Lemniscate constant

From testwiki
Jump to navigation Jump to search

Template:Short description Template:Redirect

Lemniscate of Bernoulli

In mathematics, the lemniscate constant Template:Mvar is a transcendental mathematical constant that is the ratio of the perimeter of Bernoulli's lemniscate to its diameter, analogous to the definition of [[Pi|Template:Pi]] for the circle.[1] Equivalently, the perimeter of the lemniscate (x2+y2)2=x2−y2 is 2ϖ. The lemniscate constant is closely related to the lemniscate elliptic functions and is approximately equal to 2.62205755.[2] It also appears in evaluation of the gamma and beta function at certain rational values. The symbol Template:Mvar is a cursive variant of Template:Mvar known as variant pi represented in Unicode by the character Template:Unichar.

Sometimes the quantities Template:Math or Template:Math are referred to as the lemniscate constant.[3][4]

History

Gauss's constant, denoted by G, is equal to Template:Math[5] and named after Carl Friedrich Gauss, who calculated it via the arithmetic–geometric mean as 1/M(1,2).Template:Sfn By 1799, Gauss had two proofs of the theorem that M(1,2)=π/ϖ where ϖ is the lemniscate constant.[6]

John Todd named two more lemniscate constants, the first lemniscate constant Template:Math and the second lemniscate constant Template:Math.[7][8][9]

The lemniscate constant ϖ and Todd's first lemniscate constant A were proven transcendental by Carl Ludwig Siegel in 1932 and later by Theodor Schneider in 1937 and Todd's second lemniscate constant B and Gauss's constant G were proven transcendental by Theodor Schneider in 1941.Template:R[10][11] In 1975, Gregory Chudnovsky proved that π and ϖ are algebraically independent over ℚ, which implies that A and B are algebraically independent as well.[12][13] But the set {π,M(1,1/2),M′(1,1/2)} (where the prime denotes the derivative with respect to the second variable) is not algebraically independent over ℚ.[14] In fact, π=22M3(1,12)M′(1,12)=1G3M′(1,12). In 1996, Yuri Nesterenko proved that Template:Math, π, and ϖ are algebraically independent over ℚ.[15]

As of 2025 over 2 trillion digits of this constant have been calculated using y-cruncher.[16]

Forms

Usually, ϖ is defined by the first equality below, but it has many equivalent forms:[17]

ϖ=2∫01dt1−t4=2∫0∞dt1+t4=∫01dtt−t3=∫1∞dtt3−t=4∫0∞(1+t44−t)dt=22∫011−t44dt=3∫011−t4dt=2K(i)=12B(14,12)=122B(14,14)=Γ(1/4)222π=2−24ζ(3/4)2ζ(1/4)2=2.62205755429211981046483958989111941…,

where Template:Mvar is the complete elliptic integral of the first kind with modulus Template:Mvar, Template:Math is the beta function, Template:Math is the gamma function and Template:Mvar is the Riemann zeta function.

The lemniscate constant can also be computed by the arithmetic–geometric mean M,

ϖ=πM(1,2).

Gauss's constant is typically defined as the reciprocal of the arithmetic–geometric mean of 1 and the square root of 2, after his calculation of M(1,2) published in 1800:Template:SfnG=1M(1,2)John Todd's lemniscate constants may be given in terms of the beta function B: A=ϖ2=14B(14,12),B=π2ϖ=14B(12,34).

As a special value of L-functions

β′(0)=log⁡ϖπ

which is analogous to

ζ′(0)=log⁡12π

where β is the Dirichlet beta function and ζ is the Riemann zeta function.[18]

Analogously to the Leibniz formula for π, β(1)=∑n=1∞χ(n)n=π4, we have[19][20][21][22][23] L(E,1)=∑n=1∞ν(n)n=ϖ4 where L is the L-function of the elliptic curve E:y2=x3−x over ℚ; this means that ν is the multiplicative function given by ν(pn)={p−𝒩p,p∈ℙ,n=10,p=2,n≥2ν(p)ν(pn−1)−pν(pn−2),p∈ℙ∖{2},n≥2 where 𝒩p is the number of solutions of the congruence a3−a≡b2(mod⁡p),p∈ℙ in variables a,b that are non-negative integers (ℙ is the set of all primes). Equivalently, ν is given by F(τ)=η(4τ)2η(8τ)2=∑n=1∞ν(n)qn,q=e2πiτ where τ∈ℂ such that ℑ⁡τ>0 and η is the eta function.[24][25][26] The above result can be equivalently written as ∑n=1∞ν(n)ne−2πn/32=ϖ8 (the number 32 is the conductor of E) and also tells us that the BSD conjecture is true for the above E.[27] The first few values of ν are given by the following table; if 1≤n≤113 such that n doesn't appear in the table, then ν(n)=0: nν(n)nν(n)1153145−261−109−365−1213673−617281925−185−429−10891037−297184110101−2456109649−7113−14

As a special value of other functions

Let Δ be the minimal weight level 1 new form. Then[28] Δ(i)=164(ϖπ)12. The q-coefficient of Δ is the Ramanujan tau function.

Series

Viète's formula for Template:Mvar can be written:

2π=12⋅12+1212⋅12+1212+1212⋯

An analogous formula for Template:Mvar is:[29]

2ϖ=12⋅12+12/12⋅12+12/12+12/12⋯

The Wallis product for Template:Mvar is:

π2=∏n=1∞(1+1n)(−1)n+1=∏n=1∞(2n2n−1⋅2n2n+1)=(21⋅23)(43⋅45)(65⋅67)⋯

An analogous formula for Template:Mvar is:[30]

ϖ2=∏n=1∞(1+12n)(−1)n+1=∏n=1∞(4n−14n−2⋅4n4n+1)=(32⋅45)(76⋅89)(1110⋅1213)⋯

A related result for Gauss's constant (G=ϖ/π) is:[31]

ϖπ=∏n=1∞(4n−14n⋅4n+24n+1)=(34⋅65)(78⋅109)(1112⋅1413)⋯

An infinite series discovered by Gauss is:[32]

ϖπ=∑n=0∞(−1)n∏k=1n(2k−1)2(2k)2=1−1222+12⋅3222⋅42−12⋅32⋅5222⋅42⋅62+⋯

The Machin formula for Template:Mvar is 14π=4arctan⁡15−arctan⁡1239, and several similar formulas for Template:Mvar can be developed using trigonometric angle sum identities, e.g. Euler's formula 14π=arctan⁡12+arctan⁡13. Analogous formulas can be developed for Template:Mvar, including the following found by Gauss: 12ϖ=2arcsl⁡12+arcsl⁡723, where arcsl is the lemniscate arcsine.[33]

The lemniscate constant can be rapidly computed by the series[34][35]

ϖ=2−1/2π(∑n∈ℤe−πn2)2=21/4πe−π/12(∑n∈ℤ(−1)ne−πpn)2

where pn=12(3n2−n) (these are the generalized pentagonal numbers). Also[36]

∑m,n∈ℤe−2π(m2+mn+n2)=1+3ϖ121/8π.

In a spirit similar to that of the Basel problem,

∑z∈ℤ[i]∖{0}1z4=G4(i)=ϖ415

where ℤ[i] are the Gaussian integers and G4 is the Eisenstein series of weight Template:Tmath (see Lemniscate elliptic functions § Hurwitz numbers for a more general result).[37]

A related result is

∑n=1∞σ3(n)e−2πn=ϖ480π4−1240

where σ3 is the sum of positive divisors function.[38]

In 1842, Malmsten found

β′(1)=∑n=1∞(−1)n+1log⁡(2n+1)2n+1=π4(γ+2log⁡πϖ2)

where γ is Euler's constant and β(s) is the Dirichlet-Beta function.

The lemniscate constant is given by the rapidly converging series

ϖ=π324e−π3(∑n=−∞∞(−1)ne−2nπ(3n+1))2.

The constant is also given by the infinite product

ϖ=π∏m=1∞tanh2(πm2).

Also[39]

∑n=0∞(−1)n6635520n(4n)!n!4=2457/4ϖ2π2.

Continued fractions

A (generalized) continued fraction for Template:Mvar is π2=1+11+1⋅21+2⋅31+3⋅41+⋱ An analogous formula for Template:Mvar isTemplate:R ϖ2=1+12+2⋅32+4⋅52+6⋅72+⋱

Define Brouncker's continued fraction by[40] b(s)=s+122s+322s+522s+⋱,s>0. Let n≥0 except for the first equality where n≥1. Then[41][42] b(4n)=(4n+1)∏k=1n(4k−1)2(4k−3)(4k+1)πϖ2b(4n+1)=(2n+1)∏k=1n(2k)2(2k−1)(2k+1)4πb(4n+2)=(4n+1)∏k=1n(4k−3)(4k+1)(4k−1)2ϖ2πb(4n+3)=(2n+1)∏k=1n(2k−1)(2k+1)(2k)2π. For example, b(1)=4π,b(2)=ϖ2π,b(3)=π,b(4)=9πϖ2.

In fact, the values of b(1) and b(2), coupled with the functional equation b(s+2)=(s+1)2b(s), determine the values of b(n) for all n.

Simple continued fractions

Simple continued fractions for the lemniscate constant and related constants include[43][44] ϖ=[2,1,1,1,1,1,4,1,2,…],2ϖ=[5,4,10,2,1,2,3,29,…],ϖ2=[1,3,4,1,1,1,5,2,…],ϖπ=[0,1,5,21,3,4,14,…].

Integrals

A geometric representation of ϖ/2 and ϖ/2

The lemniscate constant Template:Mvar is related to the area under the curve x4+y4=1. Defining πn:=B(1n,1n), twice the area in the positive quadrant under the curve xn+yn=1 is 2∫011−xnndx=1nπn. In the quartic case, 14π4=12ϖ.

In 1842, Malmsten discovered that[45]

∫01log⁡(−log⁡x)1+x2dx=π2log⁡πϖ2.

Furthermore, ∫0∞tanh⁡xxe−xdx=log⁡ϖ2π

and[46]

∫0∞e−x4dx=2ϖ2π4,analogous to∫0∞e−x2dx=π2, a form of Gaussian integral.

The lemniscate constant appears in the evaluation of the integrals

πϖ=∫0π2sin⁡(x)dx=∫0π2cos⁡(x)dx

ϖπ=∫0∞dxcosh⁡(πx)

John Todd's lemniscate constants are defined by integrals:Template:R

A=∫01dx1−x4

B=∫01x2dx1−x4

Circumference of an ellipse

The lemniscate constant satisfies the equationTemplate:Sfn

πϖ=2∫01x2dx1−x4

Euler discovered in 1738 that for the rectangular elastica (first and second lemniscate constants)[47]Template:Sfn

arc length⋅height=A⋅B=∫01dx1−x4⋅∫01x2dx1−x4=ϖ2⋅π2ϖ=π4

Now considering the circumference C of the ellipse with axes 2 and 1, satisfying 2x2+4y2=1, Stirling noted thatTemplate:Sfn

C2=∫01dx1−x4+∫01x2dx1−x4

Hence the full circumference is

C=πϖ+ϖ=3.820197789…

This is also the arc length of the sine curve on half a period:[48]

C=∫0π1+cos2(x)dx

Other limits

Analogously to 2π=limn→∞|(2n)!B2n|12n where Bn are Bernoulli numbers, we have 2ϖ=limn→∞((4n)!H4n)14n where Hn are Hurwitz numbers.

Notes

Template:Reflist

References

Template:Irrational number

  1. ↑ See:
  2. ↑ See:
  3. ↑ Template:Cite web
  4. ↑ Template:Cite web
  5. ↑ Template:Cite web
  6. ↑ Neither of these proofs was rigorous from the modern point of view. See Template:Harvnb
  7. ↑ Template:Cite journal
  8. ↑ Template:Cite web and Template:Cite web
  9. ↑ Template:Dlmf
  10. ↑ In particular, Siegel proved that if G4(ω1,ω2) and G6(ω1,ω2) with Im⁡(ω2/ω1)>0 are algebraic, then ω1 or ω2 is transcendental. Here, G4 and G6 are Eisenstein series. The fact that ϖ is transcendental follows from G4(ϖ,ϖi)=1/15 and G6(ϖ,ϖi)=0. Template:Pb Template:Cite book Template:Pb Template:Cite journal
  11. ↑ In particular, Schneider proved that the beta function B(a,b) is transcendental for all a,b∈ℚ∖ℤ such that a+b∉ℤ0−. The fact that ϖ is transcendental follows from ϖ=12B(14,12) and similarly for Template:Mvar and Template:Mvar from B(12,34). Template:Pb Template:Cite journal
  12. ↑ G. V. Choodnovsky: Algebraic independence of constants connected with the functions of analysis, Notices of the AMS 22, 1975, p. A-486
  13. ↑ G. V. Chudnovsky: Contributions to The Theory of Transcendental Numbers, American Mathematical Society, 1984, p. 6
  14. ↑ Template:Cite book p. 45
  15. ↑ Template:Cite book
  16. ↑ Template:Cite web
  17. ↑ See:
  18. ↑ Template:Cite web
  19. ↑ Template:Cite book p. 31, formula (2.8.10)
  20. ↑ In fact, the series ∑n=1∞ν(n)ns converges for ℜ⁡s>5/6.
  21. ↑ Template:Cite book
  22. ↑ Template:Cite book
  23. ↑ Template:Cite web
  24. ↑ The function F is the unique weight 2 level 32 new form and it satisfies the functional equation
    F(−1τ)=−τ232F(τ132).
  25. ↑ The ν function is closely related to the ξ function which is the multiplicative function defined by
    ξ(pn)={𝒩p′,p∈ℙ,n=1ξ(pn−1)+χ(p)n,p∈ℙ,n≥2
    where 𝒩p′ is the number of solutions of the equation
    a2+b2=p,p∈ℙ
    in variables a,b that are non-negative integers (see Fermat's theorem on sums of two squares) and χ is the Dirichlet character from the Leibniz formula for π; also
    ∑d|nχ(d)=ξ(n)
    for any positive integer n where the sum extends only over positive divisors; the relation between ν and ξ is
    ∑k=0n(−1)kξ(4k+1)ξ(4n−4k+1)=ν(2n+1)
    where n is any non-negative integer.
  26. ↑ The ν function also appears in
    ∑z∈𝔾;zz‾=nz=ν(n)
    where n is any positive integer and 𝔾 is the set of all Gaussian integers of the form
    (−1)a±b−12(a±bi)
    where a is odd and b is even. The ξ function from the previous note satisfies
    |{z:z∈𝔾∧zz‾=n}|=ξ(n)
    where n is positive odd.
  27. ↑ Template:Cite journal
  28. ↑ Template:Cite web
  29. ↑ Levin (2006)
  30. ↑ Hyde (2014) proves the validity of a more general Wallis-like formula for clover curves; here the special case of the lemniscate is slightly transformed, for clarity.
  31. ↑ Template:Cite journal
  32. ↑ Template:Cite book p. 60
  33. ↑ Todd (1975)
  34. ↑ Template:Harvnb for the first equality. The second equality can be proved by using the pentagonal number theorem.
  35. ↑ Template:Cite book p. 326
  36. ↑ This formula can be proved by hypergeometric inversion: Let
    a⁡(q)=∑m,n∈ℤqm2+mn+n2
    where q∈ℂ with |q|<1. Then
    a⁡(q)=2F1(13,23,1,z)
    where
    q=exp⁡(−2π32F1(1/3,2/3,1,1−z)2F1(1/3,2/3,1,z))
    where z∈ℂ∖{0,1}. The formula in question follows from setting z=14(33−5).
  37. ↑ Template:Cite book p. 232
  38. ↑ Template:Cite web p. 11—13
  39. ↑ The formula follows from the hypergeometric transformation
    3F2(14,12,34,1,1,16z(1−z)2(1+z)4)=(1+z)2F1(12,12,1,z)2
    where z=λ(1+5i) and λ is the modular lambda function.
  40. ↑ Template:Cite book p. 140 (eq. 3.34), p. 153. There's an error on p. 153: 4[Γ(3+s/4)/Γ(1+s/4)]2 should be 4[Γ((3+s)/4)/Γ((1+s)/4)]2.
  41. ↑ Template:Cite book p. 146, 155
  42. ↑ Template:Cite book p. 36, eq. 24
  43. ↑ Template:Cite web
  44. ↑ Template:Cite web
  45. ↑ Template:Cite journal
  46. ↑ Template:Cite web
  47. ↑ Levien (2008)
  48. ↑ Template:Cite web In this paper M=1/G=π/ϖ and L=π/M=Gπ=ϖ.