In this paper a general class of tree algorithms is analyzed. It is shown that, by using an appropriate probabilistic representation of the quantities of interest, the asymptotic behavior of these algorithms can be obtained quite easily without resorting to the usual complex analysis techniques. This approach gives a unified probabilistic treatment of these questions. It simplifies and extends some of the results known in this domain.
Publié le : 2005-11-14
Classification:
Splitting algorithms,
divide and conquer algorithms,
unusual laws of large numbers,
asymptotic oscillating behavior,
data structures,
tries,
renewal theorem,
68W40,
60K20,
90B15
@article{1133965768,
author = {Mohamed, Han\`ene and Robert, Philippe},
title = {A probabilistic analysis of some tree algorithms},
journal = {Ann. Appl. Probab.},
volume = {15},
number = {1A},
year = {2005},
pages = { 2445-2471},
language = {en},
url = {http://dml.mathdoc.fr/item/1133965768}
}
Mohamed, Hanène; Robert, Philippe. A probabilistic analysis of some tree algorithms. Ann. Appl. Probab., Tome 15 (2005) no. 1A, pp. 2445-2471. http://gdmltest.u-ga.fr/item/1133965768/