Tail probabilities of subadditive functionals of Lévy processes. (Q1872382)

Let \(X = X (t)\) denote a Lévy process without Gaussian component for \(t \geq 0\) on the real line and let \(\varphi \) denote a subadditive functional on the sample paths. The paper is concerned with the exceedance probability \(P (\varphi(X-\mu) > u)\), where \(\mu = \mu (t)\) is a deterministic curve. Under various conditions these probabilities behave like a known regularly varying function \(u \mapsto \varphi (u)\) if \(u\) tends to infinity. A proper choice of the functional \(\varphi \) yields concrete application. The authors establish the asymptotic decay of the probability of the following events: the time the process spends above zero, the last hitting time of zero, the supremum of the integral process greater than \(u\). The results include the decay of ruin probabilities.