<p>We investigate the limiting behavior of invariant measures in terms of weak convergence and the Wasserstein metric of probability measures for a stochastic fractional Ginzburg-Landau equation defined on the entire space <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\mathbb {R}^{n}\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mi>n</mi> </msup> </math></EquationSource> </InlineEquation>, where the growth rate <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(2\beta + 1\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>2</mn> <mi>β</mi> <mo>+</mo> <mn>1</mn> </mrow> </math></EquationSource> </InlineEquation> and the coefficient <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\varpi \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>ϖ</mi> </math></EquationSource> </InlineEquation> of the drift term satisfy an <i>admissible</i> condition <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(|\varpi | \le \sqrt{2\beta + 1}/\beta \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mrow> <mo stretchy="false">|</mo> <mi>ϖ</mi> <mo stretchy="false">|</mo> </mrow> <mo>≤</mo> <msqrt> <mrow> <mn>2</mn> <mi>β</mi> <mo>+</mo> <mn>1</mn> </mrow> </msqrt> <mo stretchy="false">/</mo> <mi>β</mi> </mrow> </math></EquationSource> </InlineEquation>, as in Okazawa and Yokota [<CitationRef CitationID="CR35">35</CitationRef>]. Under the assumption of <i>local</i> Lipschitz continuity of the diffusion term, we show that the family of invariant measures <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(\{\mathcal {S}^{\varepsilon }\}_{\varepsilon \in [0,1]}\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mrow> <mo stretchy="false">{</mo> <msup> <mrow> <mi mathvariant="script">S</mi> </mrow> <mi>ε</mi> </msup> <mo stretchy="false">}</mo> </mrow> <mrow> <mi>ε</mi> <mo>∈</mo> <mo stretchy="false">[</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">]</mo> </mrow> </msub> </math></EquationSource> </InlineEquation> is tight. Moreover, we show that every weak limit point of invariant measures <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(\mu ^{\varepsilon _n} \in \mathcal {S}^{\varepsilon _n}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msup> <mi>μ</mi> <msub> <mi>ε</mi> <mi>n</mi> </msub> </msup> <mo>∈</mo> <msup> <mrow> <mi mathvariant="script">S</mi> </mrow> <msub> <mi>ε</mi> <mi>n</mi> </msub> </msup> </mrow> </math></EquationSource> </InlineEquation> with <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(\varepsilon _n \rightarrow \varepsilon _0 \in [0,1]\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mi>ε</mi> <mi>n</mi> </msub> <mo stretchy="false">→</mo> <msub> <mi>ε</mi> <mn>0</mn> </msub> <mo>∈</mo> <mrow> <mo stretchy="false">[</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">]</mo> </mrow> </mrow> </math></EquationSource> </InlineEquation> must be an invariant measure of the limit equation. When the diffusion term is <i>globally</i> Lipschitz continuous, we discuss a stronger convergence of <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(\mu ^{\varepsilon } \in \mathcal {S}^{\varepsilon }\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msup> <mi>μ</mi> <mi>ε</mi> </msup> <mo>∈</mo> <msup> <mrow> <mi mathvariant="script">S</mi> </mrow> <mi>ε</mi> </msup> </mrow> </math></EquationSource> </InlineEquation> and <InlineEquation ID="IEq9"> <EquationSource Format="TEX">\(\mu ^{\varepsilon _0} \in \mathcal {S}^{\varepsilon _0}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msup> <mi>μ</mi> <msub> <mi>ε</mi> <mn>0</mn> </msub> </msup> <mo>∈</mo> <msup> <mrow> <mi mathvariant="script">S</mi> </mrow> <msub> <mi>ε</mi> <mn>0</mn> </msub> </msup> </mrow> </math></EquationSource> </InlineEquation> as <InlineEquation ID="IEq10"> <EquationSource Format="TEX">\(\varepsilon \rightarrow \varepsilon _0 \in [0,1]\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>ε</mi> <mo stretchy="false">→</mo> <msub> <mi>ε</mi> <mn>0</mn> </msub> <mo>∈</mo> <mrow> <mo stretchy="false">[</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">]</mo> </mrow> </mrow> </math></EquationSource> </InlineEquation> under the Wasserstein metric of probability measures, and prove that the Wasserstein distance between <InlineEquation ID="IEq11"> <EquationSource Format="TEX">\(\mu ^{\varepsilon }\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>μ</mi> <mi>ε</mi> </msup> </math></EquationSource> </InlineEquation> and <InlineEquation ID="IEq12"> <EquationSource Format="TEX">\(\mu ^{\varepsilon _0}\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mi>μ</mi> <msub> <mi>ε</mi> <mn>0</mn> </msub> </msup> </math></EquationSource> </InlineEquation> is bounded by <InlineEquation ID="IEq13"> <EquationSource Format="TEX">\(c |\varepsilon - \varepsilon _0|\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mrow> <mi>c</mi> <mo stretchy="false">|</mo> <mi>ε</mi> <mo>-</mo> </mrow> <msub> <mi>ε</mi> <mn>0</mn> </msub> <mrow> <mo stretchy="false">|</mo> </mrow> </mrow> </math></EquationSource> </InlineEquation> for some constant <InlineEquation ID="IEq14"> <EquationSource Format="TEX">\(c &gt; 0\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>c</mi> <mo>&gt;</mo> <mn>0</mn> </mrow> </math></EquationSource> </InlineEquation> independent of <InlineEquation ID="IEq15"> <EquationSource Format="TEX">\(\varepsilon \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>ε</mi> </math></EquationSource> </InlineEquation>. The main difficulty caused by the lack of compact Sobolev embeddings on unbounded domains, which obstructs the proof of tightness for <InlineEquation ID="IEq16"> <EquationSource Format="TEX">\(\bigcup _{\varepsilon \in [0,1]}\mathcal {S}^\varepsilon \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mo>⋃</mo> <mrow> <mi>ε</mi> <mo>∈</mo> <mo stretchy="false">[</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">]</mo> </mrow> </msub> <msup> <mrow> <mi mathvariant="script">S</mi> </mrow> <mi>ε</mi> </msup> </mrow> </math></EquationSource> </InlineEquation>, is handled by establishing uniform tail estimates for solutions outside sufficiently large balls. The challenge of proving convergence in probability and expectation of the solutions is addressed by carefully analyzing the global monotonicity of the drift term using the condition <InlineEquation ID="IEq17"> <EquationSource Format="TEX">\(|\varpi | \le \sqrt{2\beta + 1}/\beta \)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mrow> <mo stretchy="false">|</mo> <mi>ϖ</mi> <mo stretchy="false">|</mo> </mrow> <mo>≤</mo> <msqrt> <mrow> <mn>2</mn> <mi>β</mi> <mo>+</mo> <mn>1</mn> </mrow> </msqrt> <mo stretchy="false">/</mo> <mi>β</mi> </mrow> </math></EquationSource> </InlineEquation>. Our results also extend to the case where the fractional Laplace operator <InlineEquation ID="IEq18"> <EquationSource Format="TEX">\((-\varDelta )^{\alpha }\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mo stretchy="false">(</mo> <mo>-</mo> <mi>Δ</mi> <mo stretchy="false">)</mo> </mrow> <mi>α</mi> </msup> </math></EquationSource> </InlineEquation> reduces to the classical Laplacian operator.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Limiting behavior of invariant measures for stochastic fractional Ginzburg-Landau equations on unbounded domains

  • Jianren Long,
  • Dexin Li,
  • Sangui Zeng

摘要

We investigate the limiting behavior of invariant measures in terms of weak convergence and the Wasserstein metric of probability measures for a stochastic fractional Ginzburg-Landau equation defined on the entire space \(\mathbb {R}^{n}\) R n , where the growth rate \(2\beta + 1\) 2 β + 1 and the coefficient \(\varpi \) ϖ of the drift term satisfy an admissible condition \(|\varpi | \le \sqrt{2\beta + 1}/\beta \) | ϖ | 2 β + 1 / β , as in Okazawa and Yokota [35]. Under the assumption of local Lipschitz continuity of the diffusion term, we show that the family of invariant measures \(\{\mathcal {S}^{\varepsilon }\}_{\varepsilon \in [0,1]}\) { S ε } ε [ 0 , 1 ] is tight. Moreover, we show that every weak limit point of invariant measures \(\mu ^{\varepsilon _n} \in \mathcal {S}^{\varepsilon _n}\) μ ε n S ε n with \(\varepsilon _n \rightarrow \varepsilon _0 \in [0,1]\) ε n ε 0 [ 0 , 1 ] must be an invariant measure of the limit equation. When the diffusion term is globally Lipschitz continuous, we discuss a stronger convergence of \(\mu ^{\varepsilon } \in \mathcal {S}^{\varepsilon }\) μ ε S ε and \(\mu ^{\varepsilon _0} \in \mathcal {S}^{\varepsilon _0}\) μ ε 0 S ε 0 as \(\varepsilon \rightarrow \varepsilon _0 \in [0,1]\) ε ε 0 [ 0 , 1 ] under the Wasserstein metric of probability measures, and prove that the Wasserstein distance between \(\mu ^{\varepsilon }\) μ ε and \(\mu ^{\varepsilon _0}\) μ ε 0 is bounded by \(c |\varepsilon - \varepsilon _0|\) c | ε - ε 0 | for some constant \(c > 0\) c > 0 independent of \(\varepsilon \) ε . The main difficulty caused by the lack of compact Sobolev embeddings on unbounded domains, which obstructs the proof of tightness for \(\bigcup _{\varepsilon \in [0,1]}\mathcal {S}^\varepsilon \) ε [ 0 , 1 ] S ε , is handled by establishing uniform tail estimates for solutions outside sufficiently large balls. The challenge of proving convergence in probability and expectation of the solutions is addressed by carefully analyzing the global monotonicity of the drift term using the condition \(|\varpi | \le \sqrt{2\beta + 1}/\beta \) | ϖ | 2 β + 1 / β . Our results also extend to the case where the fractional Laplace operator \((-\varDelta )^{\alpha }\) ( - Δ ) α reduces to the classical Laplacian operator.