Hodge Laplacians on Weighted Simplicial Complexes: Forms, Closures, and Essential Self-Adjointness
Marwa Ennaceur, Amel Jadlaoui
TL;DR
The paper develops geometry-free operator bounds and explicit criteria guaranteeing essential self-adjointness for discrete Hodge Laplacians on weighted simplicial complexes, avoiding reliance on geometric completeness or curvature. It introduces an adjacency–potential decomposition, uses Schur tests, and employs a line-complex reduction along with a unitary intertwiner between skew and symmetric coboundaries on colorable complexes to obtain sharp norm bounds and ESA. For the edge block on unweighted d-regular graphs, it proves the universal bound ||̃Δ_{1,*}|| ≤ 4(d−1), with weighted extensions via a comparability constant, and extends ESA results to higher degrees under finite dual up/down degrees. In periodic lattices, Floquet–Bloch analysis yields exact constants of order 2d for the edge block, illustrating both the sharpness gap and the efficacy of translation-invariant techniques. Overall, the work advances a robust operator-theoretic framework for discrete Hodge Laplacians, providing concrete, geometry-free tools with broad applicability to graphs, complexes, and lattice systems.
Abstract
We establish explicit operator norm bounds and essential self-adjointness criteria for discrete Hodge Laplacians on weighted graphs and simplicial complexes. For unweighted $d$-regular graphs we prove the universal estimate $\|\widetildeΔ_{1,*}\|\le 4(d-1)$, and we provide weighted extensions with a sharp comparability constant. These bounds apply without geometric completeness or curvature assumptions and ensure essential self-adjointness on natural cores. The approach extends to higher degrees via dual up/down degrees, and we show a unitary equivalence between skew and symmetric models on colorable complexes. For periodic lattices we complement the universal bounds with exact Floquet--Bloch constants, typically of order $2d$, illustrating both the sharpness in growth and the generality of our method.
