This technical report revisits the convergence characteristics of GMRES in singular systems, confirming consistency in solutions.
In [Hayami K, Sugihara M. Numer Linear Algebra Appl. 2011; 18:449--469], the authors analyzed the convergence behaviour of the Generalized Minimal Residual (GMRES) method for the least squares problem minx ∈ Rⁿ \| b - A x \|₂², where A ∈ Rⁿˣⁿ may be singular and b ∈ Rⁿ, by decomposing the algorithm into the range R(A) $ and its orthogonal complement $ R(A)^⊥ $ components. However, we found that the proof of the fact that GMRES gives a least squares solution if $ R(A) = R(A^T) $ was not complete. In this paper, we will give a complete proof.
No takes yet. Share an insight, caveat, or question.
Hayami et al. (2020) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: