← All problems

Alexander–Orbach conjecture

Let CIIC\mathcal C_{\mathrm{IIC}} be the incipient infinite cluster of critical percolation, and let pt(x,x)p_t(x,x) denote the return probability at time tt for simple random walk on CIIC\mathcal C_{\mathrm{IIC}}. The Alexander–Orbach conjecture asserts that, almost surely, the spectral dimension satisfies ds(CIIC):=−2lim⁡t→∞log⁡pt(x,x)log⁡t=43d_s(\mathcal C_{\mathrm{IIC}}):=-2\lim_{t\to\infty}\frac{\log p_t(x,x)}{\log t}=\frac{4}{3}, equivalently pt(x,x)=t−2/3+o(1)p_t(x,x)=t^{-2/3+o(1)} as t→∞t\to\infty.

References

Progress summary

Refreshed
Open · progress reported

A new paper reports counterexamples in special hierarchical models, but the conjecture for ordinary critical percolation remains open.

The Alexander--Orbach conjecture predicts spectral dimension 4/34/3 for random walks on critical percolation clusters. Its general status is unresolved, despite proofs in several high-dimensional or model-specific settings.

Known results

  • Kozma and Nachmias (2008, revised 2009): proved ds=4/3d_s=4/3 for the incipient infinite cluster in sufficiently high-dimensional and sufficiently spread-out critical percolation.
  • Kozma and Nachmias: extended this to finite-range and long-range models satisfying the strong triangle condition, including nearest-neighbor percolation for d≥19d\ge 19.
  • Barlow, Járai, Kumagai, and Slade (1986): established the prediction for critical branching-process clusters, but only subdiffusive behavior for the two-dimensional percolation cluster.
  • Later work proves ds=4/3d_s=4/3 in additional high-dimensional models, including oriented percolation and Gaussian-free-field level-set clusters.

October 2026 hierarchical-lattice claim

Ziyu Neroli reports that spectral dimensions of limiting diffusions are dense in (1,2)(1,2) for broad families of hierarchical lattices, contradicting the universal 4/34/3 prediction in those models. This is claimed progress, not a resolution of the conjecture for general critical percolation, and the supplied evidence does not independently verify the result.

Current status (as of October 2026): The 4/34/3 prediction is proved in several high-dimensional and model-specific regimes, while the general critical-percolation conjecture remains open; the hierarchical-lattice counterexample claim is unverified.

Sources

Solutions 0

No solutions have been posted yet.