Rate of convergence of the critical point of the memory-$τ$ self-avoiding walk in dimensions $d>4$

Noe Kawamoto

Published 2023-06-24Version 1

We consider the spread-out models of the self-avoiding walk and its finite-memory version, called the memory-$\tau$ walk. For both models, each step is uniformly distributed over the d-dimensional box $\{x\in\mathbb Z^d:\|x\|_{\infty} \le L\}$. The critical point $p_c^\tau$ for the memory-$\tau$ walk is increasing in $\tau$ and converges to the critical point $p_c^\infty$ for the self-avoiding walk as $\tau\uparrow\infty$. The best estimate of the rate of convergence so far was obtained by Madras and Slade in [Lemma 6.8.6,The Self-Avoiding Walk,(Birkh\"auser, 1993)]: $p_c^{\infty}-p_c^{\tau} \le K\tau^{-(1+\delta)}$, where $\delta < (d-4)/2 \wedge 1$ and $K<\infty$ may depend on $d,~\delta,~L$ but not on $\tau$. By using the lace expansion, we show that there is a constant $C_d$ for $d>4$ and $L$ sufficiently large such that as $\tau\uparrow \infty$, \begin{equation} p_c^{\infty}-p_c^{\tau} =C_d\tau^{-\frac{d-2}{2}}+O(\tau^{-\frac{d-2}{2}}(\log \tau)^{-1}). \end{equation} Moreover, we show that as $L \uparrow \infty$, $C_d=\frac{2}{d-2}\left(\frac{d}{2\pi\Sigma_U^2}\right)^{d/2}L^{-d}+O(L^{-d-1})$ where $\Sigma_U^2$ is the variance of the uniform distribution over the $d$-dimensional box of side-length $2$.