Parabolic equations with dynamical boundary conditions and source terms on interfaces (Q466943): Difference between revisions

In this well-written article the authors study linear parabolic initial-boundary value problems of the form \[ \begin{aligned}\varepsilon \partial_t u - \nabla \cdot \mu \nabla u &= f_{\Omega}\qquad \mathrm{in } J \times (\Omega \setminus \Sigma),\\ u &= 0 \qquad \text{on } J \times (\partial \Omega \setminus \Gamma),\\ \varepsilon \partial_t u + \nu \cdot \mu \nabla u + bu &= f_{\Gamma}\qquad \text{on } J \times \Gamma,\\ \varepsilon \partial_t u + [\nu_{\Sigma} \cdot \mu \nabla u] &= f_{\Sigma} \qquad \text{on } J \times \Sigma,\\ u(0) &= u_0 \qquad \text{in } \Omega \cup \Gamma.\end{aligned} \] Here, \(J = (0,T)\) is a bounded time interval, \(\Omega \subset \mathbb{R}^d\) is a bounded domain, \(\Gamma \subseteq \partial \Omega\) is a part of the boundary with outer normal \(\nu\) and \(\Sigma \subset \Omega\) is a sufficiently regular hypersurface with normal field \(\nu_{\Sigma}\). Furthermore, for the coefficients one assumes that \(\mu\) is bounded, measurable and elliptic, \(\varepsilon\) is positive, bounded, measurable and elliptic away from zero, and that \(b\) lies in an \(L^p\)-space. The problem is treated with generalized form methods, as introduced in [\textit{W. Arendt} and the first author, J. Oper. Theory 67, No. 1, 32--72 (2012; Zbl 1243.47009)], which yield that a proper realization of the differential operator involved in the problem generates a holomorphic semigroup on the whole reflexive \(L^p\)-scale. For this, the case \(\varepsilon \equiv 1\) and \(b \equiv 0\) is studied first and the general case is then deduced by perturbation methods. As a consequence of these results, one obtains maximal regularity of the problem by a result of \textit{D. Lamberton} [J. Funct. Anal. 72, 252--262 (1987; Zbl 0621.47039)]. In the final section, maximal regularity is then used to study quasilinear variants of the equation.