Smooth numbers are integers whose prime factors are smaller than a threshold y. In the 80s they became important outside of pure math, as Pomerance’s quadratic sieve for factoring integers relied on their distribution. The density of smooth numbers up to x can be approximated, in some range, using a peculiar function ρ called Dickman’s function, defined via a delay-differential equation. All of the above is also true for smooth polynomials over finite fields. We’ll survey these topics and discuss recent results concerning the range of validity of the approximation of the density of smooth numbers by ρ, whose proofs rely on relating the counting function of smooth numbers to the Riemann zeta function and the counting function of primes. In particular, we uncover phase transitions in the behavior of the density at the points y=(log x)2 (as conjectured by Hildebrand) and y=(log x)3/2, when previously only a transition at y=log x was known and understood. These transitions also occur in the polynomial setting. We’ll also show that a standard conjecture on the error in the Prime Number Theorem implies ρ is always a lower bound for the density, addressing a conjecture of Pomerance.
This video is part of the Number Theory Web Seminar series.
