Uniform and Normal Distribution

1. Uniform or Rectangular Distribution

Let α and β be two real numbers such that −∞ < α < β < ∞. A continuous random
variable X is said to have a uniform (or rectangular) distribution over the interval (α, β)
(written as X ∼ U (α, β)) if probability density function of X is given by
(
1
β−α
,if α < x < β
fX (x) =
0, otherwise

Now, the r-th moment of X ∼ U (α, β) is

Z∞
r
E(X ) = xr fX (x)dx
−∞

Zβ
xr
= dx
β−α
α
β r+1 − αr+1
=
(r + 1)(β − α)
β r + β r−1 α + · · · + βαr−1 + αr
=
r+1

Hence

α+β
E(X) = ;
2
β 2 + βα + α2
E(X 2 ) = ;
3
(β − α)2
V ar(X) = E(X 2 ) − (E(X))2 = .
12

The m.g.f. of X ∼ U (α, β) is

MX (t) = E(etX )
Z∞
= etx fX (x)dx
−∞

Zβ
etx
= dx
β−α
α
(
etβ −etα
(β−α)t
if t 6= 0
,
= .
1, if t = 0
1
The d.f. of X ∼ U (α, β) is

Zx
FX (x) = fX (t)dt
−∞

0, if x < α

x−α
= β−α , if α ≤ x < β

1, if x ≥ β


X−α
Remark 1. Let X ∼ U (α, β) and Y = β−α
. Then the d.f. of Y is

FY (y) = P (Y ≤ y)
= P (X ≤ α + (β − α)y)

0, if α + (β − α)y < α

= α+(β−α)y−α
β−α
, if α ≤ α + (β − α)y < β

1, if α + (β − α)y ≥ β


0, if y < 0

= y, if 0 ≤ y < 1

1, if y ≥ 1

Clearly, FY is not differentiable at 0 and 1. Hence, the p.d.f. of Y is

(
1, if 0 < y < 1
fY (y) =
0, otherwise

Therefore, Y ∼ U (0, 1).

Example 2. Let a > 0 be a real constant. A point X is chosen at random on the interval
(0, a) (i.e., X ∼ U (0, a)).

(1) If Y denotes the area of equilateral triangle having sides of length X, find the
mean and variance of Y .
(2) If the point X divides the interval (0, a) into subintervals I1 = (0, X) and I2 =
[X, a), find the probability that the larger of these two subintervals is at least the
double of the size of the smaller subinterval.

Solution:
√
3 2
(1) We have Y = 4
X . Then
√ √
3 2 3 2
E(Y ) = E(X ) = a;
4 12
3 3
E(Y 2 ) = E(X 4 ) = a4 ;
16 80
a4
V ar(Y ) = E(Y 2 ) − (E(Y ))2 = .
80
2
 (2) The required probability is
p = P ({max(X, a − X) ≥ 2 min(X, a − X)})
a a
= P ({a − X ≥ 2X, X ≤ }) + P ({X ≥ 2(a − X), X > })
2 2
a 2a
= P (X ≤ }) + P ({X ≥ })
3 3
a 2a
= FX ( ) + 1 − FX ( )
3 3
1 2 2
= +1− =
3 3 3

2. Normal or Gaussian Distribution

(1) Let µ ∈ R and σ > 0 be real constants. A continuous random variable X is said
to have a normal (or Gaussian) distribution with parameters µ and σ 2 (written as
X ∼ N (µ, σ 2 )) if probability density function of X is given by
1 (x−µ)2
fX (x) = √ e− 2σ2 , −∞ < x < ∞
σ 2π
(2) The N (0, 1) distribution is called the standard normal distribution. The p.d.f.
and the d.f. of N (0, 1) distributions will be denoted by φ and Φ respectively, i.e.,
1 z2
φ(z) = √ e− 2 , −∞ < z < ∞
2π
Zz Zz
1 x2
Φ(z) = φ(x)dx = √ e− 2 dx.
2π
−∞ −∞
R∞ 2 √ R∞ x2 √
(3) We know that e−x dx = π and e− 2 dx = 2π.
−∞ −∞

Clearly if X ∼ N (µ, σ 2 ), then

1 x2
fX (µ − x) = fX (µ + x) = √ e− 2σ2 , ∀ x ∈ R
σ 2π
Thus the distribution of X is symmetric about µ. Hence,
1
X ∼ N (µ, σ 2 ) ⇒ FX (µ − x) + FX (µ + x) = 1, ∀ x ∈ R and FX (µ) = .
2
In particular,
1
Φ(−z) = 1 − Φ(z), ∀ z ∈ R and Φ(0) = .
2
X−µ
Suppose that X ∼ N (µ, σ 2 ). Then the p.d.f. of Z = σ is given by
fZ (z) = fX (µ + σz)|σ|, −∞ < z < ∞
1 z2
= √ e− 2 , −∞ < z < ∞
2π
i.e.,
X −µ
X ∼ N (µ, σ 2 ) ⇒ Z = ∼ N (0, 1).
σ
Thus
   
2 x−µ x−µ
X ∼ N (µ, σ ) ⇒ FX (x) = P ({X ≤ x}) = P Z≤ =Φ , ∀ x ∈ R.
σ σ
3
 Now, the m.g.f. of X ∼ N (µ, σ 2 ) is
MX (t) = E(etX )
Z∞
= etx fX (x)dx
−∞
Z∞
1 (x−µ)2
= √ etx e− 2σ 2 dx
σ 2π
−∞
∞
eµt √ x−µ
Z
2+
=√ e−y 2σty
dy (by putting √ = y)
π 2σ
−∞
σ 2 t2 Z∞
e(µt+ 2
)
−(y−
√
2σt 2
)
= √ e 2 dy
π
−∞
2 2
(µt+ σ 2t )
=e , ∀ t ∈ R.
Therefore,
(1) σ 2 t2
MX (t) = (µ + σ 2 t)e(µt+ 2
)
, ∀ t ∈ R;
(2) σ 2 t2
MX (t) = (σ 2 + (µ + σ 2 t)2 )e(µt+ 2
)
, ∀ t ∈ R;
(1)
E(X) = MX (0) = µ;
(2)
E(X 2 ) = MX (0) = µ2 + σ2;
and V ar(X) = E(X ) − (E(X))2 = σ 2 .
2