← All problems

Fuglede’s conjecture for finite abelian groups

For every finite abelian group GG and every nonempty subset A⊆GA\subseteq G, AA tiles GG by translation if and only if AA is spectral. Here, AA tiles GG if there exists a set T⊆GT\subseteq G such that every element of GG has a unique representation a+ta+t with a∈Aa\in A and t∈Tt\in T; and AA is spectral if there exists a set Λ⊆G^\Lambda\subseteq\widehat{G} with ∣Λ∣=∣A∣|\Lambda|=|A| such that the restrictions {χ∣A:χ∈Λ}\{\chi|_A:\chi\in\Lambda\} form an orthogonal basis of the space of complex-valued functions on AA.

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

A new unrefereed preprint claims the conjecture in binary dimension eight, but the general finite-group problem and binary dimension nine remain open.

The conjecture asks whether, for every finite abelian group, having an orthogonal character basis is equivalent to tiling by translations. The all-dimensions problem remains unresolved.

Known results

  • For (Z2)n(\mathbb{Z}_2)^n, the conjecture holds for n≥10n \ge 10, fails for n≤6n \le 6, and is open for 7≤n≤97 \le n \le 9 (Fallon, Kiss, Mayeli, and Somlai, 2023).
  • The tile-to-spectral implication holds for every finite cyclic group of square-free order (Somlai, 2026).
  • The spectral-to-tiling implication is known for an infinite family of cyclic groups whose prime factors grow sufficiently rapidly (Somlai, 2026).
  • Z34\mathbb{Z}_3^4 is tile-to-spectral but not spectral-to-tiling (Fan and Zhang).

October 2026 dimension-eight advance

Samuel J. Ferguson and Joshua Iosevich claim the conjecture for binary dimension eight, with linear-subspace tiling complements for spectral sets and tiles, and identify an obstruction to extending their method to dimension nine. This is an unrefereed preprint and does not settle the remaining cases.

Current status (as of October 2026): The binary cases through dimension six and from dimension ten onward are settled, while the dimension-eight result is an unverified claim and dimensions seven and nine, along with the general finite-abelian-group problem, remain open.

Sources

Solutions 0

No solutions have been posted yet.