The analysis of many number theoretic algorithms turns on the
role played by integers which have only small prime factors; such integers are
known as "smooth numbers". To be able to determine which algorithm is faster
than which, it has turned out to be important to have accurate estimates for the
number of smooth numbers in various sequences. In this chapter, we will first
survey the important estimates for application to computational number theory
questions, results as well as conjectures, before moving on to give sketches of
the proofs of many of the most important results. After this, we will describe
applications of smooth numbers to various problems in different areas of number
theory. More complete surveys, with many more references, though with a
different focus, were given by Norton in 1971, and by Hildebrand and Tenenbaum
in 1993.
Article