{"id":"arxiv/2610.19326","url":"https://arxiv.org/abs/2610.19326","tags":["math.NA"],"score":0,"title":"Solving a sequence of eigenvalue problems using snapshot and row subsampling","domain":"math.NA","source":"arxiv","authors":["Francesco Hrobat","Yuji Nakatsukasa"],"reviews":0,"subject":"math.NA","summary":"A sequence of eigenvalue problems requires solving $A_i\\vec{x}_i= λ_i\\vec{x}_i$ where $i \\in \\{1, \\dots, N\\}$, $N \\gg 1$ and (usually) $λ_i$ is the largest or smallest eigenvalue of $A_i$. Often, the matrices $A_i$ come from a discretization of a smooth map depending on a parameter $p \\rightarrow A(p)$, that is $A_i = A(p_i)$, so the matrices $A_i$ change gradually along the sequence. Under this latter assumption, we introduce an efficient algorithm for solving sequences of eigenvalue problems. Our algorithm combines the ideas of reduced order model and row subsampling, as done in [E. Jones…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-14","reactions":{}}