math523-3
DESCRIPTION
Math523-3TRANSCRIPT
Homework # 3
7.1. By additivity∞∑
n=1
p(An) = P(
∞⋃
n=1
An
)
≤ 1
This leads to limn→∞
P (An) = 0.
7.11. First, the integral
P (A) =
∫
A
f(x)dx A ∈ B(R)
defines an probability measure on (R,B(R)) with the continuous distribution function
F (x) =
∫ x
−∞
f(u)du x ∈ R
Second, for any integer n ≥ 1, x0 ⊂ (x0 − n−1, x0]. Subsequently,
0 ≤ P (x0) ≤ P(
(x0 − n−1, x0])
= F (x0) − F (x0 − n−1) n = 1, 2, · · ·
By continuity of F at x0, the right hand side tends to 0 as n → ∞. So we have P (x0) = 0.
7.14. Null set in a complete probability space (Ω,A, P ) means a set A ⊂ A withP (A) = 0. In this case, we have P (An) = 0 for n = 1, 2, · · ·. By sub-additivity
P (B) ≤∞∑
n=1
P (An) = 0
So B is null set.
Without completness of (Ω,A, P ), there are Cn ∈ A such that An ⊂ Cn and P (Cn) =
0 for n = 1, 2, · · ·. Let C =
∞⋃
n=1
Cn. Then C ∈ A and B ⊂ C. By sub-additivity
P (C) ≤
∞∑
n=1
P (Cn) = 0
So B is null set.
7.15. Since Ω is countable, X has at most countable many values, say, x1, · · · , xn, · · ·.
0 = E|X | =∑
n
|xn|PX = xn
1
This leads to PX = xn = 0 for xn 6= 0. In other words,
PX 6= 0 =∑
xn 6=0
PX = xn = 0
In general, we can not conclude X 6= 0 = φ. For example, randomly pick up a pointin (−∞,∞) and assume the picked number ξ obeys the standard normal distribution.Write the sample space Ω = ω0, ω1, where
ω0 = ξ is not an integer and ω1 = ξ is an integer,
Define X(ω0) = 0 and X(ω1) = 1. Then E|X | = 0 and X = 1 = ξ is an integer 6= φ.
9.2* Let x ∈ R be fixed but arbiotrary. By assumption X ≤ x ∈ F and X ≤ x ∈G. Therefore,
F (x) = PX ≤ x = P(
X ≤ x ∩ X ≤ x)
= PX ≤ x · PX ≤ x = F 2(x)
Thus, for any x, F (x) = 0 or 1. By the property of distribution function, F (·) is non-decreasing and right continuous on R and
limx→−∞
F (x) = 0 and limx→∞
F (x) = 1
Therefore, there is a constant c such that F (x) = 0 for x < c and F (x) = 1 for x > c.
We now claim that PX = c = 1. Indeed,
X = c =∞⋂
n=1
c −1
n< X ≤ c +
1
n
By continuity theorem,
PX = c = P(
∞⋂
n=1
c −1
n< X ≤ c +
1
n
)
= limn→∞
P
c −1
n< X ≤ c +
1
n
= limn→∞
F(
c +1
n
)
− F(
c −1
n
)
= limn→∞
(1 − 0) = 1
2