TOPICS
Search

Ramanujan g- and G-Functions


The Ramanujan g- and G-functions are defined for n>0 by the infinite products

 product_(k=1,3,5,...)^infty(1+e^(-kpisqrt(n)))=2^(1/4)e^(-pisqrt(n)/24)G_n
(1)
 product_(k=1,3,5,...)^infty(1-e^(-kpisqrt(n)))=2^(1/4)e^(-pisqrt(n)/24)g_n
(2)

(Ramanujan 1913-1914). Their values for positive rational numbers n are called Ramanujan's class invariants or Ramanujan-Weber class invariants (Andrews and Berndt 2005, p. 66). In terms of the Ramanujan chi-function, these definitions become

G_n=2^(-1/4)q^(-1/24)chi(q)
(3)
g_n=2^(-1/4)q^(-1/24)chi(-q)
(4)

(Andrews and Berndt 2005, p. 66), where q=e^(-pisqrt(n)). The functions satisfy the equalities

g_(4n)=2^(1/4)g_nG_n
(5)
G_n=G_(1/n)
(6)
g_n^(-1)=g_(4/n)
(7)
1/4=(g_nG_n)^8(G_n^8-g_n^8).
(8)

G_n and g_n can be derived using the theory of modular functions and can always be expressed as roots of algebraic equations when n is rational. They are related to the Weber modular functions.

For simplicity, Ramanujan (1913-1914) tabulated g_n for n even and G_n for n odd. For n>=1, (8) allows each function to be expressed in terms of the other as

g_n=2^(-1/8)(G_n^8+sqrt(G_n^(16)-G_n^(-8)))^(1/8)
(9)
G_n=2^(-1/8)(g_n^8+sqrt(g_n^(16)+g_n^(-8)))^(1/8).
(10)

For 0<n<1, the sign before the square root in the first formula is reversed, while the second formula is unchanged. Using (5) and the above two equations allows g_(4n) to be computed in terms of g_n or G_n

 g_(4n)={2^(1/8)g_n(g_n^8+sqrt(g_n^(16)+g_n^(-8)))^(1/8);  for n even; 2^(1/8)G_n(G_n^8+sqrt(G_n^(16)-G_n^(-8)))^(1/8);  for n odd.
(11)

In terms of the elliptic modulus k_n and complementary modulus k_n^',

G_n=(2k_nk_n^')^(-1/12)
(12)
g_n=((k_n^('2))/(2k_n))^(1/12).
(13)

Here,

 k_n=lambda^*(n)
(14)

is the elliptic lambda function, which gives the value of k for which

 (K^'(k))/(K(k))=sqrt(n).
(15)

For n>=1, solving for lambda^*(n) gives

lambda^*(n)=1/2[sqrt(1+G_n^(-12))-sqrt(1-G_n^(-12))]
(16)
lambda^*(n)=g_n^6[sqrt(g_n^(12)+g_n^(-12))-g_n^6].
(17)

For 0<n<1, the sign between the square roots in the first formula is positive, while the second formula is unchanged. Solving for G_n and g_n directly in terms of lambda^*(n) then gives

G_n=2^(-1/12)[lambda^*^2(n)-lambda^*^4(n)]^(-1/24)
(18)
g_n=2^(-1/12)[1/(lambda^*(n))-lambda^*(n)]^(1/12).
(19)

Examples of exact values include

G_5=((1+sqrt(5))/2)^(1/4)
(20)
G_9=((1+sqrt(3))/(sqrt(2)))^(1/3)
(21)
g_(10)=sqrt((1+sqrt(5))/2)
(22)
g_(18)=(sqrt(2)+sqrt(3))^(1/3)
(23)

(Ramanujan 1913-1914, p. 351). Further analytic values for small values of n can be found in Ramanujan (1913-1914), Borwein and Borwein (1987), and Berndt (1998, pp. 183-322), and have been compiled by Weisstein. Ramanujan (1913-1914) contains a typographical error labeling G_(465) as G_(265).


See also

Barnes G-Function, Elliptic Lambda Function, Weber Modular Function

Explore with Wolfram|Alpha

References

Andrews, G. E. and Berndt, B. C. Ramanujan's Lost Notebook, Part I. New York: Springer, p. 66, 2005.Berndt, B. C. "Class Invariants and Singular Moduli." Ch. 34 in Ramanujan's Notebooks, Part V. New York: Springer-Verlag, pp. 183-322, 1998.Borwein, J. M. and Borwein, P. B. Pi & the AGM: A Study in Analytic Number Theory and Computational Complexity. New York: Wiley, pp. 139 and 298, 1987.Ramanujan, S. "Modular Equations and Approximations to pi." Quart. J. Pure. Appl. Math. 45, 350-372, 1913-1914.

Referenced on Wolfram|Alpha

Ramanujan g- and G-Functions

Cite this as:

Weisstein, Eric W. "Ramanujan g- and G-Functions." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/Ramanujang-andG-Functions.html

Subject classifications