Abstract:Suppose that (/>") is an infinitely divisible distribution on the non-negative integers having Levy measure (v n ). In this paper we derive a necessary and sufficient condition for the existence of the limit lim ; , -ooPn/"n-We also derive some other results on the asymptotic behaviour of the sequence (p n ) and apply some of our results to the theory of fluctuations of random walks. We obtain a necessary and sufficient condition for the first positive ladder epoch to belong to the domain of attraction of a sp… Show more
“…increments and assume that S n → −∞ a.s. Then an analogue of τ x is the stopping time ν x = min{n ≥ 1 : S n < −x}. Asymptotics for ν x have been studied in [12] for x = 0, and in [10], [6] for x > 0. In [12], it is shown that P{ν 0 > n} ∼ P{S n ≥ 0}/n if the latter is a subexponential sequence (see definition below).…”
Section: Theorem 481]) Says That X T → −∞ If and Only Ifmentioning
confidence: 99%
“…Asymptotics for ν x have been studied in [12] for x = 0, and in [10], [6] for x > 0. In [12], it is shown that P{ν 0 > n} ∼ P{S n ≥ 0}/n if the latter is a subexponential sequence (see definition below). In [10] and [6], the asymptotics for ν x have been found when x > 0.…”
Section: Theorem 481]) Says That X T → −∞ If and Only Ifmentioning
We study the exact asymptotics for the distribution of the first time τ x a Lévy process X t crosses a negative level −x. We prove that P(τ x > t) ∼ V (x)P(X t ≥ 0)/t as t → ∞ for a certain function V (x). Using known results for the large deviations of random walks we obtain asymptotics for P(τ x > t) explicitly in both light and heavy tailed cases. We also apply our results to find asymptotics for the distribution of the busy period in an M/G/1 queue.
“…increments and assume that S n → −∞ a.s. Then an analogue of τ x is the stopping time ν x = min{n ≥ 1 : S n < −x}. Asymptotics for ν x have been studied in [12] for x = 0, and in [10], [6] for x > 0. In [12], it is shown that P{ν 0 > n} ∼ P{S n ≥ 0}/n if the latter is a subexponential sequence (see definition below).…”
Section: Theorem 481]) Says That X T → −∞ If and Only Ifmentioning
confidence: 99%
“…Asymptotics for ν x have been studied in [12] for x = 0, and in [10], [6] for x > 0. In [12], it is shown that P{ν 0 > n} ∼ P{S n ≥ 0}/n if the latter is a subexponential sequence (see definition below). In [10] and [6], the asymptotics for ν x have been found when x > 0.…”
Section: Theorem 481]) Says That X T → −∞ If and Only Ifmentioning
We study the exact asymptotics for the distribution of the first time τ x a Lévy process X t crosses a negative level −x. We prove that P(τ x > t) ∼ V (x)P(X t ≥ 0)/t as t → ∞ for a certain function V (x). Using known results for the large deviations of random walks we obtain asymptotics for P(τ x > t) explicitly in both light and heavy tailed cases. We also apply our results to find asymptotics for the distribution of the busy period in an M/G/1 queue.
“…The situation when Eτ < ∞, which is a particular case of the relative stability, was considered by Embrechts and Hawkes [5]. There it has been shown that If the expectation EX is finite, then the condition ∞ k=1 k −1 P(S k ≤ 0) = ∞ is equivalent to the inequality EX ≤ 0, see again [19,Theorem 17.1].…”
Abstract. For a family of random walks {S (a) } satisfying ES (a) 1 = −a < 0 we consider ladder epochs τ (a) = min{k ≥ 1 : S (a) k < 0}. We study the asymptotic, as a → 0, behaviour of P(τ (a) > n) in the case when n = n(a) → ∞. As a consequence we obtain also the growth rates of the moments of τ (a) .
“…:In (nPn )-1--+ 00 as n --+ 00 • As in Embrechts and Hawkes (1982), we have that LC n is absolutely convergent, LC n =0 and c n ---qn ,(n _(0). It follows from (3.3) and (3.4) that…”
Section: The Levy Measure Of Glsdmentioning
confidence: 76%
“…We will therefore concentrate on asymptotic expressions for CY n as 11 _ 00. The following theorem of Embrechts and Hawkes (1982) illustrates that such expressions are closely related to the convolution behavior of Pn as n _ 00.…”
Section: Introduction the Generalized Logarithmic Series Distributionmentioning
It is shown that the generalized logarithmic series distribution is log-convex and hence infinitely divisible, and its Levy measure is determined asymptotically up to second order. An application to risk theory is given.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.