mtommila / apfloat

Apfloat is a high performance arbitrary precision arithmetic library.
MIT License
56 stars 10 forks source link

Generalized Riemann zeta function #32

Open axkr opened 1 year ago

axkr commented 1 year ago

Is the "generalized Riemann zeta function" zeta(s,a) also implemented?

The generalized Riemann zeta function is identical to the hurwitz zeta function for Re(a)>0

for Re(a)<0 it is defined as: $$\sum_{k=0}^{\infty} { ((k+a)^2)^{-s/2} }$$

See:

mtommila commented 1 year ago

Not sure if that definition of "generalized Riemann zeta function" is used anywhere else than in Mathematica. Do you have any other references?

axkr commented 1 year ago

Don't know if this counts as another reference:

mtommila commented 1 year ago

Well, it's at least mentioned in some of the references there, although I'm not quite sure if the references are talking about the Hurwitz zeta function or that "generalized" zeta function

Is the "generalized zeta function" useful for something where the Hurwitz zeta function can't be used?

mtommila commented 8 months ago

There is apparently a formula that connects the Hurwitz zeta function and the "generalized Riemann zeta function"

https://functions.wolfram.com/ZetaFunctionsandPolylogarithms/Zeta2/02/

(Where "zeta" is the "generalized Riemann zeta" and "classical zeta" is the Hurwitz zeta)

But overall it's somewhat confusing that even on that page it claims that the "Hurwitz zeta function" has Mathematica notation Zeta[s, a] when this is not actually true.

Having gone through several of the algorithms that use ζ(s, a) on the functions.wolfram.com site, they seem to use the Hurwitz zeta function and not the "generalized Riemann zeta function" (which seems pretty bogus anyways at least due to the rather arbitrary looking rule "where any term with k + a = 0 is excluded").

axkr commented 8 months ago

Having gone through several of the algorithms that use ζ(s, a) on the functions.wolfram.com site, they seem to use the Hurwitz zeta function and not the "generalized Riemann zeta function" (which seems pretty bogus anyways at least due to the rather arbitrary looking rule "where any term with k + a = 0 is excluded").

I think there are historical reasons, Zeta is used in Mathematica since 1988, HurwitzZeta is introduced relatively late in 2008.