Ramanujan's summation method (Q1373337)

\textit{G. H. Hardy} [Divergent series, Clarendon Press, Oxford (1949; Zbl 0032.05801), p. 327] comments briefly on Ramanujan's summation method that assigns to a series the associated Euler-Maclaurin constant that arises by applying the Euler-Maclaurin summation formula to the partial sums of the series. Hardy points out that the Ramanujan sum of a series is ambiguous because it depends on a parameter that appears as the lower limit of integration. This paper formulates Ramanujan summation in a manner that removes the ambiguity in Hardy's treatment, and paves the way for a rigorous and thorough investigation that is carried out in detail. Among many interesting results, it is shown that the Ramanujan sum assigned to a series that converges in the usual sense is not necessarily the same as the limit of its partial sums. For example, if \(s>1\) the convergent series \(\sum^\infty_{n=1} n^{-s}\) has sum \(\zeta(s)\), but the Ramanujan method assigns to this series the value \(\zeta (s)- 1/(s-1)\).