Abstract
dc:descriptionThis paper comprises an attempt of investigation in the frame of asymptotic relations for sums of random variables whose distributions belong in the wider class of heavy tail distributions.More specifically, as a first result, we find the lower bound of asymptotic of large deviations for random sums of non negative random variables with long tails distributions.Moreover, for the non negative, independent and identical distributed random variables, we study the centered asymptotic relation of large deviations for non random sums as well as random sums in the frame of dominately varying tail distributions.Furthermore, for the non negative, independent and identical distributed random variables we investigate the asymptotic relation of large deviations for random sums in a frame of the subexponetial varying tail distributions.Finally, for the non negative, independent and identical distributed random variables, we prove the asymptotic relation of large deviation for random variables in finite time in the frame of the subexponetial varying tail distributions.Furthermore, we investigate the asymptotic relations of large deviations in the case of dependent random variables ( extended negative dependent ).Specifically, we find the lower bound of the asymptotic relation of large deviations for non random sums as well as random sums of negative random variables with long tail distributions.Additionally, we study the asymptotic relation of large deviations for non random sums as well as for random sums of dominately varying tail distributions for non negative random variables.Finally, we prove the asymptotic relation of large deviations for non random sums as well as for random sums in the frame of the intersection of long tail and of dominately varying tail distributions for non negative random variables.In the last part of the work, we provide applications of the above results in three stochastic models. In the renewal, the compound renewal and the negatively associated claim occurrences. Furthermore, for the conpound renewal model, we show that it satisfies the symptotic relation of the large deviation for non negative random variables with consistently varying tails.
Degree
thesis:*- Grantor dc:publisher
- University of the Aegean
- Year dc:date
- 2013
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Loukissas, Fotios
- Λουκίσσας, Φώτιος
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/28591