In this paper we consider heavy tailed Markov renewal processes and we prove that, suitably renormalised, they converge in law towards the α-stable regenerative set. We then apply these results to the strip wetting model which is a random walk S constrained above a wall and rewarded or penalized when it hits the strip [0, ∞) × [0, a] where a is a given positive number. The convergence result that we establish allows to characterize the scaling limit of this process at criticality.
Résumé.Dans cet article, nous considérons des processus de renouvellement markovien à queues lourdes. Nous montrons que, convenablement renormalisés, ils convergent vers l'ensemble régénératif d'indice α. Nous appliquons ces résultats à un modèle d'accrochage dans une bande. Dans ce modèle, une marche aléatoire S, contrainte à rester au-dessus d'un mur, est récompensée ou pénalisée lorsqu'est atteinte la bande [0, ∞) × [0, a] où a est un réel strictement positif. La convergence que nous établissons permet de caractériser les limites d'échelle de ce modèle au point critique. MSC: 60F77; 60K15; 60K20; 60K05; 82B27 Keywords: Heavy tailed Markov renewals processes; Scaling limits; Fluctuation theory for random walks; Regenerative sets k(x, y) := 1 μ(dy) n≥1 K x,dy (n) = n≥1 k x,y (n)is well defined.
484J. Sohier 2. Then sample J as a Markov chain on E starting from some initial point J 0 = x 0 ∈ E and with transition kernel3. Finally sample the increments (τ i − τ i−1 ) i≥1 as a sequence of independent, but not identically distributed random variables according to the conditional law:Markov renewal processes have been introduced independently by Lévy [18] and Smith [21], and their basic properties have been studied by Pyke [19], Cinlar [9] and Pyke and Schauffele [20] among others. A modern reference is Asmussen [1], VII, 4, which describes some applications related to these processes, in particular in the field of queuing theory. More recently, they have been widely used as a convenient tool to describe phenomena arising in models issued from statistical mechanics, and more particularly in models related to pinning models, such as the periodic wetting model ([8] and the monograph [16], Chapter 3) or the 1 + 1 dimensional field with Laplacian interaction ([4] and [5]).We will show our results in the case where the kernel K satisfies the following assumptions.Assumption 1.1. We make the following assumptions on the kernel K: