Critical logarithmic central limit conjecture for the on-line nearest-neighbour graph

About 19 years old · traced to

Let \d835𝒪d,α(\d835𝒰n)\d835𝒪^{d,α}(\d835𝒰_n) be the total power-weighted edge length of the on-line nearest-neighbour graph on nn uniform points in the unit cube, and let \d835𝒩(0,σ2)\d835𝒩(0,σ^2) denote the normal distribution with mean zero and variance σ2σ^2. Write \d835⟶d\d835\stackrel{d}{\longrightarrow} for convergence in distribution.

Critical logarithmic CLT conjecture. Let d∈Nd\in\mathbb{N}. There exists a constant σd,d/22∈(0,∞)σ_{d,d/2}^2\in(0,\infty) such that

(log⁡n)−1/2𝒪~d,d/2(𝒰n)⟶d𝒩(0,σd,d/22),n→∞.(\log n)^{-1/2}\tilde{𝒪}^{d,d/2}(𝒰_n)\stackrel{d}{\longrightarrow}𝒩(0,σ_{d,d/2}^2),\qquad n\to\infty.

The conjecture concerns the critical exponent α=d/2α=d/2, which the surrounding results identify as exceptional. In particular, it predicts Gaussian fluctuations with logarithmic rather than polynomial normalization; the supplied text gives no resolution.

References

Primary source

Andrew R. Wade, “Asymptotic theory for the multidimensional random on-line nearest-neighbour graph”, arXiv:math/0702414 (2008).

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.