Smoothing properties of McKean-Vlasov SDEs

derivative with respect to measure Kusuoka-Stroock functions Malliavin calculus McKean-Vlasov stochastic differential equation uniformly elliptic coefficients Wiener space

Mathematics Subject Classification ID

Ordinary differential equations and systems with randomness (34F05) Nonlinear parabolic equations (35K55) Degenerate parabolic equations (35K65) Stochastic calculus of variations and the Malliavin calculus (60H07) Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Applications of stochastic analysis (to PDEs, etc.) (60H30)

Abstract: In this article, we develop integration by parts formulae on Wiener space for solutions of SDEs with general McKean-Vlasov interaction and uniformly elliptic coefficients. These integration by parts formulae hold both for derivatives with respect to a real variable and derivatives with respect to a measure understood in the sense of Lions. They allows us to prove the existence of a classical solution to a related PDE with irregular terminal condition. We also develop bounds for the derivatives of the density of the solutions of McKean-Vlasov SDEs.

Review text

This paper deals with the McKean-Vlasov stochastic differential equation \[ X_{t}^{x,[\theta]}=x+\int_0^ t V_0(X_{s}^{x,[\theta]},[X_{s}^{\theta}])\,ds+\sum_{i=1}^{d}\int_0^ t V_{i}(X_{s}^{x,[\theta]},[X_{s}^{\theta}])\,dB_{i}(s), \] where \(B_{i}(t), i=\overline{1,d}\) are the Brownian motions, the coefficients \(V_{i}: \mathbb{R}^{N}\times P_2(\mathbb{R}^{N})\to \mathbb{R}^{N}\), \(i=\overline{0,N}\). Here, \([\xi]\) is the law of the random variable \(\xi\) and \(P_2(\mathbb{R}^{N})\) is the set of probability measures on \(\mathbb{R}^{N}\) with finite second moment. Some results on differentiability of \(X_{t}^{x,[\theta]}\) and the solution of the considered equation with respect to the parameters \((x,[\theta])\) are proved. It is proved that if \(V_0,\ldots,V_{d}\in C^{k,k}_{b,\mathrm{Lip}}(\mathbb{R}^{N}\times P_2(\mathbb{R}^{N});\mathbb{R}^{N})\), then \((t,x,[\theta])\to X_{t}^{x,[\theta]}\in {\mathbf K}^1_0(\mathbb{R}^{N},k)\), where \(C^{k,k}_{b,\mathrm{Lip}}(\mathbb{R}^{N}\times P_2(\mathbb{R}^{N});\mathbb{R}^{N})\) is the class of functions \(k\)-times differentiable with bounded Lipschitz derivatives, \({\mathbf K}^1_0(\mathbb{R}^{N},k)\) is the class of Kusuoka-Stroock processes. Then, the authors develop an integration by parts formulae for derivatives of \((x,[\theta])\to {\mathbf E}[f(X_{t}^{x,[\theta]})]\) under the uniform ellipticity assumption and the assumption that \(V_0,\ldots,V_{d}\in C^{k,k}_{b,\mathrm{Lip}}(\mathbb{R}^{N}\times P_2(\mathbb{R}^{N});\mathbb{R}^{N})\). The existence of a classical solution to a related partial differential equation with an irregular terminal condition is proved, and bounds for the derivatives of the density of the solutions to the McKean-Vlasov stochastic differential equation are obtained.

Recommendations

Cites work

Cited in

(63)

This page was built for publication: Smoothing properties of McKean-Vlasov SDEs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1647925)