5
$\begingroup$

This question is cross-posted from https://math.stackexchange.com/questions/4580314/asymptotic-expansion-for-the-number-of-self-avoiding-random-walks.

Let $c_n$ be the number of self-avoiding random walks on $\mathbb{Z}^2$ of length $n$ starting at $0$. It is conjectured that $$ c_n \ \sim \ A\mu^nn^{\frac{11}{32}}, $$ for suitable real numbers $A$ and $\mu$, where $f(n)\sim g(n)$ means $\lim_{n\to \infty} \frac{f(n)}{g(n)}=1$.

My question is whether anything is known or conjectured about additional terms of an asymptotic expansion for $c_n$. In other words, if we rewrite the conjecture above as stating
$$ c_n = A\mu^nn^{\frac{11}{32}} + o(\mu^nn^{\frac{11}{32}})$$ one could ask whether it would be reasonable to conjecture that $$ c_n = A\mu^nn^{\frac{11}{32}} + B\mu^nn^{\gamma_2-1} + o(\mu^nn^{\gamma_2-1})$$ for a suitable rational number $\gamma_2<\frac{11}{32}$, and if so whether a value for this number is known ? Maybe more complicated functions involing oscillatory behaviour needs to be introduced in the subleading terms ? Any thoughts or references on this problem would be greatly appreciated.

$\endgroup$

1 Answer 1

2
$\begingroup$

Allow me to answer my own question since the reference I have found might be of use to others.

It turns out that A. Guttmann and A.R. Conway have analyzed this problem in som detail in the following article

  • Guttmann, A., Conway, A. "Square Lattice Self-Avoiding Walks and Polygons", Annals of Combinatorics 5, 319–345 (2001).

In this article the two authors are led to conjecture the asymtotic expansion

$$ c_nx_c^n \ \sim \ n^{\frac{11}{32}}(a_1+ a_2n^{-1} + a_3n^{-2} + ...) + (-1)^nn^{-\frac{3}{2}}(b_1 + b_3n^{-1} + ...)$$ where $x_c=1/\mu_c$. The conjecture is backed up by numerical evidence.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.