{"id":"arxiv/2607.24138","url":"https://arxiv.org/abs/2607.24138","tags":["math.NA"],"score":0,"title":"Randomized Krylov-Projected Iterated Tikhonov Regularization for Large-Scale Ill-posed Problems Under A Posteriori Stopping Rule","domain":"math.NA","source":"arxiv","authors":["Ravi Verma","Harshit Bajpai","Ankik Kumar Giri"],"reviews":0,"subject":"math.NA","summary":"We introduce two novel randomized iterative regularization frameworks, termed \\texttt{RIGKT} and \\texttt{RIAT}, for solving large-scale linear ill-posed inverse problems governed by systems of equations. The proposed methods combine randomized iterated Tikhonov regularization with Krylov subspace projection techniques, utilizing Golub--Kahan bidiagonalization for general rectangular systems (\\texttt{RIGKT}) and Arnoldi decomposition for square systems (\\texttt{RIAT}). Unlike existing deterministic schemes that rely on fixed iteration counts, our framework incorporates randomized equation…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-27","reactions":{}}