Exact and truncated difference schemes for boundary value ODEs (Q2430557): Difference between revisions

The authors present a first unified theory of finite difference methods for the solution of linear and nonlinear boundary value problems (BVPs) of ordinary differential equations (ODEs). New numerical algorithms for BVPs are developed and evaluated which have the same efficiency and accuracy as the well-known and contemporary methods for the solution of initial value problems (IVPs). The present book is based on the theory of difference schemes established by the famous Russian mathematician Aleksandr Andreevich Samarskii. Some of the authors were Samarskii's students. The main focus of this text is the theory of numerical methods for BVPs that are based on exact difference schemes (EDS) as well as on well studied solvers for IVPs. It is shown that the combination of the EDS with the modern IVP-solvers results in numerical truncated difference schemes (TDS)-algorithms, which are highly efficient. The theory of the EDS and TDS permits the construction of a posteriori error estimators, which in turn are the basis for adaptive algorithms. The book has six chapters. The first chapter is designed to acquaint the reader with some types of difference schemes that occur in the literature. Chapter 2 deals with BVPs for systems of first-order nonlinear ODEs on a finite interval. The existence of two-point EDS on an arbitrary non-equidistant grid is shown and its structure is described. Two a posteriori error estimators are proposed along with appropriate grid generators. Some numerical examples confirming the theoretical results are given. Chapter 3 is devoted to BVPs for second-order nonlinear ODEs with a monotone operator. The appropriate EDS and TDS are constructed, theoretically analyzed and practically tested. Chapter 4 generalizes the results of Chapter 3 to BVPs for systems of second order nonlinear ODEs with monotone operators. The developed algorithms have been tested on a variety of numerical examples and the corresponding results indicate a good conformity with the theoretical predictions. Chapter 5 gives some results on EDS and TDS for BVPs on the half-axis. These techniques are a real alternative to those methods which are based on transformation of the original problems into a singular BVPs on a finite interval, and the subsequent solution of this singular problem by various special approaches. Chapter 6 proposes 32 exercises with the corresponding solutions. The book is addressed at graduate students of mathematics and physics, as well as to working scientists and engineers as a self-study tool and reference. Researchers working with BVBs will find appropriate and effective numerical algorithms for their needs. The book can also be used as a textbook for a one- or two- semester course on numerical methods for ODEs.