Well-posed, ill-posed, and intermediate problems with applications (Q2368518)

This book is intended as a tutorial for students and post-graduate students, as a guide for lecturers, and as a monograph for scientific researchers in pure and applied mathematics and dealing with algebraic, differential, integral, and operator equations. More specifically, the book treats one of the key problems in applied mathematics, i.e., how to investigate stability in solving equations allowing inaccuracies in the set of initial data, parameters and coefficients of a mathematical model and also with allowance for miscalculations, including roundoff errors. The authors consider a third class of problems which are intermediate between well- and ill-posed problems. These are problems that change their property of being well- or ill-posed under equivalent transformations of the governing equations, and also problems that display the property of being either well- or ill-posed depending on the type of the functional space used. From the standpoint of applications, problems that change their property of being well- or ill-posed under equivalent transformations should be paid special attention since their solution by classical methods may give rise to crude errors in calculations, resulting in accidents and crashes, and even in catastrophes of ships, control systems, etc. The book consists of two complementary parts. In the first part, general properties of all the three classes of mathematical, physical and engineering problems are considered, together with the approaches used to solve them. The first part has three chapters. The first chapter is devoted to examples of simplest ill-posed problems. The second chapter introduces the concept of problems intermediate between well- and ill-posed problems. In the third chapter the change of sensitivity to measurement errors under integral transformation used in modeling of ships and marine control systems is considered. In the second part, several stable methods for solving inverse ill-posed problems are described, illustrated with numerical examples. The second part has two chapters. The forth chapter presents regular methods for solving ill-posed problems: classical methods for solving Fredholm integral equations of the first kind, Gauss least-squares method, Moore-Penrose inverse-matrix method, Tikhonov regularization method, compactness method. The fifth chapter cites numerical results in reconstructions of blurred and defocused images, and presents numerical results for problems of X-ray tomography and of magnetic-field synthesis in an NMR tomograph.