Almost sure convergence of the Kaczmarz algorithm with random measurements (Q1934652)

The authors consider the classical Kaczmarz projection algorithm (cf. [\textit{S. Kaczmarz}, Bull. Int. Acad. Polon. Sci. A 1937, 355--357 (1937; Zbl 0017.31703)]) in which the measurement vectors (corresponding to the right hand side components) are random vectors with some additional properties. Under these hypothesis the authors prove that Kaczmarz algorithm almost surely converges exponentially fast and provide quantitative bounds for the associated convergence rates.