Degenerate integrodifferential equations of parabolic type with Robin boundary conditions: \(L^2\)-theory (Q1006183): Difference between revisions

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 \((3)\wedge (4)\).