Rate of convergence for certain families of summation-integral type operators (Q1877808)

The authors consider a class of linear positive operators on \(L_p[0,\infty)\) defined by \(V_n(f,x) := n \sum_{\nu=1}^\infty {n+\nu-1 \choose \nu} x^\nu (1+x)^{-n-\nu} \int_0^\infty \exp(-nt) (nt)^\nu (\nu !)^{-1} f(t) \,dt + (1+x)^{-n} f(0)\). In view of the positivity of the operators it is not surprising that they converge slowly towards the identity \(I\). The observation that the first monomial is mapped to itself allows to consider the iterative combinations \(V_{n,k} := I - (I-V_n)^k = \sum_{r=1}^k (-1)^{r+1} {k \choose r} V_n^r\) of the \(V_n\). The approximation properties of these operators, measured in the \(L_p[0,\infty)\) norm with \(1 \leq p < \infty\), are the main subject of the paper. In fact the authors claim error estimates for \(V_{n,k}\) in terms of high order moduli of smoothness of the function to be approximated. The proof is based on Steklov means and on a number of Lemmas given in the paper that are not evident because their proofs use estimates for quantities that have not been defined, such as \(\mu_{n,m}^{(1)}\) in Lemma 1 or \(i\) in eq.\ (4). The paper is concluded by a section called ``Some Examples''. However most of the operators considered here are not covered by the theory of the previous sections. Indeed the authors give saturation results and error estimates for the operator \(V_n\) mentioned above and for three similar but not identical operators. The proofs are once again not verifiable because of unexplained notation (\(S_{n,m}^{(1)}\), \(P_{n,m}^{(1)}\), and \(B_{n,m}^{(1)}\)).