Polarization and moment tensors. With applications to inverse problems and effective medium theory (Q884839)

The present book deals with one of the most interesting topics in the theory of inverse problems: the detection of (small) inclusions in a (complex) medium and the localization of their positions inside it, when a set of non-invasive pieces of information is available. This means that measurements inside the body are to be avoided and that only boundary measurements are allowed. It is obvious to everyone the importance of such studies: the applications spread from biology and medicine (e.g. research and localizations of tumors), to the determinations of cracks in materials of interest in technology. From a mathematical point of view it is elliptic PDE's that play a basic role in this theory, since the analysis of the previous questions is naturally concerned with stationary problems. More exactly, transmission problems are the starting and the central point of the investigations. To give an idea of a typical problem, assume that we are given a linear elliptic equation of the form \[ \nabla \cdot \sigma \nabla u = 0,\quad \text{ in\;}\Omega, \eqno(1) \] related to a bounded domain \(\Omega\subset {\mathbb R}^d\), \(d\geq 2\), where the conductivity \(\sigma\) admits the representation \(\sigma = 1 + (k - 1)\chi(D)\). Here \(\chi\) is the characteristic function of a Lipschitz domain \(D\) compactly embedded into the body \(\Omega\) with a constant electrical conductivity \(k\neq 1\), \(1\) being the constant electrical conductivity of \(\Omega\setminus D\). Equation (1) is endowed with a boundary condition on \(\partial \Omega\), e.g. Neumann's one, and function \(u\) -- the meaning of which is an electrical potential -- and its normal derivative \(D_\nu u\) are assumed to satisfy continuity conditions when crossing \(\partial D\). Finally, the inverse problem consists just in recovering the inclusion \(D\). As an additional information it is supplied a measurement of \(u\) on the (external) boundary \(\partial \Omega\). Of course, this identification problem can be generalized to the case of several compactly embedded inclusions as well as to (constant) anisotropic media and even to composite media, where also homogeneization theory intervenes. Similar problems can be set in the case of elastic bodies. Then equation (1) has to be replaced with \[ \sum_{i,k,l=1}^d\,D_{x_j}(C_{i,j,k,l}D_{x_j}u_k) = 0, \quad \text{ in\;}\Omega,\;i=1,\ldots,d, \eqno(2) \] where the elasticity tensor \(C=(C_{i,j,k,l})\) is given by \[ C_{i,j,k,l}=[\lambda\chi(\Omega\setminus D)+ \widetilde\lambda\chi(D)]\delta_{i,j}\delta_{k,l} + [\mu\chi(\Omega\setminus D)+ \widetilde\mu\chi(D)](\delta_{i,k}\delta_{j,l}+\delta_{i,l}\delta_{j,k}). \] The thread unifying such seemingly different problems is the use of generalized polarization and elastic moment tensors. The authors' main approach is based on layer potential techniques. In case of equation (1) electrical impedance imaging uses measurements of boundary voltage potentials and associated boundary currents to infer information about the internal conductivity profile of an object. As is well-known the class of problems to be investigated is severely ill-posed and highly nonlinear. This is the main obstacle to finding non-iterative reconstruction algorithms with limited data. If, however, structural information about the conductivity profile -- in case of equation (1) -- are available, then it is possible to determine specific features about the conductivity distribution with a satisfactory resolution. A fundamental tool consists in finding out suitable asymptotic expansions related to small volume inclusions by making use of generalized polarization tensors (GPT's) associated with a bounded Lipschitz domain and isotropic or anisotropic conductivities. This procedure allows to accurately and efficiently reconstruct the location and geometric features of the inclusions in a stable way. Moreover, stable and accurate algorithms can be constructed for the numerical computations of the steady-state voltage in the presence of small conductivity inclusions. The last part of the book is devoted to isotropic linear elasticity inverse problems. There some fundamental properties of elastic moment tensors (EMT's) are introduced and analyzed In particular, the authors estimates the eigenvalues of EMT's and compute EMT's associated with ellipses, elliptic holes, and hard inclusions of elliptic shape. Finally, they apply such results to the detection of small inclusions and effective properties of dilute elastic composites. It is worth remembering that any chapter of the volume contains an interesting section devoted to ``Further results and open problems''. To conclude, in the reviewer's opinion the volume is extremely well-written and organized, as well as (reasonably) self-contained, so that any reader with a good knowledge in elliptic PDE's and interest in inverse problems will enjoy it -- as the reviewer did --.