• Aucun résultat trouvé

On trigonometric expansions of the Weierstrass function $\wp(z)$ and applications

N/A
N/A
Protected

Academic year: 2021

Partager "On trigonometric expansions of the Weierstrass function $\wp(z)$ and applications"

Copied!
10
0
0

Texte intégral

(1)

HAL Id: hal-03226340

https://hal.archives-ouvertes.fr/hal-03226340

Preprint submitted on 17 May 2021

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

On trigonometric expansions of the Weierstrass function

℘(z) and applications

Abd Raouf Chouikha

To cite this version:

Abd Raouf Chouikha. On trigonometric expansions of the Weierstrass function ℘(z) and applications.

2021. �hal-03226340�

(2)

1

ON TRIGONOMETRIC EXPANSIONS OF THE WEIERSTRASS FUNCTION ℘(z) AND APPLICATIONS

ABD RAOUF CHOUIKHA

Abstract. Thanks to representation of the Jacobi theta functions

θ

j

(v, τ ) = f

j

(v, τ)f

j

(v + 1, τ ) we deduce a trigonometric expression of the Weierstrass function ℘(z, τ) with primitive periods (2, 2τ).

It allows us to find again more directly many known identities. In particular we derive the following identity

4℘(2z, 2τ) + 4η(2τ) = ℘(z, τ) + ℘(z + 1, τ) + 2η(τ) where

η(τ) = π

2

2 [ 1

6 + X

n≥1

1 (sinnπτ )

2

].

1. Introduction

In a preceding paper [2] we state the following representation for the Jacobi theta function

θ j (v, τ) = θ j (0, τ )f j (v, τ)f j (v + 1, τ ), j = 3, 4;

θ 1 (v, τ) = (π sin(πv) θ 1

(0, τ )f 1 (v, τ)f 1 (v + 1, τ );

θ 2 (v, τ) = cos(πv) θ 2 (0, τ) f 2 (v, τ ) f 2 (v + 1, τ ).

We derive their expansions as infinite products, their Fourier series expansions, as well as for log(f j (v, τ )) and for f f

j

j

(v, τ).

Let q = e iπτ , | q |< 1, we prove in particular ([2,Cor 2-5]) the functions θ 1 , θ 4

may be expressed as infinite products

θ 1 (v, τ) = (π sin πv) θ 1

(0, τ ) Y

k≥ 1

"

1 −

sin πv sin kπτ

2 #

θ 4 (v, τ ) = θ 4 (0, τ ) Y

k≥ 0

"

1 −

sin πv sin(k + 1 2 )πτ

2 #

here θ 4 is defined in the band | Imv |< 1 2 Imτ, while θ 1 is defined for

| Imv |< Imτ.

Moreover, we have carried out θ 4 (v, τ )

θ 4 (0, τ ) = f 4 (v, τ ) f 4 (−v, τ ) = f 4 (v, τ ) f 4 (v + 1, τ ) where

f 4 (v, τ) = Y

k≥ 0

1 −

sin πv sin(k + 1 2 )πτ

.

1

chouikha@math.univ-paris13.fr. 4, Cour des Quesblais 35430 Saint-Pere, France Key words and phrases. theta functions, elliptic functions, trigonometric expansions.

1

(3)

In this paper we are interested in various aspects of trigonometric expressions of the Weierstrass function ℘(z) with primitive periods (2, 2τ). We produce another one derived from the infinite product of the theta function θ 4 (v, τ )

℘(z + τ) = −η(τ) + π 2 X

k≥ 0

1 − cos πz cos(2k + 1)πτ [cos(2k + 1)πτ − cos πz] 2 . As a corollary one deduces some identities as

℘(z + τ + [

1 2, τ + 1) + η(τ + 1) = ℘(z + τ, τ) + η(τ).

2. Trigonometric expansions of ℘(z)

In this section we propose to achieve trigonometric expansions for the elliptic Weierstrass function ℘(z) analogous to that we did before for the Jacobi θ function.

Recall that ℘(z) which has primitive periods 2 and 2τ relative to g 2 and g 3 , may be written as ([1,4,6]

℘(z) = ℘(z; g 2 , g 3 ) = ℘(z, τ ) = 1 z 2 + X

m,n

1

(z − 2m − 2nτ) 2 − 1 (2m + 2nτ) 2

.

A direct consequence of the preceding definition is the fact that the Weierstrass elliptic function is an even function ℘(−z, τ) = ℘(z, τ ).

The original constructions of elliptic functions are due to Weierstrass and Jacobi [1]. Excellent approaches on the subject of elliptic functions are the classic book by Watson and Whittaker [7]. Useful reference handbooks with many details on tran- scendental functions including those used in this paper are provided by Bateman and Erdelyi, [4], which is freely available online.

We define the values of the Weierstrass elliptic function at the half-periods 1, 1 + τ, τ

e 1 (τ) = ℘(1, τ ), e 2 (τ) = ℘(1 + τ, τ), e 3 (τ) = ℘(τ, τ).

These e i obey the relations

e 1 + e 2 + e 3 = 0, e 1 e 2 + e 3 e 1 + e 2 e 3 = −g 2

4 , e 1 e 2 e 3 = g 3 4 . The Weierstrass elliptic function verifies the homogeneity relation

℘(z; g 2 , g 3 ) = µ 2 ℘(µz; g 2 µ 4 , g 3

µ 6 ).

Finally, when two of the roots e 1 , e 2 and e 3 coincide, the Weierstrass elliptic func- tion degenerates to a simply periodic function.

2

(4)

On the other hand, the Weierstrass function ℘(z, τ) is related to the theta func- tions θ i (v) where v = z 2 :

℘(z) = ( 1

2 ) 2 [−4η − d 2 logθ 1 (v) dv 2 ] η = η(τ) = − 1

12 θ

′′′

1 (0)

θ 1

(0) = π 2 2 [ 1

6 + X

n≥ 1

1 (sin nπτ) 2 ].

In the same way we have

℘(z + τ) = ( 1

2 ) 2 [−4η − d 2 logθ 4 (v) dv 2 ].

= ( 1

2 ) 2 [−4η − d 2 logf 4 (v)

dv 2 − d 2 logf 4 (−v) dv 2 ].

Therefore, we may deduce

e 3 (τ) = ℘(τ) = −η(τ) + π 2 X

k≥ 0

1

1 − cos(2k + 1)πτ , e 3 (τ) = ( π

2 ) 2

θ 1

′′′

(0)

3θ 1

(0) − θ 4

′′

(0) θ 4 (0)

. By the same way

e 2 (τ) = −η(τ + 1) + π 2 X

k≥ 0

1

1 + cos(2k + 1)πτ = ( π 2 ) 2

θ 1

′′′

(0)

3θ 1

(0) − θ

′′

3 (0) θ 3 (0)

.

e 1 (τ) = η(τ) + η(τ + 1) − 2π 2 X

k≥ 0

1

(sin(2k + 1)πτ) 2 = ( π 2 ) 2

θ

′′′

1 (0)

1 (0) − θ 2

′′

(0) θ 2 (0)

.

Notice the following theta function identity [7]

θ 1

′′′

(0)

θ

1 (0) = θ 2

′′

(0)

θ 2 (0) + θ 3

′′

(0)

θ 3 (0) + θ

′′

4 (0) θ 4 (0) .

On the other hand the Weierstrass function ℘(z, τ) may be expressed in different ways (see [5], [7]). In particular, we obtain the Fourier expansions

℘(z + 1, τ ) = −η + (π) 2

 1

4(cos πz) 2 − 2 X

k≥ 1

k(−1) k q 2 k

1 − q 2k cos kπz

℘(z + 1 + τ, τ) = −η − 2(π) 2 X

k≥ 1

k(−1) k q 2 k

1 − q 2 k cos kπz

℘(z + τ, τ) = −η − 2(π) 2 X

k≥ 1

kq 2k

1 − q 2 k cos kπz.

Other expansions will be obtained differently (see [7, p.183])

℘(z, τ ) = −η + (π) 2 4

X

k∈IN

1 [sin( πz 2 − kπτ)] 2

3

(5)

One trigonometric expansion called by [5] a first q-expansion where q = e iπτ : 1

2 ℘(z, τ ) = − 1

12 − 1

sin 2 πz − X

m,n≥ 1

nq

mn2

(cos nπz − 1).

This series is defined in the band | q |< e iπz <

|q|

1 , Imτ > 0.

[5] also defined a second q-expansion : 1

2 ℘(z, τ ) = − 1

12 + 2 X

n≥ 1

q 2 n

(1 − q 2n ) 2 − 2 X

n≥ 1

(−1) n nq n 1 − q 2n cos nπz .

Notice that the last expression yields an isomorphism between the multiplicative group of complex numbers and the complex points of the Tate curve parametrized by (℘(z, τ), ℘

(z, τ )).

Another alternative trigonometric expansion of ℘(z) is given by

Theorem 2-1 The Weierstrass elliptic function ℘(z) = ℘(z, τ ) with primi- tive periods 2 and 2τ and Imz < 1 2 Imτ may be expressed under the form

℘(z + τ ) = ℘ 1 (z) + ℘ 1 (−z) = −η(τ ) + π 2 X

k≥ 0

1 − cos πz cos(2k + 1)πτ [cos(2k + 1)πτ − cos πz] 2

℘(z) = −η(τ) + π 2 X

k≥ 0

1 − cos π(z − τ) cos(2k + 1)πτ [cos(2k + 1)πτ − cos π(z − τ)] 2 where

℘ 1 (z) = −η(τ) 2 + π 2

4 X

k≥ 0

1 − sin( πz 2 ) sin(k + 1 2 )πτ (sin( πz 2 ) − sin(k + 1 2 )πτ) 2 .

Proof Start from the following where 2v = z and ω = 1

℘(z + τ) = ( 1

2ω ) 2 [−4ηω − d 2 logθ 4 (v)

dv 2 ] = ℘ 1 (z) + ℘ 2 (z)

= ( 1

2ω ) 2 [−4ηω − d 2 log f 4 (v)

dv 2 + d 2 log f 4 (−v) dv 2 ].

By Corollary 3-5 of [2], deriving 1

θ 4

∂θ 4

∂v (v, τ ) = 1 f 4

∂f 4

∂v (v, τ ) + 1 f 4

∂f 4

∂v (−v, τ)

= −π X

k≥ 0

cos πv

sin(k + 1 2 )πτ − sin πv + π X

k≥ 0

cos πv sin(k + 1 2 )πτ + sin πv

= −π X

k≥ 0

sin 2πv sin(k + 1 2 )πτ 2

− (sin πv) 2 . We then obtain (using Maple for example)

d 2 logf 4 (v)

dv 2 = − π 2 sin (k + 1 2 )πτ

sin (π v) − 1

−2 + cos (k + 1 2 )πτ 2

+ 2 sin (k + 1 2 )πτ

sin (π v) + (cos (π v)) 2

4

(6)

= π 2 X

k≥ 0

1 − sin(πv) sin(k + 1 2 )πτ (sin(πv) − sin(k + 1 2 )πτ) 2 d 2 logf 4 (−v)

dv 2 = π 2 sin (k + 1 2 )πτ

sin (π v) + 1

−2 sin (k + 1 2 )πτ

sin (π v) − 2 + cos (k + 1 2 )πτ 2

+ (cos (π v)) 2

= π 2 X

k≥ 0

1 + sin(πv) sin(k + 1 2 )πτ (sin(πv) + sin(k + 1 2 )πτ) 2 .

d 2 logθ 4 (v) dv 2 =

2

− (cos (π v)) 2 − cos (k + 1 2 )πτ 2

+ 2 cos (k + 1 2 )πτ 2

(cos (π v)) 2 cos (k + 1 2 )πτ 4

− 2 cos (k + 1 2 )πτ 2

(cos (π v)) 2 + (cos (π v)) 4 . Since

−2 (cos (π v)) 2 − 2

cos

(k + 1 2 )πτ

2 + 4

cos

(k + 1

2 )πτ 2

(cos (π v)) 2

= −1 + cos (2 π v) cos ((2 k + 1) π τ ) therefore

d 2 logθ 4 (v)

dv 2 = X

k≥ 0

4 π 2 (cos (2 π v) cos ((2k + 1)π τ) − 1)

(cos ((2k + 1)π τ )) 2 − 2 cos (2 π v) cos ((2k + 1)π τ ) + (cos (2 π v)) 2

= 4π 2 X

k≥ 0

(cos (2 π v) cos ((2k + 1)τ) − 1) [cos(2k + 1)πτ − cos 2πv] 2 Then

℘(z+τ)+η = −π 2 X

k≥ 0

(cos (2 π v) cos ((2k + 1)πτ) − 1)

[cos(2k + 1)πτ − cos 2πv] 2 = π 2 X

k≥ 0

1 − (cos πz cos ((2k + 1)πτ)) [cos(2k + 1)πτ − cos πz] 2 . On the other hand one has ℘ 1 (−z) = ℘ 2 (z) and the following equality hold

4 X

k≥ 0

1 − (cos (π z) cos ((2k + 1)πτ )) [cos(2k + 1)πτ − cos πz] 2 = X

k≥ 0

1 − sin(π z 2 ) sin(k + 1 2 )πτ (sin(π z 2 ) − sin(k + 1 2 )πτ) 2 + X

k≥ 0

1 + sin(π z 2 ) sin(k + 1 2 )πτ (sin(π z 2 ) + sin(k + 1 2 )πτ) 2 .

(1) Or equivalently replacing z by z + 1 and τ by τ + 1

4 X

k≥ 0

1 − (cos (π z ) cos ((2k + 1)πτ)) [cos(2k + 1)πτ − cos πz] 2 = X

k≥ 0

1 − (−1) k cos(π z 2 ) cos(k + 1 2 )πτ (cos(π z 2 ) − (−1) k cos(k + 1 2 )πτ) 2

+ X

k≥ 0

1 + (−1) k cos(π z 2 ) cos(k + 1 2 )πτ (cos(π z 2 ) + (−1) k cos(k + 1 2 )πτ) 2

= X

k≥ 0

1 − cos(π z 2 ) cos(k + 1 2 )πτ (cos(π z 2 ) − cos(k + 1 2 )πτ) 2 + X

k≥ 0

1 + cos(π z 2 ) cos(k + 1 2 )πτ (cos(π z 2 ) + cos(k + 1 2 )πτ) 2 .

5

(7)

Replacing τ by τ + 1 2 one also obtains the identity 4 X

k≥ 1

1 − (cos (π z) cos (2kπτ)) [cos 2kπτ − cos πz] 2 = X

k≥ 1

1 − cos(π z 2 ) cos kπτ (cos(π z 2 ) − cos kπτ) 2 + X

k≥ 1

1 + cos(π z 2 ) cos kπτ (cos(π z 2 ) + cos kπτ ) 2 .

(2) Corollary 2-2 Consider the function

φ(z, τ) = X

k≥ 0

1 − (cos (π z) cos ((2k + 1)πτ)) [cos(2k + 1)πτ − cos πz] 2 = − 1

2

d 2 logθ 4 (v, τ ) dv 2

defined for Imτ > 0, and Imz < Imτ 2 . Then φ verifies the following functional equation :

4φ(2z, 2τ) = φ(z, τ ) + φ(z + 1, τ ) = φ(z, τ ) + φ(z, τ + 1).

and θ verifies the following identities (i) 4 d 2 logθ 4 (2v, 2τ)

dv 2 = d 2 logθ 4 (v, τ )

dv 2 + d 2 logθ 4 (v + 1 2 , τ ) dv 2

= d 2 logθ 4 (v, τ)

dv 2 + d 2 logθ 4 (v, τ + 1) dv 2 (ii) d 2 logθ 4 (v, τ )

dv 2 = d 2 logθ 4 (v + 1 2 , τ + 1)

dv 2 .

Indeed, notice that by ([2,Cor 2-5]) θ 4 and θ 3 may be expressed as infinite products

θ 4 (v, τ ) = θ 4 (0, τ ) Y

k≥ 0

"

1 −

sin πv sin(k + 1 2 )πτ

2 #

θ 3 (v, τ ) = θ 3 (0, τ) Y

k≥ 0

"

1 −

sin πv cos(k + 1 2 )πτ

2 # .

We then deduce θ 4 (v, τ ) θ 4 (0, τ )

θ 3 (v, τ) θ 3 (0, τ) = Y

k≥ 0

"

1 −

sin πv sin(k + 1 2 )πτ

2 # "

1 −

sin πv cos(k + 1 2 )πτ

2 # .

However,

"

1 −

sin πv sin(k + 1 2 )πτ

2 # "

1 −

sin πv cos(k + 1 2 )πτ

2 #

= 1−

sin πv sin(k + 1 2 )πτ

2

sin πv cos(k + 1 2 )πτ

2

+

sin πv sin(k + 1 2 )πτ

2

sin πv cos(k + 1 2 )πτ

2

= 1 −

sin πv sin(k + 1 2 )πτ

2

sin πv cos(k + 1 2 )πτ

2 + 4

(sin πv) 2 sin(2k + 1)πτ

2

= 1 −

(sin πv)

(sin(k + 1 2 )πτ)(cos(k + 1 2 )πτ ) 2

+ 4

(sin πv) 2 sin(2k + 1)πτ

2

6

(8)

= 1 −

2(sin πv) sin(2k + 1)πτ

2 +

2(sin πv) 2 sin(2k + 1)πτ

2

=

"

1 −

sin 2πv sin(2k + 1)πτ

2 #

Thus we derive the well known identity ([4,6,7]) θ 4 (2v, 2τ )

θ 4 (0, 2τ) = θ 3 (v, τ ) θ 3 (0, τ )

θ 4 (v, τ) θ 4 (0, τ) = Y

k≥ 0

"

1 −

sin 2πv sin(2k + 1)πτ

2 # .

Corollary 2-2 may be deduced from the Landen transformation by logarithmic differentiation of this identity.

Or equivalently,

θ 4 (2v, 2τ) = θ 3 (v, τ)θ 4 (v, τ)

θ 4 (0, 2τ) = θ 4 (v + 1/2, τ )θ 4 (v, τ) θ 4 (0, 2τ) since (θ 4 (0, 2τ)) 2 = θ 3 (0, τ )θ 4 (0, τ ).

By the same way we may prove

"

1 −

sin πv sin kπτ

2 # "

1 −

sin πv cos kπτ

2 #

=

"

1 −

sin 2πv sin 2kπτ

2 # .

Since

θ 4 (0, 2τ) = p

θ 3 (0, τ )θ 2 (0, τ) = θ

1 (0, τ )θ 2 (0, τ ) 2θ

1 (0, 2τ) we then deduce

θ 1 (2v, 2τ)

θ 4 (0, 2τ) = θ 1 (v, τ ) θ 3 (0, τ )

θ 2 (v, τ)

θ 4 (0, τ ) = (π sin πv) Y

k≥ 0

"

1 −

sin 2πv sin 2kπτ

2 # .

Corollary 2-3 The Weierstrass function ℘(z) = ℘(z, τ) with primitive periods 2 and 2τ satisfies the following identities

(i) ℘(z + τ + 1, τ + 1) + η(τ + 1) = ℘(z + τ, τ) + η(τ), (ii) 4℘(z + τ, τ) + 4η(τ) = ℘( z + τ

2 , τ

2 ) + ℘( z + τ + 2 2 , τ

2 ) + 2η( τ 2 )

= ℘( z + τ 2 , τ

2 ) + ℘( z + τ 2 , τ

2 + 1) + η( τ 2 ) + η( τ

2 + 1).

Replacing z + τ by z one gets

℘(z + 1/2, τ + 1) + η(τ + 1) = ℘(z, τ) + η(τ), 4℘(z, τ ) + 4η(τ) = ℘( z

2 , τ

2 ) + ℘( z + 2 2 , τ

2 ) + 2η( τ 2 )

= ℘( z 2 , τ

2 ) + ℘( z 2 , τ

2 + 1) + η( τ 2 ) + η( τ

2 + 1).

7

(9)

Corollary 2-4 The Weierstrass function ℘(z) = ℘(z, τ) with primitive periods 2 and 2τ verifies the following identity

4℘(2z, 2τ) + 4η(2τ) = ℘(z, 2τ ) − e 1 (τ) + 2η(τ).

Indeed, notice that since [4, p.372]

℘(z + τ, τ) = e 3 + (e 3 − e 1 )(e 3 − e 2 )

℘(z) − e 3

and ℘(z, τ

2 ) = ℘(z, τ) + (e 3 − e 1 )(e 3 − e 2 )

℘(z) − e 3

. Therefore

℘(z, τ

2 ) = ℘(z, τ ) + ℘(z + τ, τ) − e 3 (τ).

By the same way we have

℘(z, τ + 1

2 ) = ℘(z, τ ) + ℘(z + τ + 1, τ) − e 2 (τ),

℘(z, 2τ) = ℘(z, τ ) + ℘(z + 1, τ) − e 1 (τ).

Replacing the last equality in (ii) of Corollary 2-3 we obtain Corollary 2-4. Recall the duplication formula

℘(2z, 2τ) + ℘(z, 2τ ) =

′′

(z, 2τ) 2℘(z, 2τ)

2 .

8

(10)

References

[1] P. Appell, E. Lacour, Fonctions elliptiques et applications Gauthiers- Villard ed., Paris (1922).

[2] A.R. Chouikha Functions related to Jacobi Theta Functions and applica- tions https://hal.archives-ouvertes.fr/hal-03170818. (2021)

[3] A.R. Chouikha Expansions of Theta Functions and Applications ArXiv, math/0112137, http://front.math.ucdavis.edu/0112.5137, (2011).

[4] A.Erdelyi, W.Magnus, F.Oberhettinger, F.Tricomi, Higher transcendental functions Vol II. Based on notes left by H. Bateman. Robert E. Krieger Publish.

Co., Inc., Melbourne, Fla., (1981).

[5] S. Lang, Elliptic functions Addison-Wesley, Springer, 1970.

[6] D.F. Lawden, Elliptic functions and applications, Springer-Verlag, vol 80, 1980.

[7] E.T. Whittaker,G.N. Watson A course of Modern Analysis Cambridge (1963).

9

Références

Documents relatifs

Later in the clinical course, measurement of newly established indicators of erythropoiesis disclosed in 2 patients with persistent anemia inappropriately low

Households’ livelihood and adaptive capacity in peri- urban interfaces : A case study on the city of Khon Kaen, Thailand.. Architecture,

3 Assez logiquement, cette double caractéristique se retrouve également chez la plupart des hommes peuplant la maison d’arrêt étudiée. 111-113) qu’est la surreprésentation

Et si, d’un côté, nous avons chez Carrier des contes plus construits, plus « littéraires », tendant vers la nouvelle (le titre le dit : jolis deuils. L’auteur fait d’une

la RCP n’est pas forcément utilisée comme elle devrait l’être, c'est-à-dire un lieu de coordination et de concertation mais elle peut être utilisée par certains comme un lieu

The change of sound attenuation at the separation line between the two zones, to- gether with the sound attenuation slopes, are equally well predicted by the room-acoustic diffusion

Si certains travaux ont abordé le sujet des dermatoses à l’officine sur un plan théorique, aucun n’a concerné, à notre connaissance, les demandes d’avis

Using the Fo¨rster formulation of FRET and combining the PM3 calculations of the dipole moments of the aromatic portions of the chromophores, docking process via BiGGER software,