Scribd
Upload
23 views4 pages

Quiz 2 With Solutions

Uploaded by

islam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
23 views4 pages

Quiz 2 With Solutions

Uploaded by

islam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Massachusetts Institute of Technology

Department of Electrical Engineering & Computer Science


6.041/6.431: Probabilistic Systems Analysis
(Fall 2010)

6.041/6.431 Fall 2010 Quiz 2 Solutions


Problem 1. (80 points) In this problem:
(i) X is a (continuous) uniform random variable on [0, 4].
(ii) Y is an exponential random variable, independent from X, with parameter λ = 2.

1. (10 points) Find the mean and variance of X − 3Y .

E[X − 3Y ] = E[X] − 3E[Y ]


1
= 2−3·
2
1
= .
2

var(X − 3Y ) = var(X) + 9var(Y )


(4 − 0)2 1
= +9· 2
12 2
43
= .
12

2. (10 points) Find the probability that Y ≥ X.


(Let c be the answer to this question.)
The PDFs for X and Y are:

1/4, if 0 ≤ x ≤ 4,
fX (x) =
0, otherwise.

2e−2y , if y ≥ 0,

fY (y) =
0, otherwise.

Using the total probability theorem,


Z
P(Y ≥ X) = fX (x)P(Y ≥ X | X = x) dx
x
Z 4
1
= (1 − FY (x)) dx
0 4
Z 4
1 −2x
= e dx
0 4
1 4 −2x
Z
= 2e dx
8 0
1
= (1 − e−8 ).
8

Page 1 of 4
Massachusetts Institute of Technology
Department of Electrical Engineering & Computer Science
6.041/6.431: Probabilistic Systems Analysis
(Fall 2010)

3. (10 points) Find the conditional joint PDF of X and Y , given that the event Y ≥ X has
occurred.
(You may express your answer in terms of the constant c from the previous part.)
Let A be the event that Y ≥ X. Since X and Y are independent,

fX,Y (x, y) fX (x)fY (y)


fX,Y |A (x, y) = = for (x, y) ∈ A
P(A) P(A)
( −2y
4e
1−e−8
, if 0 ≤ x ≤ 4, y ≥ x
=
0, otherwise.

4. (10 points) Find the PDF of Z = X + Y .


Since X and Y are independent, the convolution integral can be used to find fZ (z).
Z z
1 −2t
fZ (z) = 2e dt
max(0,z−4) 4

 1/4 · 1 − e−2z
 
 , if 0 ≤ z ≤ 4,
= 1/4 · e8 − 1 e−2z , if z > 4,
0, otherwise.

5. (10 points) Provide a fully labeled sketch of the conditional PDF of Z given that Y = 3.
Given that Y = 3, Z = X + 3 and the conditional PDF of Z is a shifted version of the PDF of
X. The conditional PDF of Z and its sketch are:
fZ|Y =3 (z)
6
1
4

1/4, if 3 ≤ z ≤ 7,
fZ|{Y =3} (z) = -
0, otherwise. z
3 7

6. (10 points) Find E[Z | Y = y] and E[Z | Y ].


The conditional PDF fZ|Y =y (z) is a uniform distribution between y and y + 4. Therefore,

E[Z | Y = y] = y + 2.

The above expression holds true for all possible values of y, so

E[Z | Y ] = Y + 2.

7. (10 points) Find the joint PDF fZ,Y of Z and Y .


The joint PDF of Z and Y can be expressed as:

fZ,Y (z, y) = fY (y)fZ|Y (z | y)


1/2 · e−2y , if y ≥ 0, y ≤ z ≤ y + 4,

=
0, otherwise.

Page 2 of 4
Massachusetts Institute of Technology
Department of Electrical Engineering & Computer Science
6.041/6.431: Probabilistic Systems Analysis
(Fall 2010)

8. (10 points) A random variable W is defined as follows. We toss a fair coin (independent of Y ).
If the result is “heads”, we let W = Y ; if it is tails, we let W = 2 + Y . Find the probability of
“heads” given that W = 3.
Let X be a Bernoulli random variable for the result of the fair coin where X = 1 if the coin lands
“heads”. Because the coin is fair, P(X = 1) = P(X = 0) = 1/2. Furthermore, the conditional
PDFs of W given the value of X are:

fW |X=1 (w) = fY (w)


fW |X=0 (w) = fY (w − 2).

Using the appropriate variation of Bayes’ Rule:


P(X = 1)fW |X=1 (3)
P(X = 1 | W = 3) =
P(X = 1)fW |X=1 (3) + P(X = 0)fW |X=0 (3)
P(X = 1)fY (3)
=
P(X = 1)fY (3) + P(X = 0)fY (1)
P(X = 1)fY (3)
=
P(X = 1)fY (3) + P(X = 0)fY (1)
e−6
= .
e−6 + e−2

Problem 2. (30 points) Let X, X1 , X2 , . . . be independent normal random variables with mean 0
and variance 9. Let N be a positive integer random variable with E[N ] = P 2 and E[N 2 ] = 5. We
assume that the random variables N, X, X1 , X2 , . . . are independent. Let S = Ni=1 Xi .

1. (10 points) If δ is a small positive number, we have P(1 ≤ |X| ≤ 1 + δ) ≈ αδ, for some constant
α. Find the value of α.

P(1 ≤ |X| ≤ 1 + δ) = 2P(1 ≤ X ≤ 1 + δ)


≈ 2fX (1)δ.

Therefore,

α = 2fX (1)
2
1 1 (1−0)
= 2· √ e− 2 · 9
9 · 2π
2 1
= √ e− 18 .
3 2π

2. (10 points) Find the variance of S.


Using the Law of Total Variance,

var(S) = E[var(S | N )] + var(E[S | N ])


= E[9 · N ] + var(0 · N )
= 9E[N ] = 18.

Page 3 of 4
Massachusetts Institute of Technology
Department of Electrical Engineering & Computer Science
6.041/6.431: Probabilistic Systems Analysis
(Fall 2010)

3. (5 points) Are N and S uncorrelated? Justify your answer.


The covariance of S and N is

cov(S, N ) = E[SN ] − E[S]E[N ]


= E[E[SN | N ]] − E[E[S | N ]]E[N ]
XN XN
= E[E[ Xi N | N ]] − E[E[ Xi | N ]]E[N ]
i=1 i=1
= E[X1 ]E[N 2 ] − E[X1 ]E[N ]
= 0

since the E[X1 ] is 0. Therefore, S and N are uncorrelated.

4. (5 points) Are N and S independent? Justify your answer.


S and N are not independent.
Proof : We have var(S | N ) = 9N and var(S) = 18, or, more generally, fS|N (s | n) = N (0, 9n)
and fS (s) = N (0, 18) since a sum of an independent normal random variables is also a normal
random variable. Furthermore, since E[N 2 ] = 5 6= (E[N ])2 = 4, N must take more than one
value and is not simply a degenerate random variable equal to the number 2. In this case, N can
take at least one value (with non-zero probability) that satisfies var(S | N ) = 9N 6= var(S) = 18
and hence fS|N (s | n) 6= fS (s). Therefore, S and N are not independent.

Page 4 of 4

You might also like