Some generalizations of the Euler-Knopp transformation (Q1073530): Difference between revisions

Let \(\{a_ k\}\) and \(\{b_ k\}\) be two sequences such that the series \(f(t)=\sum^{\infty}_{k=0}a_ kt^ k\) represents an analytic function within a circle of convergence with finite radius R and \[ b_ k=b_ k^{(\alpha,\nu)}(\lambda,\mu)=(k!/(k+\nu)!)\int^{\infty}_{0}e^{- \lambda t}t^{\alpha}L_ k^{(\alpha)}(t)g(\mu t)dt,\quad Re \lambda >-1, \] where \(L_ k^{(\alpha)}(t)\) are Laguerre polynomials and g(t) is a given function. The authors prove that \[ \sum^{\infty}_{k=0}a_ kb_ k^{(\alpha,\nu)}(\lambda,\mu)=\sum^{\infty}_{k=0}A_ k^{(\nu)}(p)(1-p)^{-k}b_ k^{(\alpha,\nu)}(\lambda ',\mu '), \] where \(\lambda '=p+\lambda (1-p),\mu '=\mu (1-p)\) and \(A_ k^{(\nu)}(p)=\sum^{k}_{j=0}\left( \begin{matrix} k+\nu \\ k-j\end{matrix} \right)(-p)^{k-j}a_ j,\) for all values of \(\alpha,\nu,\lambda,\mu,a_ k\) and \(p<1\) for which both sums exist and converge. Five special cases of the function g(t) are considered and the corresponding series acceleration formulas are obtained. The choice of the acceleration parameter p which ensures the most rapid rate of convergence is discussed and examples, showing some interesting applications of the above transformation, are given.