Correction TD3

Correction exercice 1.

1. P (X = 0) = 38 , P (X = 1) = 5
8
P (Y = 0 / X = 0) = 27
P (Y = 0 / X = 1) = 37
P (Y = 1 / X = 0) = 57
P (Y = 1 / X = 1) = 47
(X, Y ) (Ω) = {(0, 0) (0, 1) (1, 0) (1, 1)}
3 2 6
P (X = 0, Y = 0) = P (X = 0)P (Y = 0 / X = 0) = 8 × 7 = 56
3 5 15
P (X = 0, Y = 1) = P (X = 0)P (Y = 1 / X = 0) = 8 × 7 = 56
5 3 15
P (X = 1, Y = 0) = P (X = 1)P (Y = 0 / X = 1) = 8 × 7 = 56
5 4 20
P (X = 1, Y = 1) = P (X = 1)P (Y = 1 / X = 1) = 8 × 7 = 56
X \Y 0 1
6 15
0 56 56
15 20
1 56 56
Loi marginale de X
k 0 1
21 35
P (X = k) 56 56
Loi marginale de Y
k 0 1
21 35
P (Y = k) 56 56
2. Loi de X / Y = 0
P (X=k,Y =0)
P (X = k/Y = 0) = P (Y =0)
X /Y =0 0 1
6 15
P (X = k / Y = 0) 21 21
Loi de X / Y = 1
P (X=k,Y =1)
P (X = k / Y = 1) = P (Y =1)
X/Y = 1 0 1
15 20
P (X = k / Y = 1) 35 35
Loi de Y / X = 0
P (X=0,Y =k)
P (Y = k / X = 0) = P (X=0)
Y /X=0 0 1
6 15
P (Y = k / X = 0) 21 21
Loi de Y / X = 1
P (X=1,Y =k)
P (Y = k / X = 1) = P (X=1)
Y /X=1 0 1
15 20
P (Y = k / X = 1) 35 35
3. si x < 0 ou y < 0, FX,Y (x, y) = 0
si 0 ≤ x < 1 et 0 ≤ y < 1, FX,Y (x, y) = P (X = 0, Y = 0) = 56
6

si 0 ≤ x < 1 et y ≥ 1, FX,Y (x, y) = P (X = 0, Y = 0)+P (X = 0, Y = 1) = 6 15

56 + 56 = 21
56

1
 si x ≥ 1 et 0 ≤ y < 1, FX,Y (x, y) = P (X = 0, Y = 0) + P (X = 1, Y = 0) = 6
56 + 15
56 =
21
56
si x ≥ 1 et y ≥ 1, FX,Y (x, y) = 1
si x < 0 ou y < 0

 0
6
si 0 ≤ x < 1 et 0 ≤ y < 1



 56
FX,Y (x, y) = 21
56 si 0 ≤ x < 1 et y ≥ 1
 56 si x ≥ 1 et 0 ≤ y < 1

 21

1 si x ≥ 1 et y ≥ 1


Correction exercice 2.

1. Yn → P (λ) , λ = ln 2
Xn (Ω) = {0, 1}
0
P (Xn = 1) = P (Yn = 0) = e−λ λ0! = e−λ = 21
1
P (Xn = 0) = P (Yn > 0) = 1 − P (Yn = 0) = 2
k 0 1
1 1
P (Xn = k) 2 2
Xn suit la loi de Bernoulli de paramètre 1
2
Xn → B 12

λ
a) P ((Xn = 0) ∩ (Xm = 0)) = P ((Yn > 0) ∩ (Ym > 0)) = 1 − e− 2 = 1 − √1
2
1 1 1
P (Xn = 0) × P (Xm = 0) = 2 × 2 = 4
1− √1
2
=⇒ Xn et Xm ne sont pas indépendantes.
6= 1
4
b) E (Xn Xm ) = k×k 0 ×P ((Xn = k) ∩ (Xm = k 0 )) = 1×1×P ((Xn = 1) ∩ (Xm = 1)) =
PP
k,k0 ∈{0,1}
P ((Yn = 0) ∩ (Ym = 0))
_______________________  _________ _________
= 1−P (Yn = 0) ∩ (Ym = 0) = 1− P (Yn = 0) ∪ (Ym = 0) =
1 − (P ((Yn > 0) ∪ (Ym > 0))
c) cov (Xn , Xm ) = E (Xn Xm ) − E (Xn ) E (Xm )
E (Xn Xm ) = 1−(P ((Yn > 0) ∪ (Ym > 0)) = 1−(P (Yn > 0) + P (Ym > 0)−P ((Yn > 0) ∩ (Ym > 0))
 λ
   
= 1 − (P (Xn = 0) + P (Xm = 0) − 1 − e− 2 ) = 1 − 12 + 12 − 1 − √12 = 1 − √12
1 1 1
E (Xn ) E (Xm ) = 2 × 2 = 4
cov (Xn , Xm ) = 1 − √12 − 14 = 34 − √12

2. E (Sn ) = E ( ni=1 Xi ) = ni=1 E (Xi ) = n2

P P

3 
V (Sn ) = V ( ni=1 Xi ) = ni=1 V (Xi ) + n
+ n2 − n √1
P P PP
cov(Xi , Xj ) = 4 4 − 2
=
i6=j
   
n2 34 − √12 − n 12 − √12 .

Correction exercice 3.

1. X (Ω) = {−2, −1, 1, 2}

Y (Ω) = {1, 4}
Loi conjointe du couple (X, Y ) :

2
 X \Y 1 4
1
−2 0 4
1
−1 4 0
1
1 4 0
1
2 0 4
Loi marginale de Y :
y 1 4
1 1
P (Y = y) 2 2

2. P (Y = y / X = x) = P (X=x,Y =y)
P (X=x)
y 1 4
P (Y = y / X = 1) 1 0
P (X = x / Y = y) = P (X=x,Y =y)
P (Y =y)
x −2 −1 1 2
P (X = x / Y = 1) 0 1
2
1
2 0
3. cov(X, Y ) = E(XY ) − E(X)E(Y )
xyP (X = x,Y = y) = 0
P P
E(XY ) =
x∈X(Ω),y∈X(Ω)
E(X) = 0
E(Y ) = 52
cov(X, Y ) = 0
X et Y ne sont pas indépendantes bien que cov(X, Y ) = 0 car
P (X = −2,Y = 1) = 0 6= P (X = −2) × P (Y = 1) = 18

Correction exercice 4.

1. P (X = i, Y = j) = 2
k(k+1) pour 1 ≤ j ≤ i ≤ k
j=1 k(k+1) = k(k+1) pour 1 ≤ i ≤ k
Pi Pi 2 2i
P (X = i) = j=1 P (X = i, Y = j) =

i=j k(k+1) = k(k+1) pour 1 ≤ j ≤

Pk Pk 2 2(k−j+1)
P (Y = j) = i=j P (X = i, Y = j) = k
2. P (X = i / Y = j) = P (X=i,Y =j)
P (Y =j) = 1
k−j+1 pour 1 ≤ j ≤ i ≤ k
3. P (Y = j / X = i) = P (X=i,Y =j)
P (X=i) = 1
i pour 1 ≤ j ≤ i ≤ k
4. E (X/Y = j) = i × P (X = i / Y = j) =
Pk Pk i j+k
i=j i=j k−j+1 = 2
Y +k
E (X/Y ) = 2
V (X/Y = j) = E X 2 /Y = j − (E (X/Y = j))2

E X 2 /Y = j = ki=j i2 ×P (X = i / Y = j) = ki=j i2 1
−j + k + 2jk + 2j 2 + 2k 2

P P

k−j+1 = 6
(j−k)(j−k−2)
V [X/Y = j] = 12
Pi
P (Y = j/X = i) = ij=1 ji = i+1
P
E (Y /X = i) = j=1 j × 2
2
Y 2 /X


V (Y /X = i) = E = i − (E (Y /X = i))
E Y /X = i = j=1 j × P (Y = j / X = i) = ij=1 ji =

Pi 2 (i+1)(2i+1)
2 2
P
6
i2 −1
V (Y /X = i) = 12
j+k 2(k−j+1)
E (X) = E (E (X/Y )) = kj=1 E (X/Y ) P (Y = j) = kj=1 2k+1
P P
2 k(k+1) = 3
ou E(X) = ki=1 i × P (X = i) = ki=1 k(k+1)
2i2
= 2k+1
P P
3

3
 V (X) = E(X 2 ) − (E(X))2
2i3 k(k+1)
E(X 2 ) = ki=1 i2 × P (X = i) = ki=1
P P
k(k+1) = 2
(k+2)(k−1)
V (X) = 18
E (Y ) = E (E (Y /X)) = ki=1 E (Y /X) P (X = i) = ki=1 i+1 2i k+2
P P
2 k(k+1) = 3

ou E (Y ) = kj=1 j × P (Y = j) = kj=1 j × 2(k−j+1) k+2

P P
k(k+1) = 3
V (Y ) = E(Y 2 ) − (E(Y ))2
2(k−j+1) (k+1)(k+2)
E(Y 2 ) = kj=1 j 2 × P (Y = j) = kj=1 j 2 ×
P P
k(k+1) = 6
(k+2)(k−1)
V (Y ) = 18

5. ρX,Y = cov(X,Y )
σX σY = E(XY
√ )−E(X)E(Y
√
V (X) V (Y )
)

= i,Y = j) = (3k+1)(k+2)
Pk Pi Pk Pi 2ij
E(XY ) = i=1 j=1 ijP (X i=1 j=1 k(k+1) = 12
Remarques :
Pn n(n+1)
i=1 i = 2
Pn 2 n(n+1)(2n+1)
i=1 i = 6
Pn 3  n(n+1) 2
i=1 i = 2
(3k+1)(k+2)
− 2k+1 k+2 1
ρX,Y = q 12
(k+2)(k−1)
q 3 3
(k+2)(k−1)
= 2
18 18

Correction exercice 5.

1.
R +∞ R y
ke−y dxdy = k
R R
R R f (x, y)dxdy = 0 0
R R
R R f (x, y)dxdy = 1 ⇐⇒ k = 1
Pour k = 1 on a f (x, y) ≥ 0
R +∞ −y
e dy = e−x 1R+
R
fX (x) = R f (x, y)dy = 1R+∗ (x)
x
∗ (x)

si
 −x
e x>0
fX (x) =
0 sinon
R R y −y −y
fY (y) = R f (x, y)dx = 1R+
∗ (y)
0 e dx = ye 1R+ (y)
∗

ye−y si

y>0
fY (y) =
0 sinon
2. Les variables X et Y ne sont pas indépendantes car f (x, y) 6= fX (x)fY (y)
3. f (x, y) = e−y 10<x<y
fX (x)fY (y) = e−x ye−y 1R+ ∗ (x)1R∗ (y)
+
f (x, y) 6= fX (x)fY (y) =⇒ X et Y ne sont pas indépendantes.
( ( −y
f (x,y)
si 0 < x < y e
si 0<x<y
a) fX/Y =y (x) = fY (y) = ye−y
0 sinon 0 sinon
si
 1
y 0<x<y
=
0 sinon
b) E(X/Y = y) = R xfX/Y =y (x)dx = 0y x y1 dx = y2
R R

E(X/Y ) = Y2

4
4. U =Y − X et V = X  
U =Y −X X=V x(u, v) = v
a) ⇐⇒ =⇒
V =X Y =U +V y(u, v) = u + v
0 < x < y ⇐⇒ 0 < v < u + v ⇐⇒ u > 0 et v > 0
fX,Y (x(u, v), y(u, v)) |det J| si u > 0 et v > 0

fU,V (u, v) =
0
! 
sinon
∂x(u,v) ∂x(u,v) 
0 1
où J = ∂y(u,v)∂u ∂v
∂y(u,v) = et |det J| = 1
1 1
∂u ∂v
si u > 0 et v > 0
 −(u+v)
e
fU,V (u, v) =
0 sinon
calcul des densités marginales :

densité marginale de U : R +∞
e−(u+v) dv = e−u 1R+
R
fU (u) = f (u, v)dv = 1R+
∗ (u) ∗ (u)
R U,V 0
e−u si u>0
fU (u) =
0 sinon
U → ξ(1) (U suit une loi exponentielle de paramètre 1)
densité marginale de V : R +∞ −(u+v)
du = e−v 1R+
R
fV (v) = R fU,V (u, v)du = 1R+ ∗ (v)
0 e ∗ (v)

si
 −v
e v>0
fV (v) =
0 sinon
V → ξ(1) (V suit une loi exponentielle de paramètre 1)
b) fU,V (u, v) = e−(u+v) 1R+∗ (u)1R+∗ (v)
fU (u)fV (v) = e−u e−v 1R+ ∗ (u)1R∗ (v)
+
fU,V (u, v) = fU (u)fV (v) =⇒ U et V sont indépendantes.
5. U → ξ(1) et V → ξ(1) =⇒
V (X) = V (V ) = 1
V (Y ) = V (U + V ) = V (U ) + V (V ) = 2
cov(X, Y ) = cov(V, U + V ) = cov(V, U ) + cov(V, V ) = 0 + V (V ) = 1
 
V (X) cov(X, Y )
la matrice de covariance de (X, Y ) est ΣX,Y =
cov(X, Y ) V (Y )
D'où  
1 1
ΣX,Y =
1 2

Correction exercice 6.

1. FY (y) = P (Y ≤ y) = P (εX ≤ y) = P (εX ≤ y, ε = −1) + P (εX ≤ y, ε = 1)

= P (−X ≤ y)P (ε = −1) + P (X ≤ y)P (ε = 1) = 21 (P (X ≥ −y) + P (X ≤ y)) =
1
2 (1 − P (X ≤ −y) + P (X ≤ y))
= 21 (1 − FX (−y) + FX (y)) = 21 (FX (y) + FX (y)) = FX (y)
FY (y) = FX (y) =⇒ X et Y suivent la même loi d'où Y → N (0, 1)
2. ρ(X, Y ) = cov(X,Y )
σX σY
σX = σY = 1 car X → N (0, 1) et Y → N (0, 1)
cov(X, Y ) = E(XY ) − E(X)E(Y )
E(X) = E(Y ) = 0

5
 E(XY ) = E(εX 2 )
= E(εX 2 /ε = −1)P (ε = −1) + E(εX 2 /ε = 1)P (ε = 1) =
1 2 2
2 E(−X ) + E(X ) = 0
cov(X, Y ) = 0
a) P (X ∈ [−a, a]) = P (−a ≤ X ≤ a) = FX (a) − FX (−a) = 2FX (a) − 1
b) P (X ∈ [−a, a] , Y ∈ [−b, b]) = P (−a ≤ X ≤ a, −a ≤ Y ≤ a)
= P (−a ≤ X ≤ a, −a ≤ εX ≤ a)
= P (−a ≤ X ≤ a, −a ≤ εX ≤ a, ε = −1) + P (−a ≤ X ≤ a, −a ≤ εX ≤ a, ε = 1)
= P (−a ≤ X ≤ a, −a ≤ X ≤ a)P (ε = −1)+P (−a ≤ X ≤ a, −a ≤ X ≤ a)P (ε = 1)
= 21 (P (−a ≤ X ≤ a) + P (−a ≤ X ≤ a)) = 2FX (a) − 1
P (Y ∈ [−a, a]) = P (−a ≤ Y ≤ a) = P (−a ≤ εX ≤ a) = P (−a ≤ εX ≤ a, ε = −1)+
P (−a ≤ εX ≤ a, ε = 1)
= P (−a ≤ X ≤ a)P (ε = −1) + P (−a ≤ X ≤ a)P (ε = 1) = 2FX (a) − 1
On a :
P (X ∈ [−a, a] , Y ∈ [−b, b]) = 2FX (a) − 1
P (X ∈ [−a, a]) P (Y ∈ [−a, a]) = (2FX (a) − 1)2
X et Y ne sont pas indépendantes car
P (X ∈ [−a, a] , Y ∈ [−b, b]) 6= P (X ∈ [−a, a]) P (Y ∈ [−a, a])
c) P (X = Y ) = P (X = εX)
X = εX dans deux cas : soit (X = 0 et ε ∈ {−1, 1}) ou (X ∈ R et ε = 1)
P (X = εX) = P (X = 0, ε ∈ {−1, 1}) + P (ε = 1, X ∈ R) = P (X =0)P (ε ∈
{-1,1}) + P (ε=1)P (X ∈ R) (car X et ε sont indépendantes)
= 0 × 1 + 12 × 1 = 12
(P (X = 0) = 0 car X est une v.a.r continue)

Correction exercice 7.

1. X et Y suivent laR même

R +∞ loi et sont indépendantes on Ra alors FX (Y ) = P (X ≤ Y ) =
R +∞ −x +∞ −y 1
P (X − Y ≤ 0) = R x f(X,Y ) (x, y)dydx = 0 e x e dydx = 2
D'où P (X > Y ) = 12
2. U = inf(X, Y ) et V = sup(X, Y )
a) FU (u) = P (U ≤ u) = P (inf(X, Y ) ≤ u) = 1 − P (inf(X, Y ) > u) = 1 − P (X >
u, Y > v) = 1 − P (X > u)P (Y > v) car X et Y sont indépendantes
FU (u) = 1 − (1 − P (X ≤ u)) (1 − P (Y ≤ v)) = 1 − (1 − FX (u)) (1 − FY (v)) =
1 − (1 − F 2
(u))−λxoù F est la fonction de répartition
 de la−λx
loi exponentielle.
λe si x>0 1−e si x>0
fX (x) = et FX (x) =
0 sinon 0 sinon
D'où FU (u) = 1 − 1 − 1 + e si u > 0
−λu

2 −2λu
=1−e
C'est la fonction de répartition d'une loi exponentielle de paramètre 2λ d'où U =
inf(X, Y ) → ξ (2λ)
b) P (V ≤ v) = P (U ≤ u, V ≤ v) + P (U > u, V ≤ v) ⇐⇒ (U ≤ u, V ≤ v) = P (V ≤
v) − P (U > u, V ≤ v)
c) FU,V (u, v) = P (U ≤ u, V ≤ v) = P (V ≤ v) − P (U > u, V ≤ v) = P (sup(X, Y ) ≤
v) − P (inf(X, Y ) > u, sup(X, Y ) ≤ v)
= P (X ≤ v, Y ≤ v) − P (X > u, Y > u, X ≤ v, Y ≤ v) = P (X ≤ v)P (Y ≤
v) − P (u < X ≤ v, u < Y ≤ v)
= P (X ≤ v)P (Y ≤ v) − P (u < X ≤ v)P (u < Y ≤ v)
Si u ≤ v on aura FU,V (u, v) = F (v)2 − (F (v) − F (u))2
Si u > v on aura FU,V (u, v) = F (v)2

6
 (F (v))2 si u > v

FU,V (u, v) = 2
(F (v)) − (F (v) − F (u)) 2
si u ≤ v
densité du couple (U, V ) :   
∂2
 ∂2
∂u∂v  (F (v))2
si u > v
fU,V (u, v) = ∂u∂v (FU,V (u, v)) = 
si u ≤ v
2 2 2
 ∂
∂u∂v (F (v)) − (F (v) − F (u))
si u > v

0
fU,V (u, v) =
2f (u)f (v) si u ≤ v
la loi conjointe fU,V (u, v) = 2f (u)f (v)1u≤v 6= f (u)f (v) donc U et V ne sont pas
indépendantes.

Correction exercice 8.

1. a) Vérier que la v.a. U + V est de loi G(a + b, λ)

U → G(a, λ) , V → G(b, λ)
  a
λ
ΦU (t) = λ−it

En eet, ΦU (t) = E eitU =

R +∞ a R +∞ λa a−1 −(λ−it)u
itu f (u)du λ
eitu Γ(a) ua−1 e−λu du =

R
Re U = 0 0 Γ(a) u e du
Soit z = λ−it
λ u
 a−1  a R  a
λa +∞ λa a−1 −λz
R +∞
ΦU (t) = 0
λz
Γ(a) λ−it e−λz λ−it
λ λ
dz = λ−it 0 Γ(a) z e dz = λ
λ−it
eit(U +V ) itU itV itU E(e ) (car U et V sont indépen-
itV




ΦU +V (t) = E =E e e =E e
dantes)
 a  b  a+b
λ λ λ
ΦU +V (t) == ΦU (t)ΦV (t) = λ−it λ−it = λ−it
=⇒ U + V → G(a + b, λ)
b) Vérier que que la v.a. cU est de loi G a, λc

Soit Z = cU  a
 a
λ 1
eitcU


ΦZ (t) = ΦcU (t) = E = ΦU (ct) = λ−ict = 1−i( λc )t
c
=⇒ Z = cU → G(a, λ)
2. X → N (0, 1) et U → G(a, λ).
Les fonctions de densité des v.a. X et U sont :
x2
fX (x) = √12π e− 2 ∀x ∈ R et fU (u) = Γ(a)
λa a−1 −λu
u e ∀x > 0
a) Vérier que la densité de la v.a. X est de loi G 21 , 12
2

Soit Z = X 2
√ √ √
FZ (z) = P (Z ≤ z) = P (X 2 ≤ z) = P (− z ≤ X ≤ z) = 2FX ( z) − 1
√ z
∂
fZ (z) = ∂z 2
FZ (z) = 2√ f ( z) = √1z √12π e− 2 ∀z > 0
z X
1
( 12 ) 2 12 −1 − 21 u
fZ (z) =Γ( 12 )
z e ∀z > 0
Z = X 2 → G 21 , 21


√
b) i. On pose T = √XU et S = U , Déterminer la densité du couple (T, S)
Soit fT,S (t, s) la densité du couple (T, S)
(
X 
T = √ X = TS
√U ⇐⇒
S= U U = S2

7
 x ∈ R =⇒ t = √xu ∈ R
√
s= u>0
fT,S (t, s) = fX,U (ts, s2 ) |det J| pour t ∈ R et s > 0
! 
∂x(t,s) ∂x(t,s) 
s t
où J = ∂u(t,s) ∂u(t,s) =
∂t ∂s
0 2s
∂t ∂s
|det J| = 2s2
X et U deux variables aléatoires réelle indépendantes donc fX,U (ts, s2 ) = fX (ts)fU (s2 )
t 2 s2
a−1 −λs2 2 t2 2
λa 2λa
fT,S (t, s) = fX,U (ts, s2 ) |det J| = √12π e− 2 Γ(a) s2 e 2s = √2πΓ(a) s2a e−(λ+ 2 )s
pour t ∈ R et s > 0
b) ii. Déterminer la densité de la v.a. T .
fT (t) est la densité marginale de la v.a. T
R +∞ 2λa 2a −(λ+ t2 )s2 z=s2 R +∞ 2λa t2
z a e−(λ+ 2 )z 2√ 1
R
fT (t) = R fT,S (t, s)ds = 0 √
2πΓ(a)
s e 2 ds = 0 √
2πΓ(a) z
dz =
R +∞ λa 1 t2

0
√
2πΓ(a)
z a− 2 e−(λ+ 2 )z dz
λa
R +∞ (a+ 1 )−1 −(λ+ t2 )z a Γ(a+ 12 ) a Γ(a+ 12 )
= √2πΓ(a) z 2 e 2 dz = √ λ 1 = √ λ a+ 1 =
0 2πΓ(a)  a+ 2πΓ(a) 1 2 2 2 2
λ+ t2 λa+ 2 1+ 2λ
t

Γ(a+ 1 ) 1
√ √ 2 a+ 1
λ 2πΓ(a)  t2 2
1+ 2λ
Γ(a+ 1 ) 1
=⇒ fT (t) = √ √ 2
λ 2πΓ(a)  a+ 1 ∀t ∈ R.
t2 2
1+ 2λ

3. déduire sans calcul à partir de 1.b) et 2.b) la densité de la v.a. X12 + ... + Xn2
X → N (0, 1) alors X 2 → G 12 , 12

U → G(a, λ) , V → G(b, λ) alors U + V → G(a + b, λ) et cU → G(a, λc )

Donc Xi2 → G( 21 , 12 ) et X12 + ... + Xn2 → G( n2 , 12 )
4. déduire sans calcul la densité de la v.a. r 2 Y 2 (loi de Student)
X1 +...+Xn
n

U → G(a, λ) alors cU → G(a, λc )

X12 + ... + Xn2 → G( n2 , 21 ) =⇒ n1 X12 + ... + Xn2 → G( n2 , n2 )

Γ(a+ 1 )
La densité de T = X
√
U
où X → N (0, 1) et U → G(a, λ) est fT (t) = √ √ 2
λ 2πΓ(a) 
1
a+ 1
t2 2
1+ 2λ
∀t ∈ R
X12 +...+Xn2
La densité de la variable r Y
2 +...+X 2
où Y → N (0, 1) et n → G( n2 , n2 ) est la
X1 n
n

densité de T = X
√
U
avec a = n
2 et λ = n
2
Γ( + ) n 1
1 Γ( n+1 ) 1
fr Y (t) = √ n √2 2 n n+1 = √
nΓ( n
2
)Γ( 1 
)  n+1 ∀t ∈ R
2 +...+X 2 2πΓ( 2 )  2 2 2 t2 2
X1 n 2 1+ tn 2 2
1+ n
n

Correction exercice 9.

1. X → G(λ)
P (X = x) = (1 − λ)x−1 λ pour x ∈ N∗
∀x ∈ [n, n + 1[ , FX (x) = P (X < x) = nx=1 (1 − λ)x−1 λ = 1 − (1 − λ)n
P

8
 sλ
a) GX (s) = E(sX ) = 1−s(1−λ)
00
E(X(X − 1)) = GX (1)
G0X (s) = λ
(1−s(1−λ))2
, G0X (1) = 1
λ

, GX (1) =
00 2λ(1−λ) 00 2(1−λ)
GX (s) = (1−s(1−λ))3 λ2
2(1−λ)
=⇒ E(X(X − 1)) = λ2
 
b) E(Y ) = E(e−X ) = E 1 X 1
< ∞ (nie)=⇒ E(Y ) existe



e = GX e
1 λ


E(Y ) = GX e = e−(1−λ)

2. Z = (−1)X
Z (Ω) = {−1, 1}
x 1 2 3 4 5 6 7 8 ...
z −1 1 −1 1 −1 1 −1 1 ...
P (Z = −1) = ∞ P (X = 2k + 1) = ∞ (1 − λ)2k+1−1 λ = λ ∞ 2k
P P P
k=0 k=0 k=0 (1 − λ) =
P∞   k
λ k=0 (1 − λ)2 = 2−λ 1

P∞ P∞ 2k−1 λ P∞ 2k
P (Z = 1) = k=1 P (X = 2k) = k=0 (1 − λ) λ = 1−λ k=0 (1 − λ) =
  k
λ ∞ 2
= 1−λ
P
1−λ k=0 (1 − λ) 2−λ
z −1 1
1 1−λ
P (Z = z) 2−λ 2−λ
−λ
E(Z) = 2−λ
Correction exercice 11.
X → G(p) et Y → G(p)
1) GX (z) = E(z X ) = ∞ k
P∞ k k−1 p = p P∞
− p))k =
P
z=1 z P(Z = k) = z=1 z (1 − p) 1−p z=1 (z(1
p P∞ k

p 1 pz
1−p z=0 (z(1 − p)) − 1 = 1−p 1−z(1−p) − 1 = 1−z(1−p)
De même GY (z) = E(z Y ) = 1−z(1−p) .
pz

GS (z) = E (z s ) = E z X+Y = E(z X z Y ) = E(z X )E(z Y ) = p2 z 2

.


(1−z(1−p))2
2) S(Ω) = N∗ \ {1}
P (S = n) = P (X + Y = n) = n−1 P (X = i et Y = n − i)
P
i=1
= i=1 P (X = i )(Y = n − i) = i=1 (1 − p)i−1 (1 − p)n−i−1 p = (n − 1)p2 (1 − p)n−2
Pn−1 Pn−1

3) ∀X ≥ 1 et n > k
P (X = k et S = n) P (X = k et X + Y = n) P (X = k et Y = n − k)
P (X = k / S = n) = = =
P (S = n) P (S = n) P (S = n)
P (X = k )P (Y = n − k) k−1
(1 − p) p(1 − p) n−k−1 p 1
= = n−2 =
P (S = n) 2
(n − 1)p (1 − p) n−1
1 Pn−1 1 (n − 1) n n
E(X / S = n) = k=1 kP (X = k / S = n) =
Pn−1
k=1 k = =
n−1 n−1 2 2
S
⇒ E(X / S) =
2  
S
4) E(E(X/S)) = E = 12 ∞ 1 P∞ n−2 2
P
n=2 nP (S = n) = 2 n=2 n(n − 1) (1 − p) p =
2 !
1 2 ∂
2 P
∞ n 1 2 ∂
2 (1 − p)2 1
2 p 2
( n=2 (1 − p) ) = 2 p 2
= = E(X)
∂p ∂p p p

9
 D'où E(E(X/S)) = E(X)

Correction exercice 12.

1. 2
a) (a − bρ) + 1 − ρ2 b2 = a2 − 2ρab + b2


1
− (x2 +2ρxy+y2 )
b) fY (y) = R fX,Y (x, y)dx = R √1 2 e 2(1−ρ2 )
R R
dx
2π 1−ρ
− 1
((x−ρy)2 +(1−ρ2 )y2 )
√1 2(1−ρ2 )
R
= R 2π e dx
1−ρ2
(x−ρy)2
y2 −
= e− √1 2(1−ρ2 )
R
2
R 2π 2
e dx
1−ρ
 2
2 − 21 √x−ρy t= √x−ρy2
y2 2 2
− y2 1−ρ 1 − t2 y
√1 e− √1 e− 2
R R
1−ρ2
=e R 2π 1−ρ2 e dx = 2
R 2π e dt = 2π
2
De même fX (x) =
x
√1 e− 2
2π
X et Y → N (0, 1)
c) fX,Y (x, y) 6= fX (x)fY (y) =⇒ X et Y ne sont pas indépendantes
2. Pour ρ = 0 on aura fX,Y (x, y) = exp − 12 x2 + y 2 ∀ (x, y) ∈ R 2 1


2π

a) ΦX+Y (t) = E eit(X+Y ) = E eitX eitY = E eitX )E(eitY car pour ρ = 0,

fX,Y (x, y) = fX (x)fY (y) et donc X et Y sont indépendantes.

t2 t2 2
ΦX+Y (t) = ΦX (t)ΦY (t) = e− 2 e− 2 = e−t
Démonstration :
x 2 x 2 1 2 t 2
z=x−it
E(eitX ) √1 eitx e− 2 √1 e− 2 +itx dx √1 e− 2 (x−it) − 2
R R R
ΦX (t) = = R 2π
dx = R 2π
= R 2π
dx =
t2 R 1 2 2
− t2
e− 2 R √1 e− 2 z dz
2π
=e
√
b) X + Y → N (0, 2)
Démonstration :
X suit la loi normale de paramètres 0 et 1 et la fonction caractéristique de X est
t2
ΦX (t) = e− 2
Soit U = σX + m , U → N (m, σ)
La fonction caractéristique de U est
ΦU (t) = ΦσX+m (t) = E(eit(σX+m) ) = E(eitσX+itm ) = eitm E(eitσX ) = eitm ΦX (tσ) =
σ 2 t2
eitm e− 2
√
ΦX+Y (t) = e−t =⇒ X + Y suit la loi normale de paramètres m = 0 et σ =
2
2
√
X + Y → N (0, 2)

10