{"id":"arxiv/2608.04360","url":"https://arxiv.org/abs/2608.04360","tags":["math.NA"],"score":0,"title":"The Refined Joint Bidiagonalization Method and an Implicitly Restarted Algorithm for Large GSVD Computations","domain":"math.NA","source":"arxiv","authors":["Kaixiao Fang","Zhongxiao Jia"],"reviews":0,"subject":"math.NA","summary":"We make a convergence analysis on the joint bidiagonalization (JBD) method that computes several extreme generalized singular value decomposition (GSVD) components of a regular matrix pair $\\{A,L\\}$, and show that the right and left Ritz vectors obtained by it may converge erratically and even may fail to converge, while Ritz values converge. These convergence results hold for a class of general Rayleigh--Ritz projection methods for the GSVD problem under the hypothesis that the deviation of a desired right generalized singular vector from the right subspace tends to zero. We prove the…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-08-05","reactions":{}}