Maximal Counts in the Stopped Occupancy Problem
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908421428609024 |
|---|---|
| author | Gnedin, Alexander Janson, Svante Malinovsky, Yaakov |
| author_facet | Gnedin, Alexander Janson, Svante Malinovsky, Yaakov |
| contents | We revisit a version of the classic occupancy scheme, where balls are thrown until almost all boxes receive a given number of balls. Special cases are widely known as coupon-collectors and dixie cup problems. We show that as the number of boxes tends to infinity, the distribution of the maximal occupancy count does not converge, but can be approximated by a convolution of two Gumbel distributions, with the approximating distribution having oscillations close to periodic on a logarithmic scale. We pursue two approaches: one relies on lattice point processes obtained by poissonisation of the number of balls and boxes, and the other employs interpolation of the multiset of occupancy counts to a point process on reals. This way we gain considerable insight in known asymptotics obtained previously by mostly analytic tools. Further results concern the moments of maximal occupancy counts and ties for the maximum. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_20411 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Maximal Counts in the Stopped Occupancy Problem Gnedin, Alexander Janson, Svante Malinovsky, Yaakov Probability Discrete Mathematics Statistics Theory 60G70, 60G55, 62G32 We revisit a version of the classic occupancy scheme, where balls are thrown until almost all boxes receive a given number of balls. Special cases are widely known as coupon-collectors and dixie cup problems. We show that as the number of boxes tends to infinity, the distribution of the maximal occupancy count does not converge, but can be approximated by a convolution of two Gumbel distributions, with the approximating distribution having oscillations close to periodic on a logarithmic scale. We pursue two approaches: one relies on lattice point processes obtained by poissonisation of the number of balls and boxes, and the other employs interpolation of the multiset of occupancy counts to a point process on reals. This way we gain considerable insight in known asymptotics obtained previously by mostly analytic tools. Further results concern the moments of maximal occupancy counts and ties for the maximum. |
| title | Maximal Counts in the Stopped Occupancy Problem |
| topic | Probability Discrete Mathematics Statistics Theory 60G70, 60G55, 62G32 |
| url | https://arxiv.org/abs/2506.20411 |