par Louchard, Guy 
Référence Random structures & algorithms, 5, 5, page (665-702)
Publication Publié, 1994
Référence Random structures & algorithms, 5, 5, page (665-702)
Publication Publié, 1994
Article révisé par les pairs
| Résumé : | This article considers a classical binary tree implementation of a set of keys: the trie. The trie size properties in a static environment are well known: The size is asymptotically Gaussian when the number of keys is large. In this article we analyze the trie in a dynamic environment, where the trie is allowed to grow and shrink in a probabilistic way. It appears that the trie size can be described by a stochastic process which is asymptotically Gaussian non‐Markovian. This also allows the complete asymptotic analysis of the trie size maximum and the trie size integrated cost. © 1994 John Wiley & Sons, Inc. Copyright © 1994 Wiley Periodicals, Inc., A Wiley Company |
