On constructing new interpolation formulas using linear operators and an operator type of quadrature rules (Q929930)

Let \(T,U\) be two linear operators mapped onto the function such that \(U(T(f))=f\), but \(T(U(f))\neq f\). In this paper, interpolating the function of type \(T(U(f))\) is presented in a general case. Let \(X\) be a linear vector space of dimension \(n\), its elements should be functions of a variable \(x\). The following interpolation holds for the arbitrary function \(g\) \[ g(x)=\sum_{i=1}^{n}S_{i}(g(x))\varphi_{i}(x)+E(x;g), \tag{1} \] where \(S_{i},\,\,i=\overline{1,n}\) are \(n\) given linear functionals defined on \(X\), \(E(x;g)\) denotes the error function and \(\{ \varphi_{i}(x)\}_{i=1}^{n}\) must be selected in such a way that \[ S_{j}(\varphi_{i}(x))=\delta_{j,i}=\begin{cases} 0,&\text{if } j\neq i,\\ 1, &\text{if } j=i.\end{cases} \] Replacing \(g(t)=U(f(t))\), respectively \(g(t)=T(f(t))\) in (1), the author obtains the following two formulas: \[ T(U(f(t)))=\sum_{i=1}^{n} S_{i}(U(f))T(\varphi_{i}(t))+T(E(t;U(f))),\tag{2} \] \[ U(T(f(t)))=f(t)=\sum_{i=1}^{n}S_{i}(T(f))U(\varphi_{i}(t))+U(E(t;T(f))).\tag{3} \] Many new formulas in theory of functions interpolation can be constructed by using these two basic rules (2) and (3). As a special case, the linear operators \(T(f)=\int_{\lambda}^{x} f(t) dt\) and \(U(f)=\frac{df(x)}{dx}\) are considered to interpolate the family of incomplete special functions. Three new examples of interpolation formulas together with their analytic error are also given as the special samples of the mentioned operator method. Finally, by using the basic interpolation (2), respectively (3), multiplying the positive measure \(w(x)\) on both sides of them and integrating along an interval, say \([a,b]\), the author obtains the following basic class of operator type quadrature rules \[ \int_{a}^{b}w(x) T(U(f(x)))\,dx=\sum_{i=1}^{n}S_{i}(U(f))\left(\int_{a}^{b}w(x) T(\varphi_{i}(x))\,dx\right)+\int_{a}^{b}w(x)T(E(x;U(f)))\,dx, \] \[ \int_{a}^{b}w(x) U(T(f(x)))\,dx= \sum_{i=1}^{n}S_{i}(T(f))\left(\int_a^bw(x)U(\varphi_i(x))\,dx\right)+ \int_a^bw(x)U(E(x;T(f)))\,dx. \]