<p>We consider variational inequalities with invertible operators <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\({{\mathcal {A}}}_s:W^{1,p}_0(\varOmega )\rightarrow W^{-1,p^\prime }(\varOmega )\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(s\in {\mathbb {N}}\)</EquationSource> </InlineEquation>, in divergence form and constraint set <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(V\subset W^{1,p}_0(\varOmega )\)</EquationSource> </InlineEquation> defined by a measurable lower constraint <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\varphi :\varOmega \rightarrow \overline{{\mathbb {R}}}\)</EquationSource> </InlineEquation> and a measurable upper constraint <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(\psi :\varOmega \rightarrow \overline{\mathbb R}\)</EquationSource> </InlineEquation>, where <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(\varOmega \)</EquationSource> </InlineEquation> is a nonempty bounded open set in&#xa0;<InlineEquation ID="IEq7"> <EquationSource Format="TEX">\({\mathbb {R}}^n\)</EquationSource> </InlineEquation> (<InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(n\geqslant 2\)</EquationSource> </InlineEquation>) and <InlineEquation ID="IEq9"> <EquationSource Format="TEX">\(p&gt;1\)</EquationSource> </InlineEquation>. We assume that the sequence <InlineEquation ID="IEq10"> <EquationSource Format="TEX">\(\{{\mathcal A}_s\}\)</EquationSource> </InlineEquation> &#xa0;<i>G</i>-converges to an invertible operator <InlineEquation ID="IEq11"> <EquationSource Format="TEX">\({\mathcal {A}}:W^{1,p}_0(\varOmega )\rightarrow W^{-1,p^\prime }(\varOmega )\)</EquationSource> </InlineEquation>. In addition, we assume that the set <InlineEquation ID="IEq12"> <EquationSource Format="TEX">\(\{\varphi =\psi \}\)</EquationSource> </InlineEquation> has nonempty interior, the measure of the intersection of the boundary of the set <InlineEquation ID="IEq13"> <EquationSource Format="TEX">\(\{\varphi =\psi \}\)</EquationSource> </InlineEquation> with <InlineEquation ID="IEq14"> <EquationSource Format="TEX">\(\varOmega \)</EquationSource> </InlineEquation> is zero, and there exist functions <InlineEquation ID="IEq15"> <EquationSource Format="TEX">\(\bar{\varphi },\bar{\psi }\in W^{1,p}_0(\varOmega )\)</EquationSource> </InlineEquation> such that <InlineEquation ID="IEq16"> <EquationSource Format="TEX">\(\varphi \leqslant {\bar{\varphi }}\leqslant {\bar{\psi }}\leqslant \psi \)</EquationSource> </InlineEquation> a.e. in&#xa0;<InlineEquation ID="IEq17"> <EquationSource Format="TEX">\(\varOmega \)</EquationSource> </InlineEquation> and the measure of the set <InlineEquation ID="IEq18"> <EquationSource Format="TEX">\(\{\varphi \ne \psi \}\setminus \{{\bar{\varphi }}\ne {\bar{\psi }}\}\)</EquationSource> </InlineEquation> is zero. Under these assumptions, we prove that the solutions of the considered variational inequalities converge weakly in <InlineEquation ID="IEq19"> <EquationSource Format="TEX">\(W^{1,p}_0(\varOmega )\)</EquationSource> </InlineEquation> to the solution of a similar variational inequality with the operator <InlineEquation ID="IEq20"> <EquationSource Format="TEX">\({\mathcal {A}}\)</EquationSource> </InlineEquation> and the constraint set&#xa0;<i>V</i>. We show that there is a fundamental difference between the considered case and the previously studied case where the measure of the set <InlineEquation ID="IEq21"> <EquationSource Format="TEX">\(\{\varphi =\psi \}\)</EquationSource> </InlineEquation> is zero.</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

Convergence of solutions of variational inequalities with measurable bilateral constraints coinciding on a set of positive measure

  • Alexander A. Kovalevsky

摘要

We consider variational inequalities with invertible operators \({{\mathcal {A}}}_s:W^{1,p}_0(\varOmega )\rightarrow W^{-1,p^\prime }(\varOmega )\) , \(s\in {\mathbb {N}}\) , in divergence form and constraint set \(V\subset W^{1,p}_0(\varOmega )\) defined by a measurable lower constraint \(\varphi :\varOmega \rightarrow \overline{{\mathbb {R}}}\) and a measurable upper constraint \(\psi :\varOmega \rightarrow \overline{\mathbb R}\) , where \(\varOmega \) is a nonempty bounded open set in  \({\mathbb {R}}^n\) ( \(n\geqslant 2\) ) and \(p>1\) . We assume that the sequence \(\{{\mathcal A}_s\}\)  G-converges to an invertible operator \({\mathcal {A}}:W^{1,p}_0(\varOmega )\rightarrow W^{-1,p^\prime }(\varOmega )\) . In addition, we assume that the set \(\{\varphi =\psi \}\) has nonempty interior, the measure of the intersection of the boundary of the set \(\{\varphi =\psi \}\) with \(\varOmega \) is zero, and there exist functions \(\bar{\varphi },\bar{\psi }\in W^{1,p}_0(\varOmega )\) such that \(\varphi \leqslant {\bar{\varphi }}\leqslant {\bar{\psi }}\leqslant \psi \) a.e. in  \(\varOmega \) and the measure of the set \(\{\varphi \ne \psi \}\setminus \{{\bar{\varphi }}\ne {\bar{\psi }}\}\) is zero. Under these assumptions, we prove that the solutions of the considered variational inequalities converge weakly in \(W^{1,p}_0(\varOmega )\) to the solution of a similar variational inequality with the operator \({\mathcal {A}}\) and the constraint set V. We show that there is a fundamental difference between the considered case and the previously studied case where the measure of the set \(\{\varphi =\psi \}\) is zero.