On the representation of measurable functions by absolutely convergent orthogonal spline series (Q2685175)

Earlier it was proved by the author that series with respect to the Franklin system cannot converge to \(+\infty\) on a set of positive measure. For the trigonometric system, a similar problem was a longstanding problem of \textit{N. N. Luzin} [Integral and trigonometric series. Moscow-Leningrad: Gostekhizdat (1951)] and was solved by \textit{S. V. Konyagin} [Math. Notes 44, No. 6, 910--920 (1988; Zbl 0688.42003); translation from Mat. Zametki 44, No. 6, 770--784 (1988)]. Afterward similar result was obtained by \textit{G. G. Gevorkyan} et al. [Acta Math. Hung. 162, No. 2, 604--617 (2020; Zbl 1474.42139)] for the general spline orthonormal systems of order \(m\) with dyadic rational knots. It is well-known that the Haar system and the Franklin system are both spline orthonormal systems of order \(1\), respectively \(2\) with dyadic rational knots. In this paper, the classical problem of representation of measurable and a.e. finite functions by series with respect to spline orthonormal systems of order \(m\) with dyadic rational knots are studied. Namely, it is proved that for an arbitrary a.e. finite measurable function \(f\), and arbitrary positive integer \(m\) there exists a series with respect to the spline orthonormal system of order \(m\) that converges absolutely a.e. to \(f\). It is proved also that the given function can be changed on a set of arbitrarily small measure so that the Fourier series will become uniformly absolutely convergent.