In 1990, breakthrough work of Fristedt introduced a probabilistic framework for deriving

limiting distributions for statistics for integer partitions, now known as a Boltzmann model

or Fristedt’s conditioning device. Subsequent authors have continued to use his ideas to find

limiting statistics for partitions, and analogues of his ideas have also been used, for example,

to generate fast sampling algorithms for other combinatorial structures. For partitions, precisely

because the generating function is an infinite product, one gains independence of the relevant

random variables under the Boltzmann model, making it easier to calculate distributions.

In this talk, I will discuss these ideas and the nontrivial task of extending them to unimodal

sequences of integers, which lack a product generating function.

This is joint work with Kathrin Bringmann.

Zoom Meeting ID: 919 6497 4060

For password see the email or contact Pieter Moree (moree$@$mpim-bonn$.$mpg$.$de)

© MPI f. Mathematik, Bonn | Impressum & Datenschutz |