scientific article; zbMATH DE number 3996275

From MaRDI portal
Publication:4724203





This paper is reviewed jointly with the two preceding ones (Zbl 0615.35081; Zbl 0615.35082). The papers deal with the interpretation of nature, i.e., measured values, and difficulties arising in interpreting nature. A special problem in applications is frequently described by means of very complex systems of partial differential equations or other equations. The coefficients and the source terms are internal parameters f inaccessible to measurement. They are to be determined by measured values g. Between f and g physical laws hold. In a formal way one can describe such inverse problems by equations of the first kind \(Af=g.\) In the case considered the operator A has a very complicated structure. In order to be able to make concrete statements about a system of partial differential equations one must often vastly simplify the system. This leads to a situation similar to that which arises when the system studied is modelled too coarsely, or the equations are discretized. Further, measurements are possible only outside an open set like in geophysics and medicine. This leads to intrinsic ambiguities, i.e., different values of the internal physical parameters f can give the same measured values g. The set of all f satisfying \(Af=g\), where g is given, can have the cardinal number of the continuum. From the mathematical point of view there does not exist an algorithm in order to uniquely determine a solution f. These facts are known if A is a matrix and det A\(=0\). Each additional mathematical piece of information must be clarified by the experiment. This situation is similar to that of Cantor's continuum hypothesis: neither provable, nor refutable. It is well known that K. Gödel (1931) and P. Cohen (1963) proved the inadequacy of Zermelo-Fraenkel's axioms to decide the continuum hypothesis. It seems that the inadequacy of measured values (lack of information) is a general principle in interpretating of nature. Therefore laboratories are necessary to get additional information by experiments. In geophysics and medicine however often one cannot get such additional information. Therefore many papers in these areas are not of interest for applications. As a first step the set of all solutions of an inverse problem can be introduced and studied. The elements of such a set satisfy an equivalence relation. In the case of the inverse source problem of the Laplace equation this relation was introduced by P. Pizetti (1909-1913), G. Lauricella (1907-1911) and studied from a systematic point of view by G. Anger (1976) using methods of functional analysis. Similar investigations are possible relative to the heat equation. In order to get more information mathematical modelling becomes more and more important. The second paper contains some history of inverse problems, the third paper was written for scientists which are not mathematicians.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4724203)