Abstract
We study the asymptotic behaviour of the relative entropy (to stationarity) for a commonly used model for riffle shuffling a deck of n cards m times. Our results establish and were motivated by a prediction in a recent numerical study of Trefethen and Trefethen. Loosely speaking, the relative entropy decays approximately linearly (in m) for in <log(2)n, and approximately exponentially for m > log(2) n. The deck becomes random in this information-theoretic sense after m = 3/2 log(2) n shuffles.
| Translated title of the contribution | Information loss in riffle shuffling |
|---|---|
| Original language | English |
| Pages (from-to) | 79 - 95 |
| Number of pages | 17 |
| Journal | Combinatorics, Probability and Computing |
| Volume | 11 (1) |
| DOIs | |
| Publication status | Published - Jan 2002 |
Bibliographical note
Publisher: Cambridge University PressFingerprint
Dive into the research topics of 'Information loss in riffle shuffling'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver