In this book, a new approach to approximation procedures is developed. This new approach is characterized by the common feature that the procedures are accurate without being convergent as the mesh size tends to zero. This lack of convergence is compensated for by the flexibility in the choice of approximating functions, the simplicity of multi-dimensional generalizations, and the possibility of obtaining explicit formulas for the values of various integral and pseudodifferential operators applied to approximating functions. The developed techniques allow the authors to design new classes of high-order quadrature formulas for integral and pseudodifferential operators, to introduce the concept of approximate wavelets, and to develop new efficient numerical and semi-numerical methods for solving boundary value problems of mathematical physics. The book is intended for researchers interested in approximation theory and numerical methods for partial differential and integral equations. Readership Graduate students and research mathematicians interested in approximation theory and numerical methods. Reviews "Altogether, this is an interesting book, most useful for the approximation theorist as well as for the practitioner who appreciates that approximate approximations are a useful and practicable alternative to the classical ideas of approximations with small stepsizes and ultimately convergence theorems." *-- Mathematical Reviews* |