Weighted approximation by modified Picard operators (Q2174937): Difference between revisions

In [Positivity 21, No. 3, 1189--1199 (2017; Zbl 1377.41011)] \textit{O. Agratini} et al. constructed the following modification of well-known Picard operators: \[ P_{n}^{\ast }\left( f;x\right) =\frac{\sqrt{n}}{2}\overset{\infty }{\underset{-\infty }{\int }}f\left( \alpha _{n}\left( x\right) +t\right) e^{-\sqrt{n}\left \vert t\right \vert }dt,\text{ \ }x\in\mathbb{R},\text{ \ }n\in\mathbb{N}\text{ with }n\geq n_{a} \tag{1} \] where \(\alpha _{n}\left( x\right) =x-\frac{1}{2a}\ln ( \frac{n}{n-(2a)^{2}}) \) and \(a>0\). Here, \(n_{a}\) stands for \(\left[ 4a^{2} \right] +1,\) where \(\left[ .\right] \) denotes the floor function. These operators reproduce the constant function \(1\) and a certain exponential function. In the paper under review, the authors prove complementary results for the operators defined in (1). The authors start by recalling the moment identities obtained in [loc. cit.]. Then, the existence of the operators is proved in weighted \(L_{p}\) spaces. Then, they give some results concerning, approximation in weighted \(L_{p}\) norm using a Korovkin-type approximation theorem and pointwise approximation. Finally, they prove approximation in a weighted \(L_{p}\) norm without using a Korovkin-type theorem.