← All problems

Lu–Zhu question on gradient estimates for the exponential m-Laplacian equation

Let (Mn,g)(M^n,g) be a complete, connected, noncompact Riemannian manifold with n≥2n\ge 2 and Ric⁡g≥0\operatorname{Ric}_g\ge 0. For every m>1m>1, q>0q>0, and λ>0\lambda>0, is every positive C1C^1 weak solution uu of −Δm,gu=λeu∣∇u∣q-\Delta_{m,g}u=\lambda e^u|\nabla u|^q on MM necessarily constant?

References

Progress summary

Refreshed
Open · progress reported

A new preprint claims to settle the Lu–Zhu question and extend the corresponding nonexistence theorem, but the advance has not been independently checked.

The question concerns whether positive entire weak solutions of the exponential gradient mm-Laplacian equation must be constant in the ranges left by earlier methods. It was raised by Lu and Zhu; the retrieved sources give no further dating.

Known results

  • Lu and Zhu’s theorem covers q>mq>m for every p∈Rp\in\mathbb{R}.
  • For m−1≤q<mm-1\leq q<m, the newer theorem claims constancy for every p∈Rp\in\mathbb{R}.
  • For 0<q<m−10<q<m-1, it proves constancy in a displayed subcritical range.
  • For m=2m=2, it claims to close the range 0<q<10<q<1 left by earlier work of Bidaut-Véron, García-Huidobro, and Véron.

October 2026 full-range preprint

Yuhua Sun, Zhen Wang, and Mengyu Zhang claim a unified Liouville theorem covering the remaining exponent ranges and thereby resolving the stated question. The result is supported here only by an unrefereed preprint, with no independent verification or error analysis retrieved.

Current status (as of October 2026): earlier results settle q>mq>m, while the claimed extension to the remaining ranges is unverified.

Sources

Solutions 0

No solutions have been posted yet.