Directed polymers in heavy-tail random environment (Q2189464)

In the paper, the authors considered the directed polymer model, which is known as an effective model for an interface in the Ising model with random interactions, and has been used to describe a stretched polymer interacting with an inhomogeneous solvent. The model is defined as follows. Let \(S\) be a nearest-neighbor simple symmetric random walk on \(\mathbb{Z}^{a}, d \geq 1,\) whose law is denoted by \(\mathbf{P},\) and let \(\left(\omega_{i, x}\right)_{i \in \mathbb{N}, \, x \in \mathbb{Z}^{d}}\) be a field of i.i.d. random variables (the environment) with law \(\mathbb{P}(\omega\) will denote a random variable which has the common distribution of the \(\omega_{i, x}\) ). The directed random walk \(\left(i, S_{i}\right)_{i \in \mathbb{N}_{0}}\) represents a polymer trajectory and interacts with its environment via a coupling parameter \(\beta>0\) (the inverse temperature). The model is defined through a Gibbs measure, \( \frac{d \mathbf{P}_{n, \beta}^{\omega}}{d \mathbf{P}}(s):=\frac{1}{\mathbf{Z}_{n \beta}^{\omega}} \exp \left(\beta \sum_{i=1}^{n} \omega_{i, s_{i}}\right)\) where \(\mathbf{Z}_{n \beta}^{\omega}\) is the partition function of the model. The paper deals with the dimension \(1+1\) when the environment is heavy-tailed, with a decay exponent \(\alpha \in(0,2).\) In this case all possible scaling limits of the model in the weak-coupling regime are given. For \(\alpha \in(1 / 2,2),\) its shown that all possible transversal fluctuations \(\sqrt{n} \leq h_{n} \leq n\) can be achieved by tuning properly \(\beta_{n},\) allowing to interpolate between all superdiffusive scales. The scaling limit of the model is determined, by identifing five different regimes. For \(\alpha<1/2,\) it is proven that there are only two regimes: the transversal fluctuations are either \(\sqrt{n}\) or \(n.\) The proofs based on the entropy-controlled last-passage percolation, introduced in [the authors, Ann. Appl. Probab. 29, No. 3, 1878--1903 (2019; Zbl 1467.60072)].