Correlation lengths for random polymer models and for some renewal sequences (Q2461983)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Correlation lengths for random polymer models and for some renewal sequences
scientific article

    Statements

    Correlation lengths for random polymer models and for some renewal sequences (English)
    0 references
    0 references
    23 November 2007
    0 references
    This article is motivated by the following two problems: critical behaviour of the correlation lengths for directed polymers with (de-)pinning interactions; geometric convergence rates for renewal sequences. With respect to the first problem the main result of this article is that, in the limit where \(F\) tends to zero (i.e. when the parameters of the model are varied in such a way that the critical point is approached from the localized phase), the correlation length \(\xi\) is at most of order \(1/F\), for almost every disordered realization. An exponentially decaying upper bound, with a ''good'' control of the sub-exponential prefactor, is derived also for the disordered average of the two-point correlation function \[ | \mathbf{P}_{\infty,\omega}(S_{n+k}=0| S_n=0)-\mathbf{P}_{\infty,\omega}(S_{n+k}=0)| . \] Concerning the second problem, consider a renewal sequence \(\tau=\{ \tau_i \}_{i=0,1,2,\ldots}\) of law \(\mathbf{P}\) defined as follows: \(\tau_0=0\), and \(\tau_i-\tau_{i-1}\) are IID random variables with values in \(\mathbb{N}\) and probability distribution \(p(\cdot)\), where \(p(n)\geq 0\) and \(\sum_{n\in\mathbb{N}}p(n)=1\). The renewal theorem states that \[ u_n:=\mathbf{P}(n\in\tau) @>n\to\infty>> u_{\infty}:= \frac{1}{\sum_{n\in\mathbb{N}}np(n)}= \frac{1}{\mathbf{E}(\tau_1)}, \] with the convention that \(1/\infty=0\). With respect to the speed of convergence rate here it is known that, if \(b:=\sup\{ s>0: \sum_{n\in\mathbb{N}}e^{sn}p(n)<\infty \}>0\), then there exist \(r>0\) and \(C<\infty\) such that \(| u_n-u_\infty| \leq Ce^{-rn}\). However, the relation between \(b\) and the largest possible \(r\) in the last estimate, i.e. \(r_{\max}\), is not known in general. As a corollary of the main result the following result, for the second problem is obtained: if the jump law \(p(\cdot)\) of the renewal sequence is of the form \(p(n)=a(b)\frac{L(n)}{n^{\alpha+1}}e^{-bn}\), with \(0\leq\alpha<\infty\), \(a(b)= (\sum_n L(n)n^{-(\alpha+1)}\exp (-bn))^{-1}\) and \(L(\cdot)\) a slowly varying function not depending on \(b\), then for \(b\) small one has that \(r_{\max}\gtrsim b\) and \(C\lesssim b^{-c}\) for some positive constant \(c\). In particular, this means that \(| u_n-u_\infty| \) states decaying exponentially with rate at least of order \(b\) as soon as \(n\geq 1/b\).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    pinning and wetting models
    0 references
    typical and average correlation lengths
    0 references
    critical exponents
    0 references
    renewal theory
    0 references
    exponential convergence rates
    0 references
    0 references