The Kazdan-Warner equation on a finite graph with a Neumann boundary condition
We study the Kazdan-Warner equation $Δu = c - h \mathrm{e}^u$ on a finite connected domain $Ω$ of a locally finite graph, with Neumann boundary data $\partial u / \partialν= g$. We reduce the boundary condition to homogeneous form and use variational methods and upper and lower solutions to study solvability. Let $c_g \mathrel{:=} |Ω|^{-1}\int_{\partialΩ} g\, \mathrm{d}μ$, and let $ψ$ solve $Δψ= c_g$ with boundary data $\partialψ/ \partialν= g$ and zero mean on $Ω$. When $c = c_g$ and $h \not\equiv 0$, a solution exists if and only if $h$ changes sign and $\int_Ωh \mathrm{e}^ψ\, \mathrm{d}μ< 0$. When $c > c_g$, a solution exists if and only if $h$ is positive somewhere in $Ω$. When $c < c_g$, solvability requires $\int_Ωh \mathrm{e}^ψ\, \mathrm{d}μ< 0$. Under this condition there is a threshold $-\infty \le c_-(h) < c_g$ such that a solution exists for $c_-(h) < c < c_g$ and no solution exists for $c < c_-(h)$. If $h \le 0$ and $h \not\equiv 0$, a solution exists for every $c < c_g$.
Authors
- Wenyuan Ma (ORCID: https://orcid.org/0009-0008-3749-9016)
- Anfeng Yuan
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.48550/arxiv.2611.23679
- Primary Topic
- Nonlinear Partial Differential Equations
- Type
- preprint