## Stochastic Hamiltonian flows with singular coefficients

• AcceptedJun 16, 2017
• PublishedMar 23, 2018
Share
Rating

### Abstract

In this paper, we study the following stochastic Hamiltonian system in ${\mathbb~R}^{2d}$ (a second order stochastic differential equation): where $b(x,\mathrm{v}):{\mathbb~R}^{2d}\to{\mathbb~R}^d$ and $\sigma(x,\mathrm{v}):{\mathbb~R}^{2d}\to{\mathbb~R}^d\otimes{\mathbb~R}^d$ are two Borel measurable functions.We show that if $\sigma$ is bounded and uniformly non-degenerate, and $b\in~H^{2/3,0}_p$ and $\nabla\sigma\in~L^p$ for some $p>2(2d+1)$,where $H^{\alpha,\beta}_p$ is the Bessel potential space with differentiability indices $\alpha$ in $x$ and $\beta$ in $\mathrm{v}$,then the above stochastic equation admits a unique strong solution so that $(x,\mathrm{v})\mapsto~Z_t(x,\mathrm{v}):=(X_t,\dot~X_t)(x,\mathrm{v})$forms a stochastic homeomorphism flow, and $(x,\mathrm{v})\mapsto~Z_t(x,\mathrm{v})$ is weakly differentiable withess.$\sup_{x,\mathrm{v}}{\rm~E}(\sup_{t\in[0,T]}|\nabla~Z_t(x,\mathrm{v})|^q~)<\infty$ for all $q\geqslant~1$ and $T\geqslant~0$. Moreover, we also show the uniqueness ofprobability measure-valued solutions for kinetic Fokker-Planck equations with rough coefficientsby showing the well-posedness of the associated martingale problem and using the superposition principle established by Figalli (2008) and Trevisan (2016).

### Supplement

Appendix

The following stochastic Gronwall's type lemma is probably well-known. Since we cannot find it in the literature, a proof is provided here for the reader's convenience.

Lemma 9. For a given $T>0$, let $(\xi_t)_{t\in~[0,~T]}$ and $(\beta_t)_{t\in~[0,~T]}$ $($resp. $(\alpha_t)_{t\in~[0,~T]})$ be two real-valued $($resp. ${\mathbb~R}^d$-valued$)$ measurable ${\mathscr~F}_t$-adapted processes. Let $\zeta_t$ be an Itô process with the form $$\zeta_t=\zeta_0+\int^t_0\zeta^{(1)}_s{{{d}}} s+\int^t_0\zeta^{(2)}_s{{{d}}} W_s.$$ Suppose that for any $\gamma>0$, \begin{align}\kappa_\gamma:={\rm E}\exp\bigg\{\gamma\int^T_0 (|\beta_s|+|\alpha_s|^2 ){{{d}}} s\bigg\}<\infty, \tag{103} \end{align} and \begin{align} 0\leqslant\xi_t\leqslant\zeta_t+\int^t_0\xi_s\beta_s{{{d}}} s+\int^t_0\xi_s\alpha_s{{{d}}} W_s. \tag{104} \end{align} Then for any $q_0\in[1,\infty)$ and $q_1,q_2,q_3>q_0$, there is a constant $C>0$ only depending on $q_i,\kappa_{\gamma},~i=0,1,2,3$ such that \begin{align}\|\xi^*_T\|_{q_0}\leqslant C\bigg(\|\zeta_0\|_{q_1}+\bigg\|\int^T_0|\zeta_s^{(1)}|{{{d}}} s\bigg\|_{q_2} +\bigg\|\int^T_0|\zeta_s^{(2)}|^2{{{d}}} s\bigg\|^{1/2}_{q_3/2}\bigg), \tag{105} \end{align} where $\xi^*_T:=\sup_{t\in[0,T]}\xi_t$ and $\|\cdot\|_{q_i}$ denotes the norm in $L^{q_i}(\Omega)$.

Proof. Write \begin{align*}\eta_t&:=\zeta_t+\int^t_0\xi_s\beta_s{{{d}}} s+\int^t_0\xi_s\alpha_s{{{d}}} W_s=\zeta_t+\int^t_0\eta_s\bar\beta_s{{{d}}} s+\int^t_0\eta_s\bar\alpha_s{{{d}}} W_s, \end{align*} where $\bar\beta_s:=\xi_s\beta_s/\eta_s$ and $\bar\alpha_s:=\xi_s\alpha_s/\eta_s$. Here, we use the convention $\frac{0}{0}:=0$.

Define $$M_t:=\exp\bigg\{\int^t_0\bar\alpha_s{{{d}}} W_s+\int^t_0(\bar\beta_s-\tfrac{1}{2}|\bar\alpha_s|^2){{{d}}} s\bigg\}.$$ By Itô's formula, we have $$M_t=1+\int^t_0M_s\bar\beta_s{{{d}}} s+\int^t_0M_s\bar\alpha_s{{{d}}} W_s$$ and \begin{align*}\eta_t=M_t\bigg[\zeta_0+\int^t_0 M^{-1}_s(\zeta^{(1)}_s-\langle\bar\alpha_s,\zeta_s^{(2)}\rangle){{{d}}} s+\int^t_0M^{-1}_s \zeta^{(2)}_s{{{d}}} W_s\bigg]. \end{align*} Hence, \begin{align*}\|\eta^*_T\|_{q_0}&\leqslant\|M^*_T\zeta_0\|_{q_0}+\bigg\|M^*_T(M^{-1})^*_T\int^T_0|\zeta^{(1)}_s|{{{d}}} s\bigg\|_{q_0} \\ & +\bigg\|M^*_T(M^{-1})^*_T\int^T_0|\bar\alpha_s|\cdot|\zeta_s^{(2)}|{{{d}}} s\bigg\|_{q_0} \\ & +\bigg\|M^*_T\sup_{t\in[0,T]}\bigg|\int^t_0M^{-1}_s\zeta_s^{(2)}{{{d}}} W_s\bigg|\bigg\|_{q_0} \\ &=:I_1+I_2+I_3+I_4. \end{align*} Noticing that by 104, \begin{align} |\bar\beta_s|\leqslant|\beta_s|, |\bar\alpha_s|\leqslant|\alpha_s|, \tag{106} \end{align} for any $p\in{\mathbb~R}$, by 103, Hölder's inequality and Doob's maximal inequality, we have \begin{align} {\rm E}\Big(\sup_{t\in[0,T]}M_t^p\Big)<\infty. \tag{107} \end{align} Thus, by Hölder's inequality and 107, we have $$I_1\leqslant C\|\zeta_0\|_{q_1}, I_2\leqslant C\bigg\|\int^T_0|\zeta^{(1)}_s|{{{d}}} s\bigg\|_{q_2},$$ and by 106, \begin{align*}I_3&\leqslant\bigg\|M^*_T(M^{-1})^*_T\bigg(\int^T_0|\alpha_s|^2{{{d}}} s\bigg)^{1/2}\bigg(\int^T_0|\zeta_s^{(2)}|^2{{{d}}} s\bigg)^{1/2}\bigg\|_{q_0} \\ &\leqslant C\bigg\|\bigg(\int^T_0|\zeta_s^{(2)}|^2{{{d}}} s\bigg)^{1/2}\bigg\|_{q_3}=C\bigg\|\int^T_0|\zeta_s^{(2)}|^2{{{d}}} s\bigg\|^{1/2}_{q_3/2}. \end{align*} Similarly, by Hölder and Burkholder's inequalities, we also have $$I_4\leqslant C \bigg\|\int^T_0|\zeta_s^{(2)}|^2{{{d}}} s \bigg\|^{1/2}_{q_3/2}.$$ Combining the above estimates, we obtain 105.

Let $f$ be a locally integrable function on ${\mathbb~R}^d$. The Hardy-Littlewood maximal function is defined by $${\mathcal M} f(x):=\sup_{0<r<\infty}\frac{1}{|B_r|}\int_{B_r}f(x+y){{{d}}} y,$$ where $B_r:=\{x\in{\mathbb~R}^{d}:~|x|<r\}$. The following result can be found in [-1].

Lemma 10. rm(i) There exists a constant $C_d>0$ such that for all $f\in~C^\infty({\mathbb~R}^{d})$ and $x,y\in~{\mathbb~R}^{d}$, \begin{align}|f(x)-f(y)|\leqslant C_d |x-y|({\mathcal M}|\nabla f|(x)+{\mathcal M}|\nabla f|(y)). \tag{108} \end{align} rm(ii) For any $p>1$, there exists a constant $C_{d,p}$ such that for all $f\in~L^p({\mathbb~R}^d)$, \begin{align}\|{\mathcal M} f\|_p\leqslant C_{d,p}\|f\|_p. \tag{109} \end{align}

### References

[1] Bass R F, Chen Z Q. Brownian motion with singular drift. Ann Probability, 2003, 31: 791-817 CrossRef Google Scholar

[2] Bergh J, Löfström J. Interpolation Spaces: An Introduction. Berlin: Springer-Verlag, 1976. Google Scholar

[3] Bogachev V, Krylov N, R?ckner M. ON REGULARITY OF TRANSITION PROBABILITIES AND INVARIANT MEASURES OF SINGULAR DIFFUSIONS UNDER MINIMAL CONDITIONS. Commun Partial Differ Equ, 2001, 26: 2037-2080 CrossRef Google Scholar

[4] Bogachev V I, Krylov N V, R?ckner M. Elliptic and parabolic equations for measures. Russ Math Surv, 2009, 64: 973-1078 CrossRef ADS Google Scholar

[5] Bouchut F. Hypoelliptic regularity in kinetic equations. J de Mathématiques Pures Appliquées, 2002, 81: 1135-1159 CrossRef Google Scholar

[6] Bramanti M, Cupini G, Lanconelli E. Global L p estimates for degenerate Ornstein?CUhlenbeck operators. Math Z, 2010, 266: 789-816 CrossRef Google Scholar

[7] Chaudru de Raynal P E. Strong existence and uniqueness for degenerate SDE with H?lder drift. Ann Inst H Poincaré Probab Statist, 2017, 53: 259-286 CrossRef ADS Google Scholar

[8] Chen Z Q, Zhang X. $L^p$-maximal hypoelliptic regularity of nonlocal kinetic Fokker-Planck operators,. arXiv Google Scholar

[9] Cherny A S. On the uniqueness in law and the pathwise uniqueness for stochastic differential equations. Theory Probab Appl, 2006, 46: 483--497. Google Scholar

[10] Crippa G, De Lellis C. Estimates and regularity results for the DiPerna-Lions flow. J Reine Angew Math, 2008, 616: 15--46. Google Scholar

[11] Fedrizzi E, Flandoli F. Noise prevents singularities in linear transport equations. J Funct Anal, 2013, 264: 1329-1354 CrossRef Google Scholar

[12] Fedrizzi E, Flandoli F. H?lder Flow and Differentiability for SDEs with Nonregular Drift. Stochastic Anal Appl, 2013, 31: 708-736 CrossRef Google Scholar

[13] Fedrizzi E, Flandoli F, Priola E, et al. Regularity of stochastic kinetic equations,. arXiv Google Scholar

[14] Figalli A. Existence and uniqueness of martingale solutions for SDEs with rough or degenerate coefficients. J Funct Anal, 2008, 254: 109-153 CrossRef Google Scholar

[15] Jin P. Brownian motion with singular time-dependent drift. J Theoret Probab, 2016, https://doi.org/10.1007/s1095. Google Scholar

[16] Karatza I, Shreve S E. Brownian Motion and Stochastic Calculus. New York: Springer-Verlag, 1988. Google Scholar

[17] Krylov N V. Lectures on Elliptic and Parabolic Equations in Sobolev Spaces. Graduate Studies in Mathematics, vol. 96. Providence: Amer Math Soc, 2008. Google Scholar

[18] Krylov N V, R?ckner M. Strong solutions of stochastic equations with singular time dependent drift. Probab Theor Relat Fields, 2005, 131: 154-196 CrossRef Google Scholar

[19] Kunita H. Stochastic flows and stochastic differential equations. Cambridge Studies in Advanced Mathematics, vol. 24. Cambridge: Cambridge University Press, 1990. Google Scholar

[20] Menoukeu-Pamen O, Meyer-Brandis T, Nilssen T. A variational approach to the construction and Malliavin differentiability of strong solutions of SDE's. Math Ann, 2013, 357: 761-799 CrossRef Google Scholar

[21] Menozzi S. Martingale problems for some degenerate Kolmogorov equations. Stochastic Process Appl, 2017, in press. Google Scholar

[22] Mohammed S E A, Nilssen T K, Proske F N. Sobolev differentiable stochastic flows for SDEs with singular coefficients: Applications to the transport equation. Ann Probab, 2015, 43: 1535-1576 CrossRef Google Scholar

[23] Priola E. On Weak Uniqueness for Some Degenerate SDEs by Global L p Estimates. Potential Anal, 2015, 42: 247-281 CrossRef Google Scholar

[24] R?ckner M, Zhang X. Weak uniqueness of Fokker?CPlanck equations with degenerate and bounded coefficients. Comptes Rendus Mathematique, 2010, 348: 435-438 CrossRef Google Scholar

[25] Soize C. The Fokker-Planck Equation for Stochastic Dynamical Systems and Its Explicit Steady State Solutions. Series on Advances in Mathematics for Applied Sciences, vol. 17. Singapore: World Scientific, 1994. Google Scholar

[26] Stein E M. Singular Integrals and Differentiability Properties of Functions. Princeton: Princeton University Press, 1970. Google Scholar

[27] Stroock D, Varadhan S R S. Multidimensional Diffusion Processes. Berlin: Springer-Verlag, 1997. Google Scholar

[28] Talay D. Stochastic Hamiltonian systems: Exponential convergence to the invariant measure and discretization by the implicit Euler scheme. Markov Process Related Fields, 2002, 8: 1--36. Google Scholar

[29] Trevisan D. Well-posedness of multidimensional diffusion processes with weakly differentiable coefficients. Electron J Probab, 2016, in press. Google Scholar

[30] Wang F Y, Zhang X. Degenerate SDE with H?lder--Dini Drift and Non-Lipschitz Noise Coefficient. SIAM J Math Anal, 2016, 48: 2189-2226 CrossRef Google Scholar

[31] Xie L, Zhang X. Sobolev differentiable flows of SDEs with local Sobolev and super-linear growth coefficients. Ann Probab, 2016, 44: 3661-3687 CrossRef Google Scholar

[32] Zhang X. Strong solutions of SDES with singular drift and Sobolev diffusion coefficients. Stochastic Processes their Appl, 2005, 115: 1805-1818 CrossRef Google Scholar

[33] Zhang X. Stochastic Homeomorphism Flows of SDEs with Singular Drifts and Sobolev Diffusion Coefficients. Electron J Probab, 2011, 16: 1096-1116 CrossRef Google Scholar

[34] Zhang X. Stochastic partial differential equations with unbounded and degenerate coefficients. J Differ Equ, 2011, 250: 1924-1966 CrossRef ADS Google Scholar

[35] Zhang X. Stochastic differential equations with Sobolev diffusion and singular drift and applications. Ann Appl Probab, 2016, 26: 2697-2732 CrossRef Google Scholar

[36] Zvonkin A K. A transformation of the phase space of a diffusion process that removes the drift. Mat Sb, 1974, 93: 129--149. Google Scholar

Citations

Altmetric