and Quantitative Risk, 2021, 6: 391-408

[38] Germain M, Mikael J, Warin X. Numerical resolution of mckean-vlasov fbsdes using neural networks. Methodology and Computing in Applied Probability, 2022, 24(4): 2557-2586

[39] Du, K., Jiang, Y. F. and Li, X. C., Sequential propagation of chaos, 2023, arXiv:2301.09913.

[40] Horowitz J, Karandikar R L. Mean rates of convergence of empirical measures in the Wasserstein metric. Journal of Computational and Applied Mathematics, 1994, 55(3): 261-273

[41] Fournier N, Guillin A. On the rate of convergence in Wasserstein distance of the empirical measure. Probability Theory and Related Fields, 2015, 162(3): 707-738

[42] Ambrosio L, Stra F, Trevisan D. A PDE approach to a 2-dimensional matching problem. Probability Theory and Related Fields, 2019, 173(1): 433-477

[43] Hao, T., Hu, Y., Tang, S. J. and Wen, J. Q., Mean-field backward stochastic differential equations and nonlocal PDEs with quadratic growth, 2022, arXiv:2211.056763.

[44] Kobylanski M. Backward stochastic differential equations and partial differential equations with quadratic growth. The Annals of Probability, 2000, 28(2): 558-602

[45] Briand P, Hu Y. BSDE with quadratic growth and unbounded terminal value. Probability Theory and Related Fields, 2005, 136: 604-618

[46] Briand P, Hu Y. Quadratic BSDEs with convex generators and unbounded terminal conditions. Probability Theory and Related Fields, 2008, 141: 543-567

[47] Bahlali K, Eddahbi M, Ouknine Y. Quadratic BSDE with ${\mathbb{L}^2}$-terminal data: Krylov’s estimate, Itȏ-Krylov’s formula and existence results. The Annals of Probability, 2017, 45(4): 2377-2397

[48] Hu Y, Tang S J. Multi-dimensional backward stochastic differential equations of diagonally quadratic generators. Stochastic Processes and their Applications, 2016, 126(4): 1066-1086

[49] Fan S J, Hu Y, Tang S J. On the uniqueness of solutions to quadratic BSDEs with non-convex generators and unbounded terminal conditions. Comptes Rendus. Mathmatique, 2020, 358: 227-235

[50] Fan, S. J., Hu, Y. and Tang, S. J., Multi-dimensional backward stochastic differential equations of diagonally quadratic generators: The general result, 2020, arXiv:2007.04481.

[51] Hibon, H., Hu, Y. and Tang, S. J., Mean-field type quadratic BSDEs, 2017, arXiv:1708.08784.

[52] Chassagneux J-F, Crisan D, Delarue F. A probabilistic approach to classical solutions of the master equation for large population equilibria. Memoirs of the American Mathematical Society, 2022, 280(1379): V+123PP

[53] Kazamaki N. Continuous Exponential Martingales and BMO, 1994, Berlin: Springer-Verlag

[54] Longstaff F, Schwartz E S. Valuing American options by simulation: A simple least-squares approach. Review of Financial Studies, 2001, 14(1): 113-147

[55] Bender C, Steiner J. Carmona R A, Del Moral P, Hu P, Oudjane N. Least-squares monte carlo for backward SDEs. Numerical Methods in Finance, 2012, Berlin, Heidelberg: Springer-Verlag 257-289

[56] Zhao W, Fu Y, Zhou T. New kinds of high-order multistep schemes for coupled forward backward stochastic differential equations. SIAM J. Sci. Comput., 2014, 36: A1731-A1751

[57] Yang J, Zhao W D, Zhou T. Explicit deferred correction methods for second-order forward backward stochastic differential equations. Journal of Scientific Computing, 2019, 79: 1409-1432