Random walk with barycentric self-interaction (Q634164)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Random walk with barycentric self-interaction
scientific article

    Statements

    Random walk with barycentric self-interaction (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    2 August 2011
    0 references
    Consider a \(d\)-dimensional self-interacting random walk \(X=(X_n)_{n=1,2,\ldots}\) with the property \[ \text{E}[X_{n+1}-X_n\,\big|\,X_n-G_n=x]\approx \rho\|x\|^{-\beta}\,\frac{x}{\|x\|},\qquad n=1,2,\ldots. \] Here, \[ G_n=\frac{1}{n}(X_1+\dots+X_n) \] is the barycentre of the previous path, and \(\varrho\in\mathbb{R}\) and \(\beta\geq0\) are constants. The symbol \(\approx\) means equality up to logarithmic corrections (in \(\|x\|\)). In addition, certain mild regularity assumptions are imposed on the transition probabilities. The authors show for \(\beta\in[0,1)\) and \(\rho>0\) that \(X\) is transient with a limiting random direction of super-diffusive escape. More precisely, they show that \(n^{-1/(1+\beta)}X_n\) converges almost surely to a random non-trivial vector. Furthermore, \(n^{-1/(1+\beta)}G_n\) converges also to a deterministic multiple of the same random vector. For all \(\beta\geq0\), the authors give (possibly sharp) upper bounds on the growth of \(\|X_n\|\). For the difference \(X_n-G_n\), the dichotomy between transience and recurrence is described rather completely. For \(\beta\in[0,1)\), the process \(X_n-G_n\) is recurrent for \(\rho>0\) and transient if \(\rho<0\). Under the additional assumption of some isotropy on the transition probabilities of \(X\), the authors show that, for \(\beta>1\), \(X_n-G_n\) is recurrent if \(d\leq2\) and transient if \(d\geq3\). For \(\beta=1\), a critical value \(\rho_0\) is computed such that \(X_n-G_n\) is recurrent for \(\rho\leq\rho_0\) and transient for \(\rho>\rho_0\).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    self-interacting random walk
    0 references
    self-avoiding walk
    0 references
    random walk avoiding its convex hull
    0 references
    random polymer
    0 references
    centre of mass
    0 references
    simple random walk
    0 references
    random walk average
    0 references
    limiting direction
    0 references
    law of large numbers
    0 references
    0 references
    0 references