Sharp Convergence Rates and Optimal Weights for Cimmino's Reflection Algorithm

Abstract

In this paper, Cimmino's classical reflection algorithm for solving the $n\times n$ nonsingular linear system $A\bx=\bb$ is analysed through the lens of spectral theory. Reformulating the weighted iteration as $\e^{(ν+1)}=M_w\,\e^{(ν)}$, where $M_w=I-A^{\top }{D}_{w}A$, the error is shown to contract by the spectral radius $\sprad(M_w)$ at every step, with a sharp, asymptotically tight bound. For $n=2$, a closed-form expression for the contraction factor is derived, \[ \sprad(M_w) \;=\; |1-μ| + \tfrac{1}{2}\sqrt{(w_1-w_2)^2 + 4w_1w_2\cos^2\!θ}, \] where $\hat{I}¼=(w_1+w_2)/2$ and $\hat{I}¸$ denotes the angle between the hyperplane normals. A central result of this paper is that the standard unit weights $w_1^{\ast }=w_2^{\ast }=1$ are \emph{globally optimal} over all positive weight pairs, uniquely achieving the minimum contraction factor $\sprad^*=|\cosθ|$ -- a quantity determined solely by the geometry of the hyperplane normals. The inter-normal angle $\hat{I}¸$ thus emerges as the single diagnostic parameter governing both convergence speed and weight selection. Extensions to a single-step convergence criterion at $\hat{I}¸=\ddot{I}€/2$ and to an exact spectral rate for general~$n$ are also established.