Documente Academic
Documente Profesional
Documente Cultură
Spring 2016
Problem Set 1
Author: Xiaobin Gao
1. Problem 1
(a) Sample space = {HH, HT, T H, T T }.
(b) Event space
F = 2 =
n
, {HH}, {HT }, {T H}, {T T }
{HH, HT }, {HH, T H}, {HH, T T }, {HT, T H}, {HT, T T }, {T H, T T }
{HH, HT, T H}, {HH, HT, T T }, {HH, T H, T T }, {HT, T H, T T }
o
{HH, HT, T H, T T }
3
4
34 , etc.
9
3
16 + 16 , etc.
1
16 , etc.
P ({HH, HT, T H, T T }) = 1
Note that
P (first toss is head) = P (first toss is head|fair coin is picked)
1
2
1
2
1
2
+1
1
2
1
2
3
4
(d) Since the outcomes of the first and the second tosses are independent,
P (first toss is head|second toss is tail) = P (first toss is head) =
3
4
2. Problem 2
(a) We use a four digit number, call it abcd, to represent case that packets 1, 2, 3, 4 are
routed to ports a, b, c, d, respectively, where a, b, c, d {1, 2, . . . , 8}. Then the sample
space
n
o
= abcd|a, b, c, d {1, 2, . . . , 8}
The event space is the power set of sample space, i.e., F = 2 . For any event A F,
P (A) =
where card(A) is the cardinality of A.
1
card(A)
84
(b)
1
,
84
4
,
84
6
,
84
12
,
84
24
,
84
P (X1 = 4) =
P (X1 = 3, X2 = 1) =
P (X1 = 2, X2 = 2) =
P (X1 = 2, X2 = 1, X3 = 1) =
P (X1 = 1, X2 = 1, X3 = 1, X4 = 1) =
etc.
etc.
etc.
etc.
etc.
P4
p=1 Xip ,
P
P4
4
X
Cov(X1 , X2 ) = Cov
X
,
q=1 2q
p=1 1p
P4 P4
=
p=1
q=1 Cov(X1p , X2q )
When p 6= q, X1p and X2q are independent. Hence Cov(X1p , X2q ) = 0. When p = q,
X1p and X2p can not be 1 simultaneously, i.e., X1p X2p = 0. Hence, Cov(X1p , X2p ) =
1
E[X1p X2p ] E[X1p ] E[X2p ] = 0 81 18 = 64
. Then,
Cov(X1 , X2 ) =
1
16
(d)
P (Xi 1 for all i) =
8765
105
=
4
8
256
(e)
P (Xi 2 for all i) =
8
4
4! +
8
1
7
2
84
4
2
2
1
8
2
4
2
=
483
512
3. Problem 3
(a) All bytes are equally probable is equivalent to All bits have Bern( 12 ) distribution,
mutually independent. Hence,
P (A) =
1 1
1
=
2 2
4
1
2
1
2
1
2
1
2
1
2
1
2
Hence, P (B) = 12 .
(c)
P (B|A) = P
8
X
!
bi is odd|b1 = b2 = 1
=P
i=1
8
X
!
bi is odd
i=3
1
2.
(d)
P (A|B) =
1
P (B|A)P (A)
=
P (B)
4
4. Problem 4
(a) P (X > 3) = 1 P (X = 3) = 0.8.
P (X > 8, X > 5)
P (X > 8)
=
= 0 6= P (X > 3).
P (X > 5)
P (X > 5)
Hence, the distribution of X does not have memoryless property.
P (Y > 8, Y > 5)
P (Y > 8)
(1 p)8
=
=
= (1 p)3 = P (Y > 3)
P (Y > 5)
P (Y > 5)
(1 p)5
2
.
2
1
6
+
3 2
where the first equality dues to the independence of U, V, W and the linearity of expectation.
(c)
Cov(2 + U, U V ) = Cov(2, U V ) + Cov(U, U V )
= 0 + E[U 2 V ] E[U ] E[U V ]
= E[U 2 ] E[V ] E[U ] E[U ] E[V ]
=
1
3
6. Problem 6
(a) The cdf of X, denoted by FX (x), is given as
0,
0.5
FX (x) =
2x
x<0
0x<1
1x<2
x2
1
1
1
+ (x)
2 {1<x<2} 2
1
E[X] = 0 +
2
1
E[X ] = 0 +
2
x
1
1
3
dx =
2
4
(c)
2
Hence, Var(X) =
E[X 2 ]
E[X]2
x2
1
7
dx =
2
6
29
48 .
7. Problem 7
(a) X1 has Bernoulli distribution with parameter 61 . Hence, E[X1 ] = 16 , Var(X1 ) =
5
36 .
(b)
"
E[X] = E
n
X
#
Xi =
i=1
Var(X) = Cov(X, X) =
n
X
E[Xi ] =
i=1
n X
n
X
n
6
Cov(Xi , Xj )
i=1 j=1
5n
36
n
36
Cov(X, Y )
n
1
p
Cov(X
,
Y
)
=
=
=
i
j
XY
j=1
36
5n = 5 .
Var(X)Var(Y )
36
Pn Pn
i=1
(e) Let Zi , Oi , Pi , Qi be binary random variables, where 1 means 3, 4, 5, 6 shows on the i-th
roll, respectively. Define Z, O, P, Q as the sum of Zi , Oi , Pi , Qi , respectively. It is clear
that X + Y + Z + O + P + Q = n, hence
E[Y + Z + O + P + Q|X = x] = n x
By symmetry, the conditional distribution of Y given X = x is same as the conditional
distribution of Z given X = x, etc. Then,
E[Y |X = x] = E[Z|X = x] = E[O|X = x] = E[P |X = x] = E[Q|X = x]
which implies that E[Y |X = x] =
nx
5 .
8. Problem 8
(a)
P (X 1, Y 2) = FXY (1, 2) = 1 +
1 7
e e1 e6
4
(b)
1
FX (x) = lim FXY (x, y) = 1 ex 1{x0}
y
4
1 3y
1{y0}
FY (y) = lim FXY (x, y) = 1 e
x
4
(c) FXY (x, y) 6= FX (x) Fy (y). Hence, X and Y are not independent.
9. Problem 9
(a) The joint pdf of (X, Y ), denoted by fXY (x, y), is described as follows
12 ,
X2 + Y 2 a
a
fXY (x, y) =
0,
otherwise
When 0 r < a,
P (R r) = P (X 2 + Y 2 r2 ) =
Z
fXY (x, y) dxdy =
X 2 +Y 2 r2
Hence,
0,
2
r
FR (r) =
,
a2
1,
r<0
0r<a
r0
(b)
2r ,
d
2
fR (r) = FR (r) = a
dr
0,
10. Problem 10
0<r<a
otherwise
r2
a2
(a)
P
P
X1 ,X2 ,X3 (u1 , u2 , u3 ) = exp j 3k=1 kuk 3k=1 k 2 u2k
= exp ju1 u21 exp 2ju2 4u22 exp 3ju3 9u23
Since the joint characteristic function factors, X1 , X2 , X3 are mutually independent, and
the characteristic function of X2 is given by
X2 (u2 ) = exp 2ju2 4u22
which is the characteristic function of Gaussian distribution N (2, 8).
(b) By part (a), E[X2 ] = 2, Var(X2 ) = 8.