Degenerate integrodifferential equations of parabolic type with Robin boundary conditions: L^2-theory

From MaRDI portal
(Redirected from Publication:1006183)
Degenerate integrodifferential equations of parabolic type with Robin boundary conditions: \(L^2\)-theory





The authors study the initial value problem for the following degenerate integrodifferential equation of parabolic type \[ {d\over dt}(M(t)u(t))+ L(t) u(t)+\int^t_0 B(t,s)u(s)\,ds= f(t),\quad 0< t\leq T,\tag{1} \] \[ M(0)u(0)= M(0) u_0,\tag{2} \] where \(L(t)\) is a realization of a second-order elliptic differential operator in \(L^2(\Omega)\) with the Robin boundary condition, \(M(t)\) is the multiplication operator by some nonnegative function and \(B(t,s)\) is a linear second-order partial differential operator for \(0\leq s\leq t\leq T\). At first the authors study the problem \[ {d\over dt}(M(t)u(t))+ L(t) u(t)= f(t),\quad 0< t\leq T,\tag{3} \] \[ M(0)u(0)= M(0)u_0.\tag{4} \] With \(A(t)= L(t)M(t)^{-1}\), \(0\leq t\leq T\), the problem \((3)\wedge (4)\) is transformed into the problem \[ {d\over dt}v(t)+ A(t)v(t)\ni f(t),\quad 0< t\leq T,\tag{5} \] \[ v(0)= v_0= M(0)u_0.\tag{6} \] Let \(\Omega\subset\mathbb{R}^n\) be a bounded open set with a \(C^2\)-boundary \(\partial\Omega\). If \[ \Phi_1(t,s)=-\Biggl({\partial\over\partial t}+{\partial\over\partial s}\Biggr) e^{-(t-s)A(t)}=- {1\over 2\pi i} \int_\Gamma e^{\lambda(t-s)}{\partial\over\partial t}(\lambda+ A(t))^{-1} d\lambda \] and \[ \Phi(t,s)= \Phi_1(t,s)+ \int^t_s \Phi_1(t,\tau)\Phi(\tau, s)\,d\tau \] for \(0\leq s< t\leq T\), then \[ U(t,s)= e^{-(t-s)A(t)}+ \int^t_s e^{-(t-s)A(t)}\Phi(\tau,s)\,d\tau,\;0\leq s< t\leq T, \] is the fundamental solution of \((5)\wedge (6)\) in the sense that \(U\) is differentiable with respect to \(t\) in \((s,T)\), \(R(U(t,s))\subset D(A(t))\) for \(0\leq s< t\leq T\) and \[ {\partial U\over\partial t}(t,s)+ A(t) U(t,s)\ni 0, \] \[ \lim_{\substack{ t\to s\\ t> s}}\| U(t,s)v_0- v_0\|_{L^2(\Omega)}= 0,\quad v_0\in D(A(s)),\;0\leq s< T. \] The authors suppose \[ f\in C^\omega([0, T], L^2(\Omega)),\quad (1-\rho)(2-\rho)^{-1}< \omega\leq 1 \] then \[ v(t)= U(t,0) v_0+ \int^t_0 U(t,s) f(s)\,ds,\quad v_0\in D(A(0)), \] is a solution of \((5)\wedge(6)\) with \[ v\in C([0, T], L^2(\Omega)\cap C^1((0, T], L^2(\Omega))), \] \[ v(t)\in D(A(t)),\quad 0< t\leq T, \] and there exists a solution \(u\) of \((3)\wedge (4)\) with \(u_0\in D(L(0))\) and \[ u\in C((0, T], L^2(\Omega)), \] \[ Mu\in C([0, T], L^2(\Omega))\cap C^1((0,T], L^2(\Omega)), \] \[ u(t)\in D(L(t)),\quad 0< t\leq T, \] \[ L(.) u(.)\in C((0, T], L^2(\Omega)). \] The authors prove also the uniqueness of solutions \(v\) and \(u\) of \((5)\wedge (6)\), respectively \((3)\wedge (4)\). The results for \((1)\wedge (2)\) are similar with the results of \((3)\wedge (4)\).











This page was built for publication: Degenerate integrodifferential equations of parabolic type with Robin boundary conditions: \(L^2\)-theory

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1006183)