The Appell interpolation problem (Q654126): Difference between revisions

Let \(X\) denote the linear space of real functions defined on the interval \([a,b]\) with continuous derivatives of all necessary orders, and let \(L\) denote a linear functional on \(X\) such that \(L(1) \neq 0\). Then for any given real values \(w_i \;(0 \leq i \leq n)\) there exists a unique polynomial \(P_n\) of degree \(\leq n\) satisfying \(L(P_n^{(i)}) = i! w_i\), and in fact (Theorem 5) \(P_n\) is the sum of the \(w_i A_{L,i}\;(\text{for } i = 0, 1, \dots, n)\), where the polynomials \(A_{L,i}\) (termed Appell polynomials related to \(L\)) are defined by three conditions: (1) \(A_{L,0} = L(1)^{-1}\), (2) \(A_{L,i}^{(1)} = i A_{L,i-1}\), (3) \(L(A_{L,i}) = 0 \text{ for all } i \geq 1\). In particular, if one sets \(w_i = L(f^{(i)}) / i! \text{ (where } f \in X\)), then \(P_n\) depends on \(f\) and satisfies \(L(P_n^{(i)}) = L(f^{(i)}), 0 \leq i \leq n\). \(P_n\) is then termed the Appell interpolation polynomial for \(f\). The authors consider various choices of the functional \(L\) and reveal that \(P_n\) coincides, for example, with the classical interpolation polynomials named after Taylor, Bernoulli, Euler, Hermite and Laguerre. They also study generalizations of these polynomials (by attaching a weight function to \(L\)) and investigate the remainder term, and finally provide three numerical calculations of the interpolation error.