Chapter 8

8.1-1
M = 12 equally likely outcomes
P(A) = 6/12, P(B) = 4/12, P(C) = 3/12 A 1 2 3 4 C
P(AB) = 2/12, P(AC) = 0, P(BC) = 1/12 5 6 7 8
P(AcB) = 2/12 B 9 10 11 12

8.1-2
M = 16 equally likely outcomes A 1,1
P(A) = 4/16, P(B) = 6/16, P(C) = 6/16 1,2 2,1 B
P(AB) = 0, P(AC) = 2/16, P(BC) = 2/16 1,3 2,2 3,1
P(AcB) = 6/16 1,4 2,3 3,2 4,1
2,4 3,3 4,2
3,4 4,3
C 4,4

8.1-3

P ( A B c ) = N A B c / N = (N A - N A B )/ N = P ( A ) - P (A B )

8.1-4
N A = N AB + N ABc N B = N A B + N Ac B

N AB + NABc + NAc B N A + N B − N AB
P( A + B ) = = = P ( A) + P ( B) − P( AB)
N N

8.1-5
N A = N AB + N ABc N B = NAB + NAc B

N ABc + N Ac B N A + N B − 2 N AB
P(C ) = = = P( A) + P( B ) − 2 P( AB)
N N

8-1
8.1-6
P(match) = P( HH + TT ) = P( HH ) + P(TT ) = P( H ) P (H ) + P (T )P (T )

1+ ∈ 1−∈  1+ ∈   1− ∈  1+ ∈ 1
2 2 2
P(T ) = P( H ) = 1 −
c
= P(match) =  +
   = >
2 2  2   2  2 2

8.1-7
Let A = “A fails,” B = “B fails,” C = “computer inoperable”
P(A) = 0.01, P(B) = 0.005, P(B|A) = 4 × 0.005 = 0.02
P(C) = P(AB) = P(B|A)P(A) = 0.0002, P(A|B) = P(AB)/P(B) = 0.04

8.1-8
Let M = “match,” H1 = “heads on first toss,” etc.
(a) P(H1 ) = ½, P(MH1 ) = P(H1 H2 ) = (½)2 , P(M|H1 ) = P(MH1 )/P(H1 ) = ½
(b) Let A = “H1 or H2 ,” P(A) = P(H1 T2 + T1 H2 + H1 H2 ) = ¾
P(MA) = ¼, P(M|A) = P(MA)/P(A) = 1/3
(c) P(M) = P(H1 H2 + T1 T2 ) = ½, P(A|M) = P(A)P(M|A)/P(M) = ½

8.1-9
Let M = “match,” H1 = “heads on first toss,” etc.
(a) P(H1 ) = ¼, P(MH1 ) = P(H1 H2 ) = (¼)2 , P(M|H1 ) = P(MH1 )/P(H1 ) = ¼
(b) Let A = “H1 or H2 ,” P(A) = P(H1 T2 + T1 H2 + H1 H2 ) = 2 × ¼ × ¾ + (¼)2 =7/16
P(MA) = P(H1 H2 ) = (¼)2 , P(M|A) = P(MA)/P(A) = 1/7
(c) P(M) = P(H1 H2 ) + P(T1 T2 ) = (¼)2 + (¾)2 = 10/16, P(A|M) = P(A)P(M|A)/P(M) = 1/10

8.1-10
Since P(AB) = P(A|B)P(B) = P(B|A)P(A), P(XYZ) = P(X)P(YZ|X) where
P ( XYZ ) P ( XY ) P ( XYZ )
P(YZ|X) = = = P(Y|X)P(Z|XY) so P(XYZ) = P(X)P(Y|X)P(Z|XY)
P( X ) P( X ) P( XY )

8.1-11
Let F = “fair coin,” L = “loaded coin,” A = “all tails,” P(F) = 1/3, P(L) = 2/3, P(A|F) = (½)2 , P(A|L) =
(¾)2 (a) P(A) = P(A|F)P(F) + P(A|L)P(L) = 11/24
(b) P(L|A) = P(L)P(A|L)/P(A) = 9/11

8-2
8.1-12
Let F = “fair coin,” L = “loaded coin,” A = “all tails;” P(F) = 1/3, P(L) = 2/3, P(A|F) = (½)2 , P(A|L) =
(¾)2 (a) P(A) = P(A|F)P(F) + P(A|L)P(L) = 31/96
(b) P(L|A) = P(L)P(A|L)/P(A) = 27/31
8.1-13
Let R1 = “first marble is red,” etc., M = “match;” P(R1 ) = 5/10, P(W1 ) = 3/10, P(G1 ) =2/10,
P(M|R1 ) = P(R2 |R1 ) = (5 – 1)/(10 – 1) = 4/9, P(M|W1 ) = 2/9, P(M|G1 ) = 1/9
(a) P(M) = P(M|R1 ) × P(R1 ) + P(M|W1 ) × P(W1 ) + P(M|G1 ) × P(G1 )
4 5 2 3 1 2 14
= × + × + × =
9 10 9 10 9 10 45
(b) P(W1 |M) = P(W1 )P(M|W1 )/P(M) = 3/14

8.1-14
Let R1 = “first marble is red,” etc., M = “match;” P(R1 ) = 5/10, P(W1 ) = 3/10,

P(G1 ) =2/10, P(M|R1 ) = P(R3 R2 |R1 ) = P(R3 |R2 R1 )P(R2 |R1 ) = 5 − 2 × 5 − 1 = 3 × 4 , P(M|W1 ) = 1 × 2 ,
10 − 2 10 − 1 8 9 8 9

P(M|G1 ) = 0 × 1
8 9

(a) P(M) = P(M|R1 ) × P(R1 ) + P(M|W1 ) × P(W1 ) + P(M|G1 ) × P(G1 )

= 12 × 5 + 2 × 3 + 0 × 2 = 11
72 10 72 10 10 120

(b) P(W1 |M) = P(W1 )P(M|W1 )/P(M) = 1/11

8-3
8.2-1
P(Ni) = 1/5 N xi PX(x i) FX(x i)
0 0 0.2 0.2
-1,1 0.5 0.4 0.6
2 2 0.2 0.8
3 4.5 0.2 1.0
FX 1.0
0.8
0.6

0.2

0 0.5 2.0 4.5 x

(cont.)

P(X W 0) = FX(0) = 0.2, P(2 < X W 3) = FX(3) – FX(2) = 0, P(X < 2) = FX(2 - ∈) = 0.6,
P(X D 2) = 1 – 0.6 = 0.4

8.2-2
P(Ni) = 1/5 N xi PX(x i) FX(x i)
3 -4 0.2 0.2
2 -2 0.2 0.4
-1,1 2 0.4 0.8
0 4 0.2 1.0

FX 1.0
0.8

0.4
0.2
-4 -2 0 2 4 x

8-4
P(X W 0) = FX(0) = 0.4, P(2 < X W 3) = FX(3) – FX(2) = 0, P(X < 2) = FX(2 - ∈) = 0.4,
P(X D 2) = 1 – 0.4 = 0.6

8.2-3

0 x≤0

x

FX ( x) = ∫ pX (λ) d λ =  x − λ
 ∫ λe d λ = 1 − ( x + 1)e
−x
−∞
x>0
0
P(X W 1) = FX(1) = 0.264, P(X > 2) = 1 – FX(2) = 0.406, P(1 < X W 2) = FX(2) – FX(1) = 0.330

8.2-4
x 1 λ 1 x
 ∫ e dλ = e x≤ 0
 −∞ 2
x
2
FX ( x) = ∫ pX (λ) d λ =  x
 1 + 1 e−λ dλ = 1 − 1 e−x
2 ∫ 2
−∞
x>0
 0
2

P(X W 0) = FX(0) = 1/2, P(X > 1) = 1 – FX(1) = 0.184, P(0 < X W 1) = FX(1) – FX(0) = 0.316

8.2-5
FX(∞) = 100K = 1 ⇒ K = 0.01 so pX(x) = dFX(x)/dx = 0.2x[u(x) – u(x – 10)]
P(X W 5) = FX(5) = K × 52 = 0.25, P(5 < X W 7) = FX(7) – FX(5) = 0.49 – 0.25 = 0.24

8.2-6
2π πx
FX(∞) = K/ 2 = 1 ⇒ K = 2 so pX(x) = dFX(x)/dx = cos [u(x) – u(x – 10)]
40 40

P(X W 5) = FX(5) = K sin π = 0.541, P(5 < X W 7) = FX(7) – FX(5) = K sin 7π – 0.541 = 0.198
8 40

8.2-7
P(Z < 0) = 0, P(Z W 0) = P(X W 0) = ½, P(Z W z) = P(X W z) for z > 0

0 z<0 0 z<0
 d 
FZ (z ) =  1 1 p Z (z) = FZ ( z ) =  1 1
 2 + π arctan z z ≥ 0 dz  2 δ( z ) + π(1 + z 2 ) z≥0


8-5
8.2-8
P(Z < -1) = 0, P(Z W -1) = P(X W 0) = ½, P(Z W z) = P(X W z) for z > 0


0 z < −1 1
 d  2 δ( z + 1) z≤0
FZ ( z ) = 1 / 2 −1 ≤ z ≤ 0 p Z (z) = F ( z) = 
1 1 dz Z  1
z>0
 + arctan z z≥0  π(1 + z 2 )
2 π

8.2-9
1 −2( z +5)/2  z + 5  − ( z+ 5)
p Z (z) = 2e u =e u ( z + 5)
2  2 

2
1 pZ pX

-5 -4 0 0.5 x, z

8.2-10
1  z − 1  ( z −1)
p Z (z) = 2e −2( z −1)/( −2) u   = e u [− (z − 1)]
−2  −2 

2
pZ 1 pX

-2 -1 0 0.5 x, z

8.2-11
Monotonic transformation with g-1 (z) = z2 – 1, dg-1 /dz = 2z, pX(x) = ¼ for –1 W x W 3, so
1 z
p Z (z) = 2 z [ u ( z ) − u ( z − 2)] = [u ( z ) − u ( z − 2)]
4 2

8-6
8.2-12
g1 (x) = -x[u(x + 1) – u(x)], gl-1 (z) = -z[u(z) – u(z – 1)], dg1 -1 /dz = -1
g2 (x) = x[u(x) – u(x – 3)], g2 -1 (z) = z[u(z) – u(z – 3)], dg2 -1 /dz = 1, pX(x) = ¼ for –1 W x W 3, so

1 1 1
 4 −1 + 4 1 = 2 0 ≤ z ≤1
p Z (z ) = 
1 1 = 1 1< z ≤ 3
 4 4

8.2-13

g1 (x) = − x [u(x + 1) – u(x)], gl-1 (z) = -z2 [u(z) – u(z – 1)], dg1 -1 /dz = -2z

g2 (x) = x [u(x) – u(x – 3)], g2 -1 (z) = z2 [u(z) – u(z – 3 )], dg2 -1 /dz = 2z, pX(x) = ¼ for –1 W x W
3, so

1 1
 4 −2 z + 4 2 z = z 0 ≤ z ≤1
p Z (z ) = 
1 2z = z 1< z ≤ 3
 4 2

8.2-14

g1 (x) = x 2 u(-x), gl-1 (z) = - z u(z), dg1 -1 /dz = -1/2 z , g2 (x) = x 2 u(x),

g2 -1 (z) = + z u(z), dg2 -1 /dz = +1/2 z (cont.)

1 1
pZ(z) = 0 for z < 0, p Z ( z ) = pX ( z ) + pX ( − z ) − for z > 0, so
2 z 2 z
1 
p Z (z) = p X ( z ) + pX ( − z )  u ( z )
2 z

8.2-15
∞ ∞
pY ( y ) = ∫ pXY ( x , y) dx = ye− yu ( y ) ∫ e− yx dx = e − yu ( y ) ,
−∞ 0

p XY ( x , y ) = ye− yx u( x) e − yu ( y ) ≠ pX ( x) pY ( y) , pX(x|y) = p XY ( x , y ) / pY ( y ) = ye− yx u( x)

8-7
8.2-16
1 y 1 2  y
Π   ∫ ( x + 2 xy + y ) dx = (1+ 3 y )Π   ,
∞ 1
pY ( y ) = ∫ pXY ( x , y) dx =
2 2
−∞ 40  6  −1 60 6

3 ( x + y )2  x  1 y
p XY ( x , y ) = Π   (1 + 3y 2 ) Π   ≠ p X ( x) pY ( y)
2 (1 + 3y )  2  60
2
6
3( x + y) 2  x 
pX(x|y) = p XY ( x , y ) / p Y ( y ) = Π
2(1 + 3y 2 )  2 

8.2-17
∞ ∞ p XY ( x , y) 1 ∞
∫−∞
pX(x|y) dx = ∫−∞ pY ( y )
dx = ∫
pY ( y) −∞
pXY ( x , y ) dx = 1

For any given Y = y, X must be somewhere in the range -∞ < x < ∞.

8.2-18
∞ ∞
pXY(x,y) = pX(x|y)pY(y), p X ( x) = ∫ pXY ( x , y) dy = ∫ pX (x|y)pY(y) dy
−∞ −∞

∞
Thus, pY(y|x) = pXY(x,y)/pX(x) = pX(x|y)pY(y)/ ∫ pX (x|y)pY(y) dy
−∞

8.3-1
2
∞ ∞ 2 1 1
mX = a ∫ xe − ax dx = 1/ a, X 2 = a ∫ x2 e− ax dx = 2 / a 2 , so σ X = −  =
a  a
2
0 0 a

8.3-2
2
∞ ∞ 6 2 2
∫ ∫
2 − ax 3 − ax
mX = a 2
xe dx = 2 / a , X = a 2 2
xe dx = 6 / a , so σ X =
2
−  =
a a
2
0 0 a

8.3-3

2 ∞ x 2 ∞ λ ∞ a 
mX =
π ∫ −∞ 1 + ( x − a) 4
dx =  ∫
π  −∞ 1 + λ 4
dλ + ∫
−∞ 1 + λ 4
d λ  = a,


2 ∞ x2 2  ∞ λ2 ∞ 2λ ∞ a2 
X2 = ∫−∞ 1 + ( x − a )4 = ∫ λ + ∫ λ + ∫ λ  = 1 + a , so
2

π
dx 
π  −∞ 1 + λ 4
d −∞ 1 + λ4
d −∞ 1 + λ 4
d


σX = (1+ a ) − a 2 2
=1

8-8
8.3-4

P(X = b) = 1 – p, mX = ap + b(1 – p), X 2 = a2 p + b2 (1 – p),

σX2 = a2 p + b2 (1 – p) – [ap + b(1 – p)]2 = (a – b)2 p(1 – p), σX = |a – b| p (1 − p)

8.3-5
1 a ( K − 1) K K − 1
m X = ∑ i =0 ai
K −1
= = a,
K K 2 2
1 a 2 ( K −1) K [2( K − 1) + 1] ( K −1)(2 K − 1) 2
X = ∑ i= 0
K −1
2
( ai) 2
= = a
K K 6 6

( K − 1) a 2 2(2 K − 1) − 3( K − 1) K 2 − 1 2 K 2 −1 a
σ X = X 2 − mX = = a , σX =
2 2

2 6 12 3 2

8.3-6
∞ a θ +2 π
mY = ∫ a cos x pX ( x) dx =
2π ∫θ
cos xdx = 0 ,
−∞

∞ a 2 θ +2 π 2 a2
Y 2 = ∫ a2 cos 2 x pX ( x) dx =
2π ∫θ
cos xdx = σY = a2
2 −0 = a
2
−∞ 2

8.3-7
∞ a θ+π 2a
mY = ∫ a cos x pX ( x) dx = ∫ cos xdx = − sin θ ,
−∞ π θ π
∞ a2 θ+π a2
Y 2 = ∫ a2 cos 2 x pX ( x) dx = ∫ cos 2 xdx =
−∞ π θ 2
2
a 2  2a  1 4
σY = −  − sin θ  = a − 2 sin 2 θ
2  π  2 π

8.3-8

mY = αmX + β, Y 2 = E ( αX + β) 2  = E α 2 X 2 + 2αβX + β2  = α 2 X 2 + 2αβ mX + β2

σY = Y 2 − mY = α 2 X 2 − mX
2 2
( 2
)=ασ 2
X
2
, σY = α σ X

8-9
8.3-9

Y 2 = E ( X + β) 2  = E  X 2 + 2β X + β2  = X 2 + 2βmX + β 2

d 2
Y = 2 mX + 2β = 0 ⇒ β = − mX
dβ

8.3-10
∞
P( X ≥ a ) = ∫ p X ( x ) dx and pX ( x ) = 0 for x < 0
a

∞ ∞ ∞
E [ X ] = ∫ p X ( x ) dx ≥ ∫ xp X ( x ) dx ≥ a ∫ p X ( x) dx = aP ( X ≥ a) , so P( X ≥ a) ≤ m X / a
0 a a

8.3-11

E ( X ± Y )  = E  X 2 ± 2 XY + Y 2  = X 2 ± 2 XY + Y 2 ≥ 0
2

 

(
so 2 XY ≥ − X + Y 2 2
) and 2 XY ≤ ( X 2
+Y 2
) , ⇒ −
X 2 +Y2
2
≤ XY ≤
X2 +Y2
2

8.3-12

CXY = E [ XY − mX Y − mY X + mX mY ] = XY − mX mY

(a) XY = XY = m X mY ⇒ C XY = 0

(b) XY = E [ X (αX + β) ] = α X 2 + βmX and mY = αmX + β so C XY = α X 2 − mX =ασ X ( ) 2

8.3-13

∈2 = E Y 2 − 2 (α X + β) Y + ( αX + β)  = Y 2 − 2α XY − 2 βY + α 2 X 2 + 2αβ X + β2
2

 

∂ ∈2 ∂ α = −2 XY + 2α X 2 + 2 βY = 0 and ∂ ∈2 ∂ β = −2Y − 2α X + 2β = 0 so

( )
α = XY − X Y / σX 2 and β = Y − α X

8.3-14

dn dn
E e jνX  = E  ( jX ) e jνX  = j n E  X ne jνX  , so
n
Φ (ν ) =
dν n X
dν n  

dn −n d
n
   
Φ X (0) = j E  X  ⇒ E  X  = j
n n n
Φ X (ν)
d νn d νn ν=0

8-10
8.3-15
a a
F  ae− atu (t )  =
= ⇒ F −1
 aae− af u ( f )  = aso
a + j 2 πf a + j 2π( −t )
−1
a a  ν
Φ X (2πt ) = and Φ X ( ν ) = = 1− j 
a − j 2π t a − jν  a
−2
dΦ X  ν  j  j 1
= − 1 − j   −  ⇒ X = j −1   =
dν  a  a a  a
−3
d 2Φ X ν  j
2 2
  j 2
= 2 1 − j   −  ⇒ X 2 = j −2 2   = 2
dν 2
 a  a a a
−4
d 3Φ X ν   j
3 3
  j 6
= −6  1 − j   −  ⇒ X 3 = j −3 ( −6 )   = 3
dν 3
 a  a a a

8.3-16
∞ ∞
Φ Y (ν ) = E  e jνx  = ∫0 e jνx 2axe − ax dx = ∫0 e jνλ ae −aλ d λ ⇒
2 2 2
pY ( y ) = ae − ay u ( y )

8.3-17
π/ 2 1
Φ Y (ν ) = E  e jνsin x  = ∫ e jν sin x dx
−π /2 π

Let λ = sin x, dλ = cos x dx where cos x = 1 − sin 2 x = 1 − λ2 , so

sin π / 2 1 dλ 1 1  y
Φ Y (ν ) = ∫ ∫
1
e jνλ = e jνλ
dλ ⇒ pY ( y ) = Π  
sin ( −π / 2 ) π 1− λ2 −1
π 1− λ2 π 1− y2  2 

8.4-1
Binomial distribution with α = (1 - α)= ½, so m = 10 × ½ = 5, σ2 = 5 × ½ = 2.5, m ± 2σ ≈ 2 to 8

10  10  10   1 

10
1 + 10 + 45
P(i < 3) = FI (2) =   +   +  2   2  = 1024 = 0.0547
 0   1     

8.4-2
Binomial distribution with α = 3/5 and (1 - α)= 2/5, so m = 10 × 3/5 = 6, σ2 = 5 × 2/5 = 2.4,
m ± 2σ ≈ 3 to 9

10  10   10   28 (1× 4 + 10 × 6 + 45 × 9)256

P(i < 3) = FI (2) =   3022 +   3121 +   3220  10 = = 0.0122
 0  1 2 5 9.87 × 106

8-11
8.4-3
Let I = number of forward steps, binomial distribution with mI = 100 × ¾ = 75, σI2 =75 × ¼,

I 2 = 75 4 + 752 , X = Il – (100 – I)l = (2I – 100)l so mX = (2mI - 100)l = 50l and

X 2 = (2 I − 100) 2l 2 = (4 I 2 − 400m I + 10 4 )l 2 = 2575l 2 , σ X = 2575l 2 − (50l ) 2 = 75l

8.4-4
Binomial distribution with 1 - α = 0.99 so
 10   10 
P( I > 1) = 1 − PI (0) − PI (1) = 1 −   0.010.99
0 10
−  1  0.010.99 = 0.0042
1 9

 0  
Poisson approximation with m = 10 × 0.01 = 0.1

(0.1)0 (0.1) 1
P( I > 1) ≈ 1 − e −0.1 − e −0.1 = 0.0047
0! 1!

8.4-5
µ = 0.5 particles/sec, T = 2 sec, µT = 1, so

11−1 10 −1 1
−1
1
(a) PI (1) = e = 0.368 (b) P( I > 1) = 1 − PI (0) − PI (1) = 1 − e −e = 0.264
1! 0! 1!

8.4-6

mi mi mi
E [ I ] = ∑ i=0 ie − m = e − m ∑ i =0 i , E  I 2  =e− m ∑ i= 0 i 2
∞ ∞ ∞
, where
i! i! i!
1 2 ∞ m
i
d m d ∞ m
i
mi−1 1 ∞ mi
em = 1+ m + m + L = ∑ i= 0 so e = ∑ = ∑ i ! = m ∑ i= 0 i i! and
i
2! i! dm dm i =0 i !
d2 m d mi −1 mi− 2 1 mi
∑ ∑ ∑
∞ ∞
e = i = i (i − 1) = ( i 2
− i )
dm2 dm i= 0 i ! i! m2 i= 0 i!
(cont.)

d m d2 m mi mi mi
∑ i =0 i ∑ i =0 (i2 − i) = ∑ i= 0 i 2
∞ ∞ ∞
But e = 2
e = em so = mem and − mem = m2 e m
dm dm i! i! i!
Thus, E[I] = e-m(mem) = m and E[I2 ] = e-m(m2 em + mem) = m2 + m

8-12
8.4-7
2
X = m = 100, σ2 = X 2 − X ⇒ X 2 = σ 2 + m 2 = 10,004
P( X < m − σ or X > m + σ) = P ( X ≤ m − σ ) + P ( X > m + σ ) = 2 Q (1) ≈ 0.32

8.4-8
2
m = X = 2, σ = X 2 − X = 3

P( X > 5) = P( X > m + σ) = Q (1) ≈ 0.16 ,

1
P(2 < X ≤ 5) = P ( X > m) − P( X > m + σ) = − Q (1) ≈ 0.34
2

8.4-9

m = 10, σ = 500 −100 = 20, P( X > 20) = P( X > m + σ / 2 ) = Q (0.5) ≈ 0.31

P(10 < X ≤ 20) = P( X > m) − P( X > m + σ /2) = 1 / 2 − Q (0.5) ≈ 0.19

P(0 < X ≤ 20) = P( X − m < σ / 2 ) =1 − 2 Q(0.5) ≈ 0.38

P( X > 0) = 1 − P ( X ≤ m − σ /2) = 1 − Q (0.5) ≈ 0.69

8.4-10
m = 100 × ½ = 50, σ2 = 50 × ½ = 25, σ = 5
(a) P( X > 70) = P( X > m + 4σ ) = Q(4) ≈ 3.5 × 10−5

(b) P(40 < X ≤ 60) = P( X − m ≤ 2σ ) =1 − 2 Q(2) = 0.95

8.4-11
Let a = m – k 1 σ and b = m + k 2 σ so

 m−a  m−b 
P( a < X ≤ b) = 1 − Q ( k1) − Q ( k2 ) = 1 − Q   − Q 
 σ   σ 
Q(k 1 ) Q(k 2 )

a m b

8-13
8.4-12

m = 0, σ = 3, P( X ≤ c) = P  X − m ≤ σ  = 1 − 2Q  
c c
 σ  σ
(a) 1 – 2Q(c/3) = 0.9 ⇒ Q(c/3) = 0.05, c ≈ 3 × 1.65 = 4.95
(b) 1 – 2Q(c/3) = 0.99 ⇒ Q(c/3) = 0.005, c ≈ 3 × 2.57 = 7.71

8.4-13

1  1 −λ 2 / 2 ∞ ∞ −λ2 / 2  1  
 1
( )
1 ∞ 1 ∞
∫ e−λ ∫ k ∫k
2
−λ 2 / 2
Q( k ) = /2
dλ = − λ d e = − e − e d  − 
2π k
2π k
  2π  λ  λ 
1 1 ∞ 1 − λ2 / 2
= e−k − ∫ dλ
2
/2
e
2π λ
2
2π k 2 k

1
so Q ( k ) < e− k
2
/2
and
2 πk 2

1 ∞ 1 −λ 2 / 2 1 1 ∞ 1
∫ dλ < ∫ e−λ dλ =
2
/2
e Q( k ) = Q( k ) i f k ? 1
2π k λ 2
2π k 2 k k2
1
Thus, Q ( k ) ≈ e−k
2
/2
for k ? 1
2π k 2

8.4-14

1 ∞ (2σ )
2 n/2
∞
E  ( X − m)  = ∫ ∫
−( x − m ) 2 / 2 σ 2
(x − m) e dx = λ ne −λ d λ
2
n n

2πσ 2 −∞
π −∞

E  ( X − m) n  = 0 for odd n since λn e−λ has odd symmetry

2
(a)

(b) E  ( X − m)  = K n 2 ∫ λ e
n
∞
n −λ 2
d λ for even n, where Kn =
(2σ ) 2 n/2

=
2 n / 2 σn
0
π π

But e −λ d λ = − 12λ d e−λ

2

( ) so 2

 2 ∞ 
( )
− ∫0 e−λ d ( λn−1 ) = K n (n − 1) ∫ λn −2e −λ d λ
∞ ∞ ∞
E  ( X − m) n  = −K n ∫0 λn −1 d e−λ = − Kn λ n−1e −λ
2 2 2

 0  0

Kn ∞
= (n − 1) K n −2 2∫ λ n− 2e −λ d λ = ( n − 1) σ2 E ( X − m) n−2 
2

2K n − 2 0

(cont.)

8-14
Thus, E  ( X − m ) 4  = (4 − 1)σ2 E  (X − m ) 2  = 3σ4 , E  ( X − m ) 6  = (6 − 1)σ2 ( 3σ4 ) = 3⋅ 5σ 6 , and

E  ( X − m ) n  = 1⋅ 3 ⋅ 5L (n − 1)σn , n = 2,4,6, K

8.4-15
1 − π[( f − m)/ b]2
pX ( f ) = e where b = 2πσ 2
b

If m = 0, Φ X (2πt ) = e −π( bt) = e−σ (2 πt )2/ 2

, so Φ X ( ν ) = e− σ ν / 2
2 2 2 2

For m ≠ 0, use frequency-translation theorem with ωc = 2πm, so

Φ X (2πt ) = e −π( bt) e jωct = e −σ (2 πt) 2 / 2 jm(2 πt )

and Φ X (ν ) = e − σ ν / 2e jmν
2 2 2 2
e

8.4-16
2 2 2 2 2 2
Φ Z ( ν ) = Φ X (ν ) ΦY ( ν ) = e−σX ν /2
e jmX ν e−σY ν / 2 jmY ν
e = e−σZ ν / 2 e jmZ ν
1 −[ z − ( mX + mY ) ] /2( σ X 2 +σ Y 2 )
2

where σ Z = σX + σY , mZ = mX + mY . Hence, p Z ( z ) =
2 2 2
e
2π(σ X + σY )
2 2

1 n X
If Z =
n
∑ i =1
X i = Y1 + Y2 + L + Yn where Yi = i is gaussian with
n

σ Xi
2 2
X i 2 X i2 X i2 − X i
Yi = , Yi = 2 , σYi = =
2

n n n2 n2
2 2
Then Φ Z ( ν ) = Φ Y1 ( ν) Φ Y2 ( ν )L Φ Yn ( ν) = e− σ Z ν / 2e jmZ ν where

1 1 n
σZ = ∑ σYi = ∑ σX i , mZ = ∑ mYi = ∑ mX
2 2 n 2
i =1
n i=1 i
2
n

8.4-17
X = ln Y ⇒ Y = eX

E [Y ] = E  e X  = E  e jνX 
2
= Φ X ( − j ) = eσX / 2 mX
e
jν = 1

E Y 2  = E  e 2 X  = E  e jνX 
2
= Φ X ( − j 2) = e 2σX e2 mX
jν = 2

8-15
8.4-18
 1 

dx = (1 − j 2σ2 ν )
∞ 2 ∞ −  2 − jν x2 −1 / 2
(a) Φ Z ( ν ) = ∫ e jνx 2
pX ( x ) dx = ∫ e  2σ 
−∞
2πσ 2 0

dΦZ
= jσ 2 (1 − j 2σ2ν )
−3 / 2
(b) ⇒ E [ Z ] = j −1( j σ2 ) = σ2
dν
d 2Φ Z
( )
−5 / 2
= −3 σ 4
1 − j 2 σ 2
ν ⇒ E  Z 2  = j −2 ( −3σ4 ) = 3σ4
d ν2
d 3Φ Z
( )
−7 / 2
= − j15σ 6
1− j 2 σ 2
ν ⇒ E  Z 3  = j −3 (− j15σ6 ) =15 σ6
d ν3
Thus, E  X 2  = E [ Z ] = σ2 , E  X 4  = E  Z 2  = 3σ 4 , E  X 6  = E  Z 3  = 15σ6

8.4-19

R 2 = 2σ 2 = 32 ⇒ σ2 = 16 so p R ( r ) =
r − r2 /32
16
e u ( r ) and P( R ≤ r ) = FR ( r ) = 1 − e
− r 2 /32
(
u (r ) )
Thus, P( R > 6) = 1 − P (R ≤ 6) = e−6 = 0.325 and
2
/32

P(4.5 < R < 5.5) = P( R ≤ 5.5) − P (R ≤ 4.5) = e −4.5 / 32 − e−5.5 / 3 2 = 0.143

2 2

8.4-20

X 2 = 2σ 2 = 18 ⇒ σ2 = 9 so p X ( x) =
x − x2 /18
9
e u ( x ) and P( X ≤ x ) = 1− e (
− x 2 /18
u( x) )
Thus, P( X < 3) = P ( X ≤ 3) = 1 − e −3 = 0.393 , P( X > 4) = 1 − P( X ≤ 4) = e−4 = 0.411 , and
2 2
/18 /18

P(3 < X ≤ 4) = P (X > 3) − P( X > 4) = [ 1− P( X ≤ 3) ] − [1− P ( X ≤ 4) ] = e −3 − e−4 = 0.195

2 2
/18 /18

8.4-21

Since Z D 0 and X D 0, monotonic transformation with g(x) = x 2 , g-1 (z) = + z , dg −1 / dz = 1

2 z

x − x 2 / 2σ 2
p X ( x) = e u ( x ), m = Z = E  X 2  = 2σ2 . Thus
σ
2

e( ) ( ) 2 1 z = m1 e
2
z − z / 2σ 2 −z / m
p Z (z) = u + z u( z)
σ 2

km 1 − z/m  0.632 k = 1
P( Z ≤ km) = ∫ e dz = 1 − e − k = 
0 m  0.095 k = 0.1

8-16
8.4-22

(a) A = R12 = X 2 +Y 2 where X and Y are gaussian with X = Y = 0, σ X 2 = σY 2 = σ2

2
Φ A ( ν ) = Φ X 2 ( ν )Φ Y 2 ( ν ) =  (1− j 2σ2 ν) −1 / 2  = (1 − j 2σ2 ν )−1

1 b 1 1
Φ A (2πt ) = = , b = 2 so p A ( a ) = be− bau (a ) = 2 e − a / 2σ u( a )
2

1 + j 2σ (−2πt ) b + j 2π( −t )
2
2σ 2σ

(b) pW ( w) = p R 2 (w ) * pR 2 ( w), pR 2 = pR 2 = p A
1 2 1 2

0 w<0

=  w 1 −λ / 2σ2 1 − ( w−λ) / 2σ2 2
 1  − w / 2 σ2 w
 ∫0 2σ2 e 2σ2
e dλ =  2  e ∫0 d λ w > 0
  2σ 
w − w / 2 σ2
so pW ( w) = e u ( w)
4σ
4

8.4-23
1 a − a 2 ( x2 − 2 ρxy +y 2 )
Let a2 = so p ( x , y ) = e
2σ2 (1 − ρ2 )
XY
2 π2 σ2

e− a e− y / 2σ 2
2
y2 2
∞ a ∞ ∞
pY ( y ) = ∫ pXY ( x , y) dx = e −a 2 y 2
∫ e − a 2 ( x 2 −2 ρxy)
dx = a 2 ρ2 y2
e 2∫ e − λ2
dλ =
−∞ −∞
2 π2 σ 2 2π 2 σ 2 0
2 πσ2
pXY ( x , y) 1
= e − ( x −ρy ) / 2σ (1−ρ )
2 2 2
pX(x|y) =
pY ( y ) 2 πσ2 (1 − ρ2 )

8.4-24
Since Z is a linear combination of gaussian RVs, pZ(z) is a gaussian PDF with

mZ = E [ X + 3Y ] = mX + 3mY = 0

σZ = E  X 2 + 6 XY + 9Y 2  = (σ X + m X ) + 6 E [ XY ] + 9(σY + mY ) = 100
2 2 2 2 2

8.4-25
∞
E  X n  = ∫ x n p X ( x ) dx = 0 for n odd, E [Y ] = E  X 2  = σX 2
−∞

E [ ( X − mX )(Y − mY )] = E  X ( X 2 − σ X 2 )  = E  X 3  − σX 2 E [ X ] = 0 ⇒ ρ = 0

8-17