A generalization of the Ostrowski integral inequality for mappings whose derivatives belong to \(L_p[a,b]\) and applications in numerical integration (Q5934257)

The author proves the following generalization of the Ostrowski inequality, which has natural applications for quadrature formulae in numerical analysis. Let \(I_{k}:a=x_{0}<x_{1}<\dots <x_{k-1}<x_{k}=b\) be a division of the interval \([a,b]\) and \(\alpha _{i}\quad (i=0,\dots ,k+1)\) be ``\(k+2\)'' points so that \(\alpha _{0}=a,\quad \alpha _{i}\in [x_{i-1},x_{i}] (i=1,\dots ,k)\) and \(\alpha _{k+1}=b\) . If \(f:[a,b]\rightarrow \mathbb{R}\) is absolutely continuous on \([a,b]\), then we have the inequality \[ \left|\int\limits_{a}^{b}f(x)dx-\sum\limits_{i=0}^{k}(\alpha _{i+1}-\alpha _{i})f(x_{i})\right|\leq \frac{\nu (h)(b-a)^{\frac{1}{q}}}{(q+1)^{\frac{1}{q }}}\left\|f'\right\|_{p}, \] where \(\nu (h)=\max \{x_{i+1}-x_{i}\mid i=0,\dots ,n\}\). \(p>1\), \(\frac{1}{p}+\frac{1}{q}=1\) and \(\|\cdot\|_{p}\) is the usual \(L_{p}[a,b]\)-norm.