← All problems

Shortest rectifiable curve of unit width

Determine the minimum of Len⁡(γ)\operatorname{Len}(\gamma) over all rectifiable continuous curves γ:[a,b]→R3\gamma:[a,b]\to\mathbb{R}^3 whose convex hull K=conv⁡(γ([a,b]))K=\operatorname{conv}(\gamma([a,b])) has width 11, where w(K)=inf⁡u∈S2(max⁡x∈K⟨x,u⟩−min⁡x∈K⟨x,u⟩)w(K)=\inf_{u\in\mathbb{S}^2}\left(\max_{x\in K}\langle x,u\rangle-\min_{x\in K}\langle x,u\rangle\right). Equivalently, determine the shortest rectifiable curve γ\gamma satisfying w(conv⁡(γ([a,b])))=1w(\operatorname{conv}(\gamma([a,b])))=1, and establish whether the proposed eight-segment, cone-geodesic, and helical curve is globally minimizing.

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

The problem remains unsolved, but a new preprint reports a better candidate and computer-checked evidence about its geometry.

The problem asks for the shortest rectifiable curve in R3\mathbb{R}^3 whose convex hull has width 11. Zalgaller’s L3L_3 candidate was conjectured to be optimal, but no proof of global minimality is known.

Known results

  • Zalgaller, 1994: L3L_3 has length approximately 3.9215453.921545 and was conjectured shortest.
  • Ghomi, 2016: every rectifiable curve satisfies L/w≥3.7669L/w\ge 3.7669, while L3L_3 gives L/w≤3.9215L/w\le 3.9215.
  • Ghomi, 2016: disproved Zalgaller’s closed candidate L5L_5 and constructed a closed curve with L/w<5.1151L/w<5.1151.
  • For the general problem, these bounds leave global optimality unresolved.

October 2026 candidate

A preprint by Tinko Sebastian Bartels, Dennis Kevin Choy, Ruixiang Wu, and Tongxin Li reports an eight-segment/helical candidate, interval-arithmetic certification of unit width,. It also reports a disproof of a prior candidate. The candidate and its formal checks are substantive claimed progress, but global minimality remains unproved and the report is unverified.

Current status (as of October 2026): The problem remains open; bounds and new computationally certified candidate constructions are known, but no global shortestness proof has been established.

Sources

Solutions 0

No solutions have been posted yet.