Cours Mi
Cours Mi
IM 26 ur
Prof H. MONAIM
NA 5 ie
O - 2 én
.M F ng
-
H P rI
I/U ou
FS s p
h
at
M
IM 26 ur
NA 5 ie
O - 2 én
1 Rappel sur les espaces vectoriels 4
1.1 Normes standards de Rn . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
.M F ng
1.2 Convergence et continuité . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3 Calcul différentiel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
-
1.4 Application differentiable . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
H P rI
2 Séries de Fourier 21
2.1 Séries trigonométriques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
FS s p
3
CHAPTER 1
IM 26 ur
NA 5 ie
O - 2 én
Définition 1
.M F ng
Soient K (R ou C) un corps. Un K-esapce vactoriel est un ensemble E, dont les éléments
-
sont appelés vecteurs, muni de deux lois:
H P rI
Exemple 1
L’espace R2 = {(x, y), x, y ∈ R} est un espace vectoriel sur R.
En effet, pour tout (x, y); (x′ , y′ ); (x′′ , y′′ ) ∈ R2 et α, β ∈ R, on a
1-
(x, y) + (x′ , y′ ) = (x + x′ , y + y′ )
= (x′ + x, y′ + y)
= (x′ , y′ ) + (x, y).
4
Mathématiques pour l’ingénieur 5/ 25
2-
(x, y) + (x , y ) + (x′′ , y′′ ) = (x + x′ , y + y′ ) + (x′′ , y′′ )
′ ′
= (x + x′ ) + x′′ , (y + y′ ) + y′′
= x + (x′ + x′′ ), y + (y′ + y′′ )
= (x, y) + x′ + x′′ , y′ + y′′
= (x, y) + (x′ , y′ ) + (x′′ , y′′ ) .
IM 26 ur
3- Cherchons (x′ , y′ ) ∈ R2 tel que (x, y) + (x′ , y′ ) = (x, y)
NA 5 ie
(x, y) + (x′ , y′ ) = (x + x′ , y + y′ ) = (x, y).
O - 2 én
Donc
(
x + x′ = x
.M F ng
y + y′ = y.
-
H P rI
Donc
(
x + x′ = 0
h
y + y′ = 0.
at
5-
′ ′ ′ ′
α (x, y) + (x , y ) = α x + x , y + y
= α(x + x′ ), α(y + y′ )
′ ′
= αx + αx , αy + αy )
= (αx, αy) + (αx′ , αy′ )
= α(x, y) + α(x′ , y′ ).
6-
(α + β)(x, y) = (α + β)x, (α + β)y
= (αx + βx, αy + βy)
= (αx, αy) + (βx, βy)
= α(x, y) + β(x, y).
7-
(αβ)(x, y) = (αβ)x, (αβ)y
= α(βx), α(βy)
= α βx, βy
= α β(x, y) .
8-
IM 26 ur
1(x, y) = 1x, 1y
NA 5 ie
= (x, y).
O - 2 én
Remarque 1
.M F ng
Les espaces Rn , pour n ∈ N∗ , sont des R-espaces vectoriels.
-
H P rI
Exemple 2
a11 a12
I/U ou
a11 a12 b11 b12 a11 + b11 a12 + b12
+ =
at
2-
a a12
b11 b12
c11 c12
a11 + b11 a12 + b12
c11 c12
11
+ + = +
a21 a22 b21 b22 c21 c22 a21 + b21 a22 + b22 c21 c22
(a11 + b11 ) + c11 (a12 + b12 ) + c12
=
(a21 + b21 ) + c21 (a22 + b22 ) + c22
a11 + (b11 + c11 ) a12 + (b12 + c12 )
=
a21 + (b21 + c21 ) a22 + (b22 + c22 )
a11 a12 b11 + c11 b12 + c12
= +
a21 a22 b21 + c21 b22 + c22
a11 a12 b11 b12 c11 c12
= + + .
a21 a22 b21 b22 c21 c22
′
a11 a′12
′
a11 a′12
a11 a12 a11 a12
3- Cherchons ′ ∈ M2 (R) tel que + ′ =
a21 a′22 a21 a22 a21 a′22 a21 a22
′
a11 a′12 a11 + a′11 a12 + a′12
a11 a12 a11 a12
+ ′ = = .
a21 a22 a21 a′22 a21 + a′21 a22 + a′22 a21 a22
Donc
IM 26 ur
a22 + a′22 = a22
NA 5 ie
′
a11 a′12
0 0
Ce qui donne ′ = .
a21 a′22 0 0
O - 2 én
′
a11 a′12
a11 a12
4- Cherchons ′ ∈ M2 (R) en fonction de tel que
a21 a′22 a21 a22
.M F ng
-
′
a11 a′12
a11 a12 0 0
+ ′ =
H P rI
a21 a22 a21 a′22 0 0
On a
I/U ou
′
a11 a′12 a11 + a′11 a12 + a′12
a11 a12 0 0
+ ′ = = .
a21 a22 a21 a′22 a21 + a′21 a22 + a′22 0 0
FS s p
Donc
a11 + a′11 = 0
h
a + a′ = 0
at
12 12
′
a 21 + a 21 = 0
a22 + a′22 = 0
M
′
a11 a′12
−a11 −a12 a11 a12
Ce qui donne ′ = =− .
a21 a′22 −a21 −a22 a21 a22
5-
a a12
b11 b12
a11 + b11 a12 + b12
11
α + =α
a21 a22 b21 b22 a21 + b21 a22 + b22
α(a11 + b11 ) α(a12 + b12 )
=
α(a21 + b21 ) α(a22 + b22 )
αa11 + αb11 αa12 + αb12
=
αa21 + αb21 αa22 + αb22
αa11 αa12 αb11 αb12
= +
αa21 αa22 αb21 αb22
a11 a12 b11 b12
=α +α .
a21 a22 b21 b22
6-
a11 a12 (α + β)a11 (α + β)a12
(α + β) =
a21 a22 (α + β)a21 (α + β)a22
αa11 + βa11 αa12 + βa12
=
αa21 + βa21 αa22 + βa22
αa11 αa12 βa11 βa12
= +
αa21 αa22 βa21 βa22
a11 a12 a11 a12
=α +β .
IM 26 ur
a21 a22 a21 a22
NA 5 ie
7-
O - 2 én
a11 a12 (αβ)a11 (αβ)a12
(αβ) =
a21 a22 (αβ)a21 (αβ)a22
.M F ng
α(βa11 ) α(βa12 )
=
α(βa21 ) α(βa22 )
-
βa11 βa12
H P rI
=α
βa21 βa22
a
βa12
11
I/U ou
=α β .
βa21 βa22
FS s p
8-
a11 a12 1a11 1a12
=
h
1
a21 a22 1a21 1a22
at
a a12
= 11 .
a21 a22
M
Définition 2 (Norme)
∥.∥ : E −→ R+ ,
telle que
IM 26 ur
Preuve:
NA 5 ie
Il suffit de vérifier les axiomes d’une norme.
O - 2 én
i- Soit u = (u1 , u2 , ..., un ) ∈ Rn telle que ∥u∥1 = 0.
Ce qui donne
n
.M F ng
∑ |uk | = 0
k=1
-
Par suite
H P rI
u = 0.
Inversement, si u = (u1 , u2 , ..., un ) = 0.
FS s p
On obtient immédiatement
∥u∥1 = 0.
ii- Soient u = (u1 , u2 , ..., un ) ∈ Rn et λ ∈ R, on a
h
n
at
∥λu∥1 = ∑ |λuk |
k=1
M
n
= |λ| ∑ |uk |
k=1
= |λ|∥u∥1 .
D’où l’homogénéité.
Donc
n n n
∑ |uk + vk | ≤ ∑ |uk | + ∑ |vk |.
k=1 k=1 k=1
(i.e.)
∥u + v∥1 ≤ ∥u∥1 + ∥u∥1 .
IM 26 ur
Preuve: Soient u = (u1 , u2 , ..., un ) ∈ Rn , v = (v1 , v2 , ..., vn ) ∈ Rn et t ∈ R. On a, d’une part
NA 5 ie
∥tu + v∥22 ≥ 0.
O - 2 én
D’autre part
.M F ng
n
∥tu + v∥22 = ∑ |tuk vk |2
-
k=1
n n n
H P rI
= t 2 ∑ |uk |2 + 2t ∑ uk vk + ∑ |vk |2
k=1 k=1 k=1
2
= at + bt + c
FS s p
n n n
avec a = ∑ |uk |2, b = 2 ∑ uk vk et c = ∑ |vk |2.
h
!2
n n n
b2 − 4ac = 4 − 4 ∑ |uk |2 ∑ |vk |2 ≤ 0.
M
∑ uk vk
k=1 k=1 k=1
D’où !2
n n n
∑ uk vk ≤ ∑ |uk |2 ∑ |vk |2.
k=1 k=1 k=1
2
Proposition 2
Preuve:
Il suffit de vérifier les axiomes d’une norme.
IM 26 ur
∀k ∈ {1, ..., n} |uk | = 0.
D’où
NA 5 ie
u = 0.
O - 2 én
Inversement, si u = (u1 , u2 , ..., un ) = 0.
On obtient immédiatement
∥u∥2 = 0.
.M F ng
n
ii- Soient u = (u1 , u2 , ..., un ) ∈ R et λ ∈ R, on a
-
s
H P rI
n
∥λu∥2 = ∑ |λuk |2
k=1
I/U ou
s
n
= |λ| ∑ |uk |2
k=1
FS s p
= |λ|∥u∥2 .
D’où l’homogénéité.
h
n
∥u + v∥22 = ∑ |uk + vk |2
M
k=1
n n n
= ∑ |uk |2 + 2 ∑ uk vk + ∑ |vk |2
k=1 k=1 k=1
(i.e.)
!1 !1
n 2 n 2
Lemme 2
Soient f et g deux fonction réelles bornées sur un intervalle [a, b], on a alors
IM 26 ur
Proposition 3
NA 5 ie
∥u∥∞ = sup |uk | avec u = (u1 , u2 , ..., un )
O - 2 én
k∈{1,..,n}
sup |uk | = 0
k∈{1,..,n}
Par suite
h
D’où
u = 0.
M
c.à.d
∥u∥∞ = 0.
ii- Soient u = (u1 , u2 , ..., un ) ∈ Rn et λ ∈ R, on a
D’où l’homogénéité.
D’après le lemme 2, on a
IM 26 ur
sup |uk | + |vk | ≤ sup |uk | + sup |vk |.
k∈{1,..,n} k∈{1,..,n} k∈{1,..,n}
NA 5 ie
(i.e.)
O - 2 én
∥u + v∥∞ ≤ ∥u∥∞ + ∥v∥∞ .
.M F ng
2
-
H P rI
Définition 3
FS s p
Soient (E, ∥.∥E ) un espace vectoriel normés et (Un )n∈N une suite de E.
On dit que (Un )n∈N converge vers a ∈ E si
h
On ecrit alors
M
lim Un = a.
n→+∞
Définition 4
Soient (E, ∥.∥E ) et (F, ∥.∥F ) deux espaces vectoriels normés et f : E 7−→ F une fonction
et a ∈ E.
On dit que f est continue en a si
On ecrit alors
lim f (x) = f (a).
x→a
f (x) − f (a)
lim
n→a x−a
existe et finie.
IM 26 ur
NA 5 ie
O - 2 én
Propriétés:
1• (xn )′ = nxn−1 .
.M F ng
-
1
H P rI
2• (ln x)′ = .
x
I/U ou
1
3• (arctan x)′ = .
1 + x2
FS s p
′
1 1
4• = − 2.
x x
h
at
Propriétés 1
2• (α f (x))′ = α f ′ (x).
Soient (E, ∥.∥E ), (F, ∥.∥F ) sont deux espaces vectoriel normés et L :−→ F une applica-
tion linéaire. f est continue si et seulement si
∃M ≥ 0, ∀x ∈ E ∥L (x)∥E ≤ M∥x∥F .
IM 26 ur
Preuve:
NA 5 ie
Notation:
O - 2 én
On note L (E, F) l’espace des applications linéaires continues.
.M F ng
Définition 6
-
H P rI
Soient (E, ∥.∥E ), (F, ∥.∥F ) deux espaces de Banach, A une partie ouverte de E, a ∈ A
et f : E −→ F une application. On dit que f est differentiable en a s’il existent une
application linéaire continue La ∈ L (E, F) et une fonction εa : A −→ F avec lim εa (h) =
I/U ou
h→0
0 telles que
f (a + h) = f (a) + La (h) + ∥h∥E εa (h).
FS s p
Remarques:
h
Proposition 5
Soient (E, ∥.∥E ), (F, ∥.∥F ) deux espaces de Banach, A une partie ouverte de E, a ∈ A et
f : E −→ F. Si f est differentiable en a alors La est unique.
Preuve:
Soient La et La deux applications linéaires telles que
f (a + h) = f (a) + La (h) + ∥h∥E ε1,a (h).
et
f (a + h) = f (a) + La (h) + ∥h∥E ε2,a (h).
La difference des deux égalités mène à
(La − La )(h) = ∥h∥E εa (h).
Définition 7
On dira que est differentiable sur A si elle est differentiable en tout point de a ∈ A. Ce qui
conduit à la définition de la differentielle de f comme suit
IM 26 ur
Da f : A −−−−−−−−→ L (E, F)
NA 5 ie
a 7−−−−−−−−→ Da f
1.4.1
O - 2 én
Le gradient et la matrice hessienne d’une fonction
.M F ng
On considère une fonction
-
H P rI
f : Rn 7−→ R
I/U ou
FS s p
Définition 8
∂f ∂f ∂f
∇ f (X) = grad f (X) = (X), (X), ..., (X)
at
Exemples:
f : R2 −→ R
(x, y) 7−→ ln(xy) + xy2 + x2 y
Donc
Définition 9
∂2 f ∂2 f ∂2 f
∂x1 ∂x1 (X) (X) ... (X)
2 ∂x2 ∂x1 ∂xn ∂x1
∂ f 2
∂ f 2
∂ f
(X) (X) ... (X)
H f (X) = ∂x1 ∂x 2 ∂x 2 ∂x 2 ∂x n ∂x 2
.
.. .
.. . .. .
..
IM 26 ur
∂2 f 2
∂ f 2
∂ f
(X) (X) ... (X)
∂x1 ∂xn ∂x2 ∂xn ∂xn ∂xn
NA 5 ie
O - 2 én
Exemples:
f : R2 −→ R
.M F ng
(x, y) 7−→ ln(xy) + xy2 + x2 y
-
H P rI
Donc
∂f ∂f
= (x, y), (x, y)
∂x ∂y
FS s p
1 2 1 2
= + y + 2xy, + 2xy + x .
x y
h
et
∂2 f ∂2 f
at
1
∂x∂x (X) (X) −
x2 + 2y 2y + 2x
H f (X) = ∂y∂x = .
2 2 1
M
∂ f ∂ f 2y + 2x − + 2x
(X) (X) y2
∂x∂y ∂y∂y
Lemme 3
∂2 f ∂2 f
(X) = (X).
∂x∂y ∂y∂x
1
(ex )′ = ex , (sin x)′ = cos x, (ln x)′ = ,
x
′
1 1 1
= − 2 , (arctan x)′ = ·
IM 26 ur
x x 1 + x2
NA 5 ie
Z b Z b
O - 2 én
′
f (t)g(t)dt = [ f (t)g(t)]ba − f (t)g′ (t)dt.
a a
.M F ng
4- Changement de variables. Pour a, b ∈ R, g : [a, b] 7−→ R et f : [g(a), g(b)] 7−→ R
-
Z b Z g(b)
H P rI
′
f ◦ g(t)g (t)dt = f (t)dt.
a g(a)
I/U ou
Z c Z b Z c
f (t)dt = f (t)dt + f (t)dt.
a a b
M
Proposition 6
Soient a, b ∈ R et f une fonction intégrable sur [a, b]. Si f est positive sur [a, b] alors
Z b
f (t)dt ≥ 0.
a
Corollaire 1
Corollaire 2
Alors
Z b Z b
f (t)dt ≤ g(t)dt.
a a
IM 26 ur
Proposition 7
NA 5 ie
Soient a, b ∈ R et f une fonction intégrable sur [a, b]. Alors
O - 2 én
Z b Z b
f (t)dt ≤ | f (t)|dt.
a a
.M F ng
-
H P rI
Proposition 8
Soient a, b ∈ R et f une fonction intégrable sur [a, b]. Si f est positive et continue sur
I/U ou
[a, b] alors
Z b
FS s p
f (t)dt = 0 =⇒ f = 0.
a
h
Proposition 9 (Cauchy-Shwartz)
at
Z b
2 Z b Z b
f (t)g(t)dt ≤ f 2 (t)dt g2 (t)dt.
a a a
Preuve: Soient a, b ∈ R et f , g deux fonctions intégrables sur [a, b]. On a, pour tout r ∈ R,
Z b Z b Z b Z b
2 2 2
[(r f (t) + g(t)] dt = r f (t)dt + 2r f (t)g(t)dt + g2 (t)dt.
a a a a
Exercices:
Calculer les integrales suivantes
Z 2 π
3 Z
3
1- √ dx. 4- (sin(6x) + 1 − cos(3x))dx.
0 5x + 1 0
Z √π Z 2p
ln(x)
2- x sin(x2 )dx. 5- dx.
0 1 x
Z e
Z 1
x
6- arctan x + x2 ln xdx.
3- xe dx. 1
0
IM 26 ur
NA 5 ie
O - 2 én
.M F ng
-
H P rI
I/U ou
FS s p
h
at
M
IM 26 ur
SÉRIES DE FOURIER
NA 5 ie
O - 2 én
.M F ng
2.1 Séries trigonométriques
-
H P rI
Historique:
On fait habituellement remonter l’origine des séries de Fourier à l’ouvrage de Joseph Fourier
Théorie analytique de la chaleur paru en 1822 où est exposée l’utilisation formelle de sommes
h
À cette époque, les notions d’intégration ou de continuité ainsi que les problèmes de conver-
gence des séries de Fourier ne sont pas clairement maı̂trisés et les séries de Fourier vont se
trouver au cœur du développement de l’analyse au 19ème siècle.
M
En analyse, les séries de Fourier sont un outil essentiel pour l’étude des fonctions périodiques.
Elles sont largement utilisées dans l’étude des signaux.
Elles interviennent par exemple en filtrage numérique, en synthèse sonore ou en traitement
d’images.
La décomposition d’une fonction f en série de Fourier consiste à trouver une série trigonométrique
dont f est la somme. Il résulte des propriétés des séries trigonométriques que seule une fonc-
tion 2π-rpériodique peut être la somme d’une série trigonométrique.
21
22/ 25 GCIVIL3-GRESE3-IAAR3
Soit f une application définie sur R, 2π-périodique et Riemann-intégrable sur [0, 2π].
1 2π
Z
a0 = f (x)dx
2π 0
1 2π
Z
an = f (x) cos(nx)dx n ∈ N∗
π 0
1 2π
Z
IM 26 ur
bn = f (x) sin(nx)dx n∈N
π 0
NA 5 ie
2- On appelle série de Fourier de f la série trigonométrique de terme général
O - 2 én
fn : x 7−→ an cos(nx) + bn sin(nx)
.M F ng
-
H P rI
Exemple:
Considérons l’application 2π-périodique f qui en restriction à l’intervalle [0, 2π[ est x 7−→
I/U ou
π − x.
On a a0 = 0 et en effectuant le changement de variable y = π − x, on obtient pour tout
n ∈ N∗ :
FS s p
Z 2π
1
an = (π − x) cos(nx)dx
h
π 0
1
at
Z π
=− y cos(ny − nπ)dx
π −π
M
1
Z π
= (−1)n+1 y cos(ny)dx
π −π
=0 Car y 7−→ y cos(ny) est impaire.
et
Z 2π
1
bn = (π − x) sin(nx)dx
π 0
cos(nx) 2π 1 2π
1
Z
= − (π − x) − cos(nx)dx
π n 0 nπ 0
2
= .
n
2
La série de Fourier de f est donc ∑ n sin(nx)
n≥1
Proposition 10
Soit f une application 2π−périodique définie sur R et intégrable sur [O, 2π]. Pour tout
réel α ∈ R, on a
1 α+2π
Z
a0 = f (x)dx
2π α
1 α+2π
Z
an = f (x) cos(nx)dx n ∈ N∗
π α
1 α+2π
Z
bn = f (x) sin(nx)dx n∈N
IM 26 ur
π α
NA 5 ie
Preuve:
Soient f une application 2π−périodique définie sur R et intégrable sur [O, 2π] et α ∈ R un
O - 2 én
réel, on a
Z α+2π Z 0 Z 2π Z α+2π
.M F ng
f (x)dx = f (x)dx + f (x)dx + f (x)dx
α α 0 2π
-
et grâce au changement de variables y = x − 2π, on obtient
H P rI
Z α+2π Z α
f (x)dx = f (y + 2π)dy.
2π 0
I/U ou
2π 0
2
Corollaire 3
h
at
Soit f une application 2π−périodique définie sur R et intégrable sur [0, 2π].
1 2
Z π Z π
a0 = f (x)dx et an = f (x) cos(nx)dx n ∈ N∗
π 0 π 0
2
Z π
bn = f (x) sin(nx)dx n ∈ N∗
π 0
Preuve:
1- Soient f une application paire 2π−périodique définie sur R et intégrable sur [0, 2π]. Pour
α ∈ −π, on a
Z 0
1 1 1
Z π Z π
bn = f (x) sin(nx)dx = f (x) sin(nx)dx + f (x) sin(nx)dx.
π −π π −π π 0
et grâce au changement de variables y = −x, on obtient
Z 0
1 1
Z π
f (x) sin(nx)dx = − f (x) sin(nx)dx.
π −π π 0
D’où bn = 0.
Comme f est paire, on a:
Z 0
1 1
Z π
f (x) cos(nx)dx = f (x) cos(nx)dx.
π −π π 0
On en tire
IM 26 ur
1
Z π
a0 = f (x)dx
π 0
NA 5 ie
et
O - 2 én
2
Z π
an = f (x) cos(nx)dx
π 0
.M F ng
2- Dans le cas où f est impaire, les relations données s’obtiennent de façon similaire.
-
2
H P rI
Exercice:
I/U ou
(
−1 si x ∈ [−π, 0]
2- f (x) =
1 si x ∈ [0, π].
h
at
x
3- f (x) = sin x ∈ [0, 2π].
2
M
Si f est une application 2π-périodique, dérivable par morceaux sur [0, 2π] alors pour
tout x0 ∈ R, la série de Fourier de f converge. Sa somme en x0 est:
IM 26 ur
Exemple:
NA 5 ie
Siot f est la fonction définie par
(
O - 2 én
1 si 0≤x≤π
f (x) =
0 si −π ≤ x < 0
.M F ng
Alors
-
lim f (x) = 0 et lim f (x) = 1.
H P rI
x0− x0+
Et
1
∑ an cos(nπ) + bn sin(nπ) = ∑(−1)nan = 2 .
n n
h
car
at