CDF Q's and Ans

University of Illinois Spring 2013
ECE 313: Problem Set 7: Problems and Solutions

CDF and pdf; Uniform and Exponential random variables
Due: Wednesday, March 6 at 6 p.m.
Reading: ECE 313 Course Notes, Sections 3.1–3.4
1. [Cumulative Distribution Function]

For each of the following functions Fi (c), state whether or not Fi (c) is the CDF of some
random variable. If not, state which of the properties of a CDF it violates. If so, find the
corresponding pmf or pdf.
(a) 

 0 c≤0
0.5c 0<c≤1

F1 (c) =

 0.25 + 0.25c 1<c≤3
1 3<c

Solution: This is a valid CDF; the corresponding pdf is


 0.5 0 < c < 1
f1 (u) = 0.25 1 < c < 3
0 otherwise

(b) 

 0 c≤0
0.5 0<c≤1

F2 (c) =

 0.75 1<c≤3
1 3<c

Solution: This is not a valid CDF, because it is not right-continuous.

(c) 
 0.5 c < 1
F3 (c) = 0.75 1 ≤ c < 3
1 3≤c

Solution: This is not a valid CDF, because it does not approach 0.0 as c → −∞.
(d) 

 0 c<0
0.5 0≤c<1

F4 (c) =

 0.75 1≤c<3
1 3≤c

Solution: This is a valid CDF. The corresponding pmf is



 0.5 k = 0
0.25 k = 1

p4 (k) =

 0.25 k = 3
0 otherwise

(e) 

 0 c<0
0.5 0≤c<1

F5 (c) =

 0.25 1≤c<3
1 3<c

Solution: This is not a valid CDF, because it is not non-decreasing.

(f) 
 0 c≤0
F6 (c) = 0.5c 0<c≤1
0.25 + 0.25c 1 < c

Solution: This is not a valid CDF, because it does not approach 1.0 as c → ∞.
2. [Continuous Random Variables]

Random variable X is distributed with the following pdf:

sin(x) 0 ≤ x ≤ A
fX (u) =
0 otherwise
(a) What is the value of the constant A? Solution: fX (u) must integrate to one. We have
that Z ∞ Z A
fX (u)du = sin(x) = 1 − cos(A)
−∞ 0
R∞
Thus, to achieve −∞ fX (u)du = 1, we need A = π2 .
(b) What is the corresponding CDF, FX (c)? Solution:

Z c  0 c≤0
π
FX (c) = fX (u)du = 1 − cos(c) 0 ≤ c ≤ 2
−∞ π
1 2 ≤c

(c) What is E[X]? Solution:

Z π/2 Z π/2
π/2
E[X] = u sin(u)du = [−u cos(u)]0 + cos(u)du = 1
0 0
(d) What is Var(X)? Solution:

Z π/2
2
E[X ] = u2 sin(u)du
0
π/2
Z π/2
2

= −u cos(u) 0
+2 u cos(u)du
0
Z π/2
π/2
= 0+ 2 [u sin(u)]0 −2 sin(u)du
0
= π−2
Therefore
Var(X) = E[X 2 ] − E 2 [X] = π − 3
2
3. [Independent intervals and the exponential pdf ]
Consider a process in which successes can occur at any time. The number of successes
occurring in any time interval is independent of the number of successes occurring in any
other time interval unless the intervals overlap. Define
G1 (u) ≡ Pr {T1 > u} , L1 (u) ≡ ln Pr {T1 > u}
Consider the two non-overlapping intervals (0, u] and (u, v]. Since these two intervals are
non-overlapping, what happens in the second interval is independent of what happens in the
first interval, as long as u ≥ 0 and v ≥ u. Let A be the event that there are no successes in
the first interval, and let B be the event that there are no successes in the second interval.
Since A and B are independent events,
P (AB) = P (A)P (B)
But notice that P (A) = G1 (u), P (B) = G1 (v − u), and P (AB) = G1 (v), therefore
G1 (v) = G1 (u)G1 (v − u), u ≥ 0, v ≥ u (1)
or equivalently,
L1 (v) = L1 (u) + L1 (v − u), u ≥ 0, v ≥ u
(a) According to Eq. 1, what must be the dependence of L1 (u) on u, for u > 0? Your answer
should be one word.
Solution: Linear.
(b) Consider the following equations:
P {T1 > 0} = 1 (2)

P {T1 > ∞} = 0 (3)
Describe the set of all functions G1 (u) that satisfy Eq. 1, Eq. 2, and Eq. 3.
Solution: The only functions that satisfy these three conditions are functions of the
form G1 (u) = e−λu for positive u, and for some real-valued constant λ.
(c) What is the relationship between the CDF, FT (u), and the function G1 (u)?
Solution: FT (u) = 1 − G1 (u).
4. [Central moments of the uniform]

A uniform random variable is defined by the pdf
1
fX (u) = b−a a ≤ u ≤ b
0 otherwise
For the purpose of calculating its central moments, though, it is more useful to define it as
1 ∆ ∆
fX (u) = ∆ µ− 2 ≤u≤µ+ 2
0 otherwise
where the relationship between the parameters (a, b) and the parameters (µ, ∆) is given by
a+b
µ= , ∆=b−a
2
3
(a) In terms of ∆, find the nth central moment, E[(X − µ)n ], for odd integer values of n.
Solution:
Z µ+∆/2
n
E[(X − µ) ] = (u − µ)n fX (u)du
µ−∆/2
Z ∆/2
1
= v n dv
∆ −∆/2
1 n+1 ∆/2
= v −∆/2
(n + 1)∆
n+1 !
∆ n+1

1 ∆
= − −
(n + 1)∆ 2 2
For odd-valued n, the last line simplifies to E[(X − µ)n ] = 0.

(b) In terms of ∆, find the nth central moment, E[(X − µ)n ], for even integer values of n.
Solution: n
n 1 ∆
E[(X − µ) ] =
n+1 2
(c) Express the standard deviation σX in terms of ∆.

Solution: σX = E[(X − µ)2 ] = √13 ∆
p
2

CDF Q's and Ans

Uploaded by

Copyright:

Available Formats

CDF Q's and Ans

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CDF Q's and Ans

Uploaded by

Copyright:

Available Formats

University of Illinois Spring 2013

ECE 313: Problem Set 7: Problems and Solutions

1. [Cumulative Distribution Function]

Solution: This is a valid CDF; the corresponding pdf is

Solution: This is not a valid CDF, because it is not right-continuous.

Solution: This is a valid CDF. The corresponding pmf is

Solution: This is not a valid CDF, because it is not non-decreasing.

2. [Continuous Random Variables]

(c) What is E[X]? Solution:

(d) What is Var(X)? Solution:

G1 (u) ≡ Pr {T1 > u} , L1 (u) ≡ ln Pr {T1 > u}

P (AB) = P (A)P (B)

G1 (v) = G1 (u)G1 (v − u), u ≥ 0, v ≥ u (1)

P {T1 > 0} = 1 (2)

4. [Central moments of the uniform]

For odd-valued n, the last line simplifies to E[(X − µ)n ] = 0.

(c) Express the standard deviation σX in terms of ∆.

You might also like