Semi-bounded differential operators, contractive semigroups and beyond (Q2438523)

Necessary and sufficient conditions for an operator \(A\) to be the infinitesimal generator of a strongly continuous semigroup \(S(t)\) satisfying \(\|S(t)\| \leq Ce^{\omega t}\) \((t \geq 0)\) were given by \textit{E. Hille} and \textit{R. S. Phillips} [Functional analysis and semi-groups. 3rd printing of rev. ed. of 1957. Providence: AMS (1974; Zbl 0392.46001)]: the resolvent \(R(\lambda; A) = (\lambda I - A)^{-1}\) must exist for \(\lambda > \omega\) and the infinite set of inequalities \(\|R(\lambda; A)^n\| \leq C/(\lambda - \omega)^n\) \((\lambda > \omega, \;n = 1, 2,\dots)\) must be verified. Application of the ``raw'' Hille-Phillips theorem to particular operators (say, differential) is difficult except in the case \(C = 1\), where the first inequality implies all the others. Moreover, it was proved by \textit{G. Lumer} and \textit{R. S. Phillips} [Pac. J. Math. 11, 679--698 (1961; Zbl 0101.09503)] that, assuming that \(\omega = 0\) (as we may via a translation of \(A)\), this first inequality is equivalent to \textit{dissipativity} of the operator \(A\) (which means \(\text{Re}\langle x^*, Ax \rangle \leq 0\) for \(x \in D(A)\), \(x^* \in E^*\) in the duality set of \(x)\) plus other easily verifiable conditions on \(A\). This result is doubly useful since dissipativity has interesting physical interpretations, especially in the Hilbert space case where it reduces to \(\text{Re}(u, Au) \leq 0\). The semigroups generated by dissipative operators are \textit{contractive,} that is, they satisfy \(\|S(t)\| \leq 1\) \((t \geq 0)\). Dissipative operators are a particular case of \textit{semi-bounded} operators characterized by the inequality \(\text{Re}\langle x^*, Ax \rangle \leq c\|u\|^2\). After going over the necessary functional analysis in Chapter 1, the authors devote the rest of the book to the application of the theory to certain classes of differential operators in \(L^p\) spaces. In Chapter 2, the operators are \[ Au = \text{div} ({\mathcal A} \nabla u) + {\mathbf b}\nabla u + \text{div}({\mathbf c}u) + au, \] where \(\mathcal A\) is a matrix whose entries are complex valued measures. There are results on the dissipativity in \(L^p\) of the sesquilinear form associated with this operator and on dissipativity of \(A\) itself; there are also theorems on more general operators and on weaker notions of dissipativity. In Chapter 3, there are results on the elasticity system \[ Eu = \Delta u + (1 - 2\nu)^{-1} \, \nabla \, \text{div}\, u, \] both in the planar and spatial case. Chapter 4 extends the setup of Chapter 2 to systems of partial differential operators. Chapter 5 deals with the \textit{angle of dissipativity} of a dissipative operator \(A\), which is the set of all \(\phi\) such that \(e^{i\phi}A\) is dissipative. The results are applied to \[ A = \nabla^t ({\mathcal A}(x) \nabla), \] where \({\mathcal A}\) is a matrix with complex locally integrable entries, and generalized to systems. Chapter 6 deals with higher order differential operators, and shows by means of counterexamples that these operators cannot be \(L^p\)-dissipative except for \(p = 2\). There are positive results, however, when the operator is given in the cone of nonnegative functions. Chapter 7 is on \(L^2\)-weighted positivity of operators \(L\), which in the scalar case means \[ \text{Re} \int_\Omega \langle Lu, u \rangle \Psi dx \geq 0. \] There are results for operators, systems and particular equations such as the 3-dimensional Lamé system, the polyharmonic operator and real positive powers of the Laplacian. The book is logically ordered and clearly written; it would be even more readable with a table of symbols. It will be of interest to specialists and graduate students working on partial differential operators in function spaces. It presupposes only a modest acquaintance with functional analysis and partial differential equations.