← All problems

Kozma’s 2007 shape conjecture for radially excited random walk

Let (Xn)n≥0(X_n)_{n\ge 0} be the once-centrally-excited random walk on Zd\mathbb{Z}^d, with d≥2d\ge 2, in which the first visit to a site has a bias of strength β>0\beta>0 toward the origin and subsequent visits use the simple symmetric random walk. Writing Vn={X0,…,Xn}V_n=\{X_0,\ldots,X_n\} for the trace, the conjecture is that there is a constant κ>0\kappa>0 such that, with ωd\omega_d denoting the volume of the unit ball in Rd\mathbb{R}^d and rn=((d+1)n2dκωd)1/(d+1)r_n=\left(\frac{(d+1)n}{2d\kappa\omega_d}\right)^{1/(d+1)}, the rescaled trace converges almost surely to the Euclidean unit ball: rn−1Vn→B‾(0,1)r_n^{-1}V_n\to \overline{B}(0,1), equivalently, the trace is asymptotically a Euclidean ball centered at the origin with radius proportional to n1/(d+1)n^{1/(d+1)}.

References

Progress summary

Refreshed
Open · solution reported

A 2026 manuscript claims the conjectured round shape and recurrence, but no independent verification was found.

Kozma posed the conjecture in 2007: the visited set of the centrally excited walk should approach a Euclidean ball, with scale n1/(d+1)n^{1/(d+1)}. The new manuscript claims this, together with recurrence and a precise local-time profile, for d≥2d\ge 2.

Known results

  • Kozma reported recurrence in 2013, citing unpublished 2006 work; related unpublished work was discussed by Dembo, Huang, and Sidoravicius in 2014, and the predicted radius was recorded by Benjamini, Duminil-Copin, Kozma, and Lucas in 2020.

2026 claimed solution

The manuscript claims almost-sure convergence of the visited set to a Euclidean ball of radius rn=((d+1)n/(2dκωd))1/(d+1)r_n=\left((d+1)n/(2d\kappa\omega_d)\right)^{1/(d+1)}, plus recurrence and uniform convergence of local times to 2dκ(rn−∣x∣)+2d\kappa(r_n-|x|)_+. It describes Claude as assisting the authors, not independently producing the result. The claim is unrefereed and lacks independent mathematical assessment.

Current status (as of October 2026): A preprint claims the conjecture solved for d≥2d\ge 2, but the result remains unverified and is not independently established.

Sources

Solutions 0

No solutions have been posted yet.