Existence, exponential mixing and convergence of periodic measures of fractional stochastic delay reaction-diffusion equations on \(\mathbb{R}^n\) (Q2168031): Difference between revisions

From MaRDI portal
Import240304020342 (talk | contribs)
Set profile property.
Set OpenAlex properties.
Property / full work available at URL
 
Property / full work available at URL: https://doi.org/10.1016/j.jde.2022.07.026 / rank
 
Normal rank
Property / OpenAlex ID
 
Property / OpenAlex ID: W4288679914 / rank
 
Normal rank

Revision as of 18:51, 19 March 2024

scientific article
Language Label Description Also known as
English
Existence, exponential mixing and convergence of periodic measures of fractional stochastic delay reaction-diffusion equations on \(\mathbb{R}^n\)
scientific article

    Statements

    Existence, exponential mixing and convergence of periodic measures of fractional stochastic delay reaction-diffusion equations on \(\mathbb{R}^n\) (English)
    0 references
    0 references
    0 references
    31 August 2022
    0 references
    The authors consider the equation \begin{align*} du+(-\Delta)^\alpha u\,dt+\lambda u&=f(t,x,u(t-\rho(t)))\,dt+\varepsilon(\sigma_1(t,x)+g(x)\sigma_2(t,u(t-\rho(t))))\,dW(t) \\ u(s)&=\varphi(s),\qquad s\in[t_0-\bar\rho,0] \end{align*} on \(\mathbb R^n\). Denote the space of Hilbert-Schmidt operators from \(X\) to \(Y\) by \(\operatorname{HS}(X,Y)\). In the above equation we have the following: \(t_0\ge 0\), \(\alpha\in(0,1)\), \(\bar\rho>0\), \(\lambda>0\), \(\rho\in C^1(\mathbb R,[0,\bar\rho])\) is such that \(\sup\rho^\prime<1\), \(\varepsilon\in[0,1]\), \(g\in H^\alpha(\mathbb R^n)\cap L^\infty(\mathbb R^n)\) with \(\nabla g\in L^\infty(\mathbb R^n)\), \(W\) is a cylindrical Wiener process on \(\ell_2\), \(\rho\), \(f\), \(\sigma_1\) and \(\sigma_2\) are \(\mathcal T\)-periodic for some \(\mathcal T\in(0,\infty)\), \(\sigma_1\) is a bounded mapping from \(\mathbb R\) to \(\operatorname{HS}(\ell_2,H^\alpha(\mathbb R^n))\) such that \[ \lim_{r\to\infty}\sup_t\|1_{[|x|>r]}\sigma(t)\|_{\operatorname{HS}(\ell_2,L^2(\mathbb R^n))}=0, \] \(f\) is jointly continuous, globally Lipschitz continuous and linearly growing in the third variable, and \(\sigma_2\) is \(\ell_2\)-valued, jointly continuous, globally Lipschitz continuous and linearly growing in the second variable. Let us define \(C_H=C([-\bar\rho,0];L^2(\mathbb R^n))\) and \(C_V=C([-\bar\rho,0];H^\alpha(\mathbb R^n))\). The authors prove that if \(\varphi\in L^2(\Omega;C_H)\) is \(\mathcal F_0\)-measurable then there exists a unique solution \(u^\varepsilon(t;t_0,\varphi)\) of the above equation with paths that are continuous in \(L^2(\mathbb R^n)\) and locally \(L^2\)-integrable in \(H^\alpha(\mathbb R^n)\). Moreover, if \(\lambda\) is larger than a given threshold \(\lambda_1\) then the \(C_H\)-valued segment process \[ u_t^\varepsilon(t_0,\varphi)(s)=u^\varepsilon(t+s;t_0,\varphi) \] defines a two-parameter Feller system \(P^\varepsilon_{r,t}\) on \(C_H\) with the transition function \[ p^\varepsilon(r,\varphi;t,\Gamma)=\mathbb P\,[u^\varepsilon_t(r,\varphi)\in\Gamma], \] and there exists a periodic Borel probability measure \(\nu\) on \(C_H\) such that \[ (P^\varepsilon_{0,t+\mathcal T})^*\nu=(P^\varepsilon_{0,t})^*\nu,\qquad t\ge 0. \] Let us denote the set of periodic Borel probability measures on \(C_H\) by \(\mathcal S^\varepsilon\). If \(\varepsilon_m\to\varepsilon\) and \(\mu^{\varepsilon_m}\in\mathcal S^{\varepsilon_m}\) then it is proved that there exists a subsequence \(m_k\) and \(\mu^{\varepsilon}\in\mathcal S^{\varepsilon}\) such that \(\mu^{\varepsilon_{m_k}}\to\mu^{\varepsilon}\) weakly. If \(\lambda\) is larger than a given threshold \(\lambda_2\) then the authors show that \(\mathcal S^\varepsilon=\{\mu^\varepsilon\}\) for every \(\varepsilon\in[0,1]\), and if \(\nu\) is a Borel probability measure on \(C_H\) with the finite second moment then \[ \|(P^\varepsilon_{0,k\mathcal T})^*\nu-\mu^\varepsilon\|\le C_\nu e^{-\eta k},\qquad k\in\mathbb N , \] where the Wasserstein metric is considered on the left-hand-side of the above estimate and \(C>0\) and \(\eta>0\) are independent of \(\varepsilon\in[0,1]\).
    0 references
    fractional reaction-diffusion equation
    0 references
    periodic measure
    0 references
    invariant measure
    0 references
    exponential mixing
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references