Gnedin–Olshanski conjecture on the Martin boundary of the zigzag graph

At least 10 years old · documented by

Let Z\mathcal{Z} be the zigzag graph, let ∂MZ\partial_{M}\mathcal{Z} and ∂min⁡Z\partial_{\min}\mathcal{Z} denote its Martin and minimal entrance boundaries, and let U(2)\mathcal{U}^{(2)} be the space of pairs (U↑,U↓)(U_{\uparrow},U_{\downarrow}) of disjoint open subsets of [0,1][0,1] arising in the boundary parametrization. For each composition λ\lambda of nn, define Uλ=(U↑(λ),U↓(λ))∈U(2)U_{\lambda}=(U_{\uparrow}(\lambda),U_{\downarrow}(\lambda))\in\mathcal{U}^{(2)} by

U↑(λ)=int⁡(⋃i∉des⁡(λ)Ii),U↓(λ)=int⁡(⋃i∈des⁡(λ)Ii),U_{\uparrow}(\lambda)=\operatorname{int}\left(\bigcup_{i\notin\operatorname{des}(\lambda)} I_i\right),\qquad U_{\downarrow}(\lambda)=\operatorname{int}\left(\bigcup_{i\in\operatorname{des}(\lambda)} I_i\right),

where Is=[(s−1)/(n−1),s/(n−1)]I_s=[(s-1)/(n-1),s/(n-1)] for s≤n−1s\leq n-1. Write Kμ(λ)K_{\mu}(\lambda) for the Martin kernel indexed by μ∈Z\mu\in\mathcal{Z}, and p(U↑,U↓)(μ)p_{(U_{\uparrow},U_{\downarrow})}(\mu) for the corresponding boundary transition probability. Gnedin–Olshanski conjecture. For every sequence (λn)n≥1(\lambda_n)_{n\geq1}, the following hold: (a) (λn)(\lambda_n) belongs to ∂MZ\partial_M\mathcal{Z} if and only if UλnU_{\lambda_n} converges in U(2)\mathcal{U}^{(2)}; (b) Uλn→(U↑,U↓)U_{\lambda_n}\to(U_{\uparrow},U_{\downarrow}) in U(2)\mathcal{U}^{(2)} if and only if Kμ(λn)→p(U↑,U↓)(μ)K_{\mu}(\lambda_n)\to p_{(U_{\uparrow},U_{\downarrow})}(\mu) for every μ∈Z\mu\in\mathcal{Z}; and (c) the Martin boundary coincides with the minimal boundary, ∂MZ=U(2)\partial_M\mathcal{Z}=\mathcal{U}^{(2)}. The conjecture identifies all Martin-boundary limits with the already parametrized minimal boundary and characterizes convergence through both the geometric parameters UλnU_{\lambda_n} and the Martin kernels. The source attributes the problem to Conjecture 45 of Gnedin and Olshanski; no resolution is given here.

References

Primary source

Pierre Tarrago, “Zigzag diagrams and Martin boundary”, arXiv:1501.07087 (2018).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.