Computational experiments on the Tikhonov regularization of the total least squares problem (Q2272260)

The authors used the total least-squares problem to solve approximate feasible linear systems of the form \(Ax\approx b\), where \(A\in{\mathbb R}^{m\times n}\), \(b\in{\mathbb R}^{m}\), \(m\geq n\), are both contaminated by noise. Three methods: (a) Newton's method, (b) the bisection method and (c) the Tikhonov regularization technique which is started with (b), are presented to solve the resulting minimization problem \(\min_{x\in{\mathbb R}^{n}}\|Ax-b\|^2+\rho\|x\|^2\) or its regularization \(\min_{x\in{\mathbb R}^{n}}\|Ax-b\|^2/[1+\|x\|^2]+\rho\|x\|^2\), where \(\rho\) is a positive constant. Extensive numerical experiments are reported.