On the measure of approximation for some linear means of trigonometric Fourier series (Q5961550)

The author presents a new general method for solving the Kolmogorov-Nikol'skij problem. It is known that the generalized Rogosinski means are defined by \(\varphi_j(t):=\cos(j- 1/2)\pi t\) \((j\in\mathbb{N})\) and these functions are orthogonal on \([0,1]\), furthermore \(\varphi_j(0)= 1\), \(\varphi_j(1)=0\) for all \(j\in\mathbb{N}\). These facts inspired the author to consider arbitrary means NEWLINE\[NEWLINEU_n(f,x):= {A_0\over 2}+\sum^n_{k=1} \varphi(k/n)(A_k\cos kx+ B_k\sin kx),\tag{1}NEWLINE\]NEWLINE where \(\varphi\) has the Fourier representation NEWLINE\[NEWLINE\varphi= \sum^\infty_{j=1} a_j\varphi_j,\quad a_j:= 2\int^1_0\varphi\cdot\varphi_j.\tag{2}NEWLINE\]NEWLINE The author shows that some properties of the generalized Rogosinski means defined by \(\varphi_j\) also hold for arbitrary summation method (1), provided the representation (2) is valid.NEWLINENEWLINENEWLINESome applications are also presented.