Uniform and optimal estimates for solutions to singularly perturbed parabolic equations (Q2644652): Difference between revisions

From MaRDI portal
RedirectionBot (talk | contribs)
Removed claim: author (P16): Item:Q855380
RedirectionBot (talk | contribs)
Changed an Item
Property / author
 
Property / author: Denis R. Akhmetov / rank
 
Normal rank

Revision as of 07:39, 21 February 2024

scientific article
Language Label Description Also known as
English
Uniform and optimal estimates for solutions to singularly perturbed parabolic equations
scientific article

    Statements

    Uniform and optimal estimates for solutions to singularly perturbed parabolic equations (English)
    0 references
    0 references
    0 references
    3 September 2007
    0 references
    Let \(L\) be the operator defined by \[ Lv=v_t+k(x,y,t)v_y-a(x,y,t)v_{xx}-b(x,y,t)v_x-c(x,y,t)v \] with \(a(x,y,t) \geq a_o>0\). This paper deals with the boundary value problem for the equation \[ Lv=f(x,y,t) \tag{1} \] in \(\mathcal Q_T=\{(x,y,t)\in \mathbb R\times[0,1]\times [0,T]\}\), with the y-periodic boundary condition \(v| _{y=0}=v| _{y=1}\) and the initial data \(v(x,y,0)=\varphi(x,y).\) The function \(f\), and the coefficients of \(L\) [resp. the initial data \(\varphi\)] are defined on \(\mathbb R\times \mathbb R\times [0,T]\) [resp. \(\mathbb R^{2}],\) smooth enough, and periodic in \(y\) with the unit period. Equation (1) may be interpreted as an ultraparabolic equation with the time variables \(t_1=y,t_2=t\), or as a parabolic equation, with space variables \(x,y,\) degenerate with respect to \(y.\) Under suitable assumptions on the coefficients of (1) and on \(f,\varphi\), the authors prove the existence of a global in time solution of this problem, which belongs to the anisotropic Sobolev space \(W^{3,2,1}_2(\mathcal Q_{T})\) as well as the anisotropic Hölder space \(C^{\lambda,\lambda,\frac{1}{12}}(\mathcal Q_T\cap\{x\in[-K,K]\})\) for \(\lambda\in (0,1)\) and \( K>0\). Moreover , for every \( P\geq 0, \) [resp. for every m=0,1,\dots] \(v\) can be estimated as \(| v(x,y,t)| \leq C_Pe^{-P| x| }\) , [resp. \(| v(x,y,t)| \leq \frac{C_m}{1+| x| ^m}]\), in \(\mathcal Q_T\) , according to the assumptions on the growth of the coefficients in \(L\) . The proof is based on the obtention of \(\varepsilon\) uniform estimates for the solution \(u^{\varepsilon}\) to the regularized parabolic problem \[ Lu^{\varepsilon} -\varepsilon u_{yy}^{\varepsilon} =f(x,y,t)\quad \text{in } \mathcal Q_T \] \[ (u^{\varepsilon},u_y^{\varepsilon})| _{y=0}=(u^{\varepsilon},u_y^{\varepsilon})| _{y=1},\quad u^{\varepsilon}(x,y,0)=\varphi(x,y,) \] An application of these results to the solution \(\rho(\theta,\omega,t,\Omega)\) to the Fokker-Plank type equation [cf. \textit{M. M. Lavrentiev jun.} and the authors [Sib. Mat. Zh. 42, No. 4, 825--848 (2001; Zbl 0983.35075)] yields to the estimate \(| \rho(\theta,\omega,t,\Omega)| \leq Ce^{-M\omega^{2}}.\) This paper completes the results obtained recently by the authors and \textit{M. M. Lavrentiev}, in particular in [Asympt. Anal. 35, No. 1, 65--89 (2003; Zbl 1043.35023), Differ. Integral Equ., 17, No. 1--2, 99--118 (2004; Zbl 1164.35312)] and loc.cit.
    0 references
    0 references
    Fokker-Planck equation
    0 references
    unbounded coefficients
    0 references
    boundary-layer
    0 references
    anisotropic Sobolev space
    0 references
    anisotropic Hölder space
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references