Book Description

In the last decade high order methods for scientific computing have been attracting increasing interest. This trend has been generated by the need for a higher accuracy in the numerical simulation of more and more complex scientific and technological problems; it is backed up by sound mathematical research, and propelled by the availability of faster supercomputers. Spectral methods have now become the methods preferred in the prediction of many highly structured phenomena. The h-p version of the finite element method has proven extremely effective in handling singularities in structural mechanics. Finite differences have been demonstrated capable of blending flexibility and accuracy in applications to non-smooth problems. Although these and other high order methods originated from different, sometimes even opposite philosophies, they exhibit common features, and share a large part of the methodologies for their mathematical investigation and their algorithmic implementation. The technical content of the 14 invited and 30 general papers presented in this volume reflect the high standard of current research being achieved in this field.