On the automatic numerical evaluation of definite integrals

Abstract
A critical examination is made of adaptive subdivision as a means of reliably and efficiently performing the automatic evaluation of definite integrals. A model is set up which embodies the basic features of adaptive schemes. Circumstances are discussed under which adaptive schemes may inspire false confidence in the results produced. The efficiency of the method is seriously impaired by any attempts to overcome this difficulty. The conclusions have been illustrated by appropriate numerical examples.

This publication has 0 references indexed in Scilit: