Chapter 5

PROBABILITY DENSITIES

5.1

{ 2e-2x for x > 0

f(x) = 0 elsewhere

Since e-2x is always positive, f (x) is always ~ o.

100 f(x)dx = _e-2xI00 = 1

-00 0

Thus, f(x) is a density.

5.2 To find k, we must integrate f(x) from x = 0 to x = 1 and set it equal to 1. Thus,

which implies k/4 = 1. Thus, k = 4.

(a) P(.25:S X :S .75) = II: 4x3dx = x41:;~ = 80/256 (b) P(X> 2/3) = fi/34x3 = x41~/3 = 65/81

123

124

Chapter 5 PROBABILITY DENSITIES

5.3 The distribution function is given by

F(x) = 1:00 f(s)ds = X4

(a) P(X> .8) = 1 - F(.8) = .5904

(b) P(.2 < X < .4) = F(.4) - F(.2) = .024

5.4 (a) Let X be a random variable with density f(x). Then,

f·8 f·8 18

P(.2 < X < .8) = f(x)dx = xdx = x2/2' = (.64 - .04)/2 = .30

.2 .2 .2

(b)

11.2 11 r1.2

P(.6 < X < 1.2) - .6 f(x)dx = .6 xdx + J1 (2 - x)dx

_ x2/211 + (2x - x2/2)I1.2 = .32 + .18 = .50

.6 1

5.5

F(x)

{ ~2 /2 ~ ~ ~ < 1

- 1/2+[2s-s2/2]1~ l<x:S2

1 x > 2

x<O O:Sx:Sl l<x:S2 x>2

(a) P(X > 1.8) = 1 - F(1.8) = 1 - [2(1.8) - (1.8)2/2 - 1] = 1 - .98 = .02

(b) P(.4 < X < 1.6) = F(1.6) - F(.4) = 2(1.6) - (1.6)2/2 - 1 - (.4)2/2 = .84

5.6 We need to integrate f(x) from x = -00 to x = 00 and set it equal to 1.

i: k/(l + x2)dx - k i: 1/(1 + x2)dx = k· arctanxl~: - k(7f/2 + 7r/2) = kat = 1

125

Thus, k = 1/7r.

5.7 Let X have distribution F(x). Then,

(a) P(X < 3) = F(3) = 1 - 4/9 = 5/9 = .556

(b) P(4::; X ::; 5) = F(5) - F(4) = 4/16 - 4/25 = .09

5.8 The density f(x) is given by f(x) = d:F(x). Thus, the density is

5.9 (a) P(O::; error j; 7r/4) = J;/4cosxdx = sinxl~/4 = sin(7r/4) = .;2/2

(b) P(phase error> 7r/3) = J:f; cosxdx = sin(7r/2) - sin(7r/3) = 1 - V3/2 = .1339

5.10 (a) P(no. of miles tires last::; 10,000 miles)

1 !olD io

= - e'-x/20dx = _e-x/2°1 = 1 - e-1/2 = .3935

20 0 0

(b) P(16,000::; no. of miles tires last::; 24,000)

1 /24

= _ e-x/20dx = e-16/20 _ e-24/20 = .1481

20 16

(c) P(no. of miles tires last ~ 30,000)

1 hoo

= _ e-x/2°dx = e-30/20 = .2231

20 30

5.11 Integrating the density function by parts shows that the distribution function is given by

1

F(x) = 1 - _xe-x/3 - e-x/3

3

126

Chapter 5 PROBABILITY DENSITIES

Thus,

P(power supply will be inadequate on any given day) P(consumption 2: 12 million kwh's)

- 1 - F(12) = 4e-4 + e-4 = 5e-4 = .0916

5.12 By the definition of variance

(J2 - I: (x - M)2 f(x)dx

I: x2 f(x)dx - 2M I: xf(x)dx + M2 I: f(x)dx

Using the fact that

100 I

-00 x2 f(x)dx = M2 ,

I: xf(x)dx = M and I: f(x)dx = 1

we have

5.13 The density is

f(x) = {40X3 0 < X < 1 elsewhere

Thus,

M = fol 4x4dx = 4x5 /51~ = 4/5

I r1 5 6 11

M2 = 10 4x dx = 4x /6 0 = 2/

and the variance is

5.14 In this case,

102 xf(x)dx = 101 x2dx + ,12 x(2 - x)dx

x3 / 31 ~ + x21: - x3 / 3 1 ~ = 1/3 + 4 - 1 - 8/3 + 1/3

and

/-l~ - 102 X2 f(x)dx = 101 x3dx + 12 x2(2 - x)dx

_ x4/41~ + 2x3/31: - X4/41:

1/4 + 16/3 - 2/3 - 16/4 + 1/4 7/6

Thus,

2 I 2 / 2 /

a = f.-l2 - f.-l = 7 6 - 1 = 1 6

5.15 The density is:

f(x) = {80X-3 X > 2 x::=;2

Thus,

and

I roo 2 -3 00

f.-l2 = } 2 X (8x ) dx = 8ln x 12 = 00

Thus, a2 does not exist.

5.16 The density is

f(x) = {coosx 0 < x < 7r/2 elsewhere

127

1

128

Cbapter 5 PROBABILITY DENSITIES

Thus,

r/2

Jo xcosxdx =

7f /2 + cos XI~/2

. 17r/2

xsmx °

r=

Jo sinxdx

= 7f/2 - 1

and

I r: 17r/2 r/2

fJ,2 - Jo x2 cosxdx = x2 sinx ° - 2 Jo xsinxdx

- x2 sin xl:/2 - 2 [ - x cos XI~/2 + fo7r/2 cos xdx 1

- 7f2/4 - 2[1 - 0] = 7f2/4 - 2

Thus,

5.17 The density is:

{ (1/20)e-x/2o x > 0

f(x) = 0 x ::; 0

Thus,

1 100

fJ, = - xe-x/20 dx

20 °

Integrating by parts gives:

fJ, - _xe-x/2°1: + L~) e-x/2odx _ 0 - 20e-X/20I:

- 20 (thousand miles)

129

5.18 On Ixl > 1, we have x2/(1 + x2) > 1. So

I 1 100 x2 2100 x2 2 1,00

I-l =- dx=- dx>- dx

2 1f -00 1 + x2 1f 0 1 + x2 1f 1

and this integration term is infinite. Thus I-l~ is infinite and the variance does not exist.

5.19 (a) P(less than 1.50) = F(1.50) = .9332

(b) P(less than - 1.20) = F( -1.20) = .1151

(c) P(greater than 2.16) = 1 - F(2.16) = 1 - .9846 = .0154 (d) P(greater than - 1.75) = F(1.75) = .9599

-3 -2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.19 (a) z

5.19 (b) z

0.0154

-3 -2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.19 (c) z

5.19 (d) z

5.20 Let Z be a random variable having a standard normal distribution.

(a) P(O < Z < 2.7) = F(2.7) - F(O) = .9965 - .50 = .4965

130

Chapter 5 PROBABILITY DENSITIES

(b) P(1.22 < Z < 2.43) = F(2.43) - F(1.22) = .9925 - .8888 = .1037

(c) P(-1.35 < Z < -.35) = F(-.35) - F(-1.35) = .3632 - .0885 = .2747 (d) P( -1.70 < Z < 1.35) = F(1.35) - F( -1.70) = .9115 - .0446 = .8669

-3 ·2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.20 (a) z

5.20 (b) z

-3 -2 -1 0 1 2 3

-3 -2 ·1 0 1 2 3

5.20 (c) z

5.20 (d) z

5.21 (a) P(Z:::; z) = F(z) = .9911. Thus z = 2.37

(b) P(Z > z) = .1093. That is, P(Z :::; z) = 1 - .1093 or F(z) = .8907. Thus, z = 1.23

(c) P(Z > z) = .6443. That is, F(z) = 1 - .6443 = .3557. Using Table 3, z = -.37

(d) P(Z < z) = .0217 so z is negative. From Table 3, z = -2.02.

(e) P( -z :::; Z :::; z) = .9298. That is, F(z) - F( -z) = .9298, which implies that F(z) - (1 - F(z)) = .9298 or F(z) = (1 + .9298)/2 = .9649. By Table 3, z = 1.81.

131

-3 -2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.21 (a) z

5.21 (b) z

0.0217

-3 -2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.21 (c) z

5.21 (d) z

-3 -2 -1 0 1 2 3

5.21 (e) z

5.22 Let X be a random variable having distribution Nt p; a2) ( normal with mean p and variance a2 ). Then

(a) P(IX - pi < a) = P(IX - pl/a < 1) = P( -1 < (X - p)/a < 1) = P(l) - 1 + F(l) = 2F(1) - 1 = .6826

(b) P(IX - pi < 2a) = 2F(2) - 1 = .9544

(c) P(IX - pi < 3a) = 2F(3) - 1 = .9974

132

Chapter 5 PROBABILITY DENSITIES

(d) P(IX - J-il < 40-) = 2F(4) - 1 = 2(.99997) - 1 = .99994

5.23 (a) P(Z> Z.005) = .005. Thus, F(Z.005) = .995 and Z = 2.575 by linear interpolation in the Table 3.

(b) P(Z> Z.025) = .025. Thus, F(Z.025) = .975 and Z = 1.96

5.24 Let X have distribution N(16.2, 1.5625).

(a) P(X> 16.8) = 1 - F((16.8 - 16.2)/1.25) = 1 - F(.48) = 1 - .6844 = .3156 (b) P(X < 14.9) = F((14.9 - 16.2)/1.25) = F( -1.04) = .1492

(c) P(13.6 < X < 18.8) = F((18.8 - 16.2)/1.25) -F((13.6 - 16.2)/1.25) = F(2.08) - F( -2.08) = .9812 - .0188 = .9624

(d) P(16.5 < X < 16.7) = F((16.7 - 16.2)/1.25) -F((16.5 - 16.2)/1.25) = F(.4) - F(.24) = .6554 - .5948 = .0606

12.4 15.0 17.4 20.0

-3 -2 -1 0 1 2 3

5.24 (a) x

z

12.4 15.0 17.4 20.0

-3 -2 -1 0 1 2 3

5.24 (b) x

z

133

12.4 15.0 17.4 20.0

-3 -2 -1 0 1 2 3

5.24 (c) x

z

12.4 15.0 17.4 20.0

-3 -2 -1 0 1 2 3

5.24 (d) x

z

5.25

x - 30 9.2

P[X > 39.2] = .20 so P[ > -] = .20

a a

That is, 1 - F(9.2/a) = .20, and F(9.2/a) = .80. But F(.842) = .80. Thus 9.2/ a = .842, so a = 10.93.

5.26 From Table 3, P((X - J.l)/1O < .92) = .8212. Thus, given

P[X < 282.5] = .8212

or P[X - J.l > 282.5 - J.l] = .8212

10 10

we maut have (282.5 - J.l)/10 = .92 or J.l = 282.5 - 9.2 = 273.3 To find P(X > 258.3), we need to find

1 - F((258.3 - 273.3)/10) = 1 - F( -15/10) = F(1.5)

134

Chapter 5 PROBABILITY DENSITIES From Table 3, F(1.5) = .9332. Thus, P(X > 258~3) = .9332

5.27 (a) We need to find P(X > 11.5), where X is normally distributed with J-L = 12.9 and a = 2.

P(X ~ 11.5)

1 - F((l1.5 - 12.9)/2) - F(.7) = .7580

F((12.9 - 11.5)/2)

(b)

P(ll :S X :S 14.8) - F( (14.8 - 12.9) /2) - F( (11 - 12.9) /2)

F(.95) - F( -.95) = .8289 - .1711 = .6578

6.9 10.9 14.9 18.9

-3 -2 -1 0 1 2 3

5.27 (a) x

z

6.9 10.9 14.9 18.9

-3 -2 -1 0 1 2 3

5.27 (b) x

z

5.28 P(Z:S -ZO.25) = .25. Thus, -ZO.25 = -.675

135

P(Z:S ZO.50) = F(ZO.50) = .50. Thus, ZO.50 = 0 P(Z:S ZO.25) = F(ZO.25) = .75. Thus, ZO.25 = .675

5.29 Let X be a random variable representing the developing time which is normally distributed with f-l = 16.28 and a = .12.

(a) P(16:S X :S 16.5) = F((16.5 - 16.28)/.12) - F((16 - 16.28)/.12) = F(1.833) - F( -2.333) = .9666 - .0098 = .9568

(These values are determined by interpolation)

(b) P(X ~ 16.20) = 1 - F((16.20 - 16.28)/.12) = 1 - F( -.667) = F(.667) = .7476

(c) P(X:S 16.35) = F((16.35 - 16.28)/.12) = F(.5833) = .7201

15.9 16.2 16.4 16.6

·3 -2 -1 0 1 2 3

5.29 (a) x

z

15.9 16.2 16.4 16.6

-3 -2 -1 0 1 2 3

5.29 (b) x

z

136

Chapter 5 PROBABILITY DENSITIES

15.9 16.2 16.4 16.6

-3 -2 -1 0 1 2 3

5.29 (c) x

z

5.30 From Table 3, we know that F(1.645) = .95. We need to find a value x such that

P(X > x)

P((X - 16.28)/.12 > (x - 16.28)/.12)

- 1 - P((X - 16.28)/.12 < (x - 16.28)/.12)

1 - F((x - 16.28)/.12) = F(( 16.28 - x)/.12) - .95

Thus, (16.28 - x)/.12 = 1.645, and x = 16.0826

5.31 P(.295::; X::; .305) = F((.305 - .302)/.003) - F((.295 - .302)/.003) = F(l) - F( -2.333) = .8413 - .0098 = .8315

Thus, 83.15 percent will meet specifications.

0.3 0.3 0.3 0.3

-3 -2 -1 0 1 2 3

5.31 x

z

5.32 We must find /1 such that F((4 - /1)/'025) = .02 or F((/1- 4)/.025) = .98. But, F(2.05) = .98. Thus, (/-l- 4)/.025 = 2.05 or /-l = 4.05.

137

5.33 We need to find 11 such that F((3 - 11);'01) _: .95. Thus, from Table 3, (3 - 11)/.01 = 1.645 or 11 = 2.98355.

5.34 (a)

P(5.9 < X :::; 6.1)

F((6.1 - 6);'06) - F((5.9 - 6);'06)

- F(1.667) - F( -1.667) = .9522 - .0478 = .9044

The proportion of rods exceeding the tolerance limits is 1 - .9044 = .0956

5.8 5.9 6.1 6.2

-3 -2 -1 0 1 2 3

5.34 x

z

(b) If 99% of the rods must' be within tolerance, a should satisfy 2F(.I/a) - 1 = .99, that is F(.I/a) = .995. Hence .1/a = 2.575 or a = 0.0388

5.35 If n = 30 and p = .60 then 11 = 30(.60) = 18 and a2 = 30(.6)(04) = 7.2 or a = 2.6833.

(a) P(14) = F((14.5 - 18)/2.6833) - F((13.5 - 18)/2.6833) = F(-1.304) - F( -1.677) = .0961 - .0468 = .0493

(b) P(less than 12) = F((I1.5 -18)/2.6833) = F(-2042) = .0078

5.36 In this case, n = 1200, p = .02, 11 = 24, a2 = 23.52, a = 4.8497. Thus,

P( at least 30 need repairs)

1 - F((29.5 - 24)/4.8497)

1 - F(1.13) = 1 - .8707 = .1292

138

Chapter 5 PROBABILITY DENSITIES

5.37 In this case, n = 200, p = .25, 11 = 50, a2 = 37.5, a = 6.1237. Thus,

P(fewer than 45 fail) - F((44.5 - 50)/6.1237) F( -.90) = .1841

5.38 In this case, n = 84, p = .3, 11 = 25.2, a2 = 17.64, a = 4.2.

F((30.5 - 25.2)/4.2) - F((19.5 - 25.2)/4.2) - F(1.26) - F( -1.36)

- .8962 - .0869 = .8093

5.39 Again, we will use the normal approximation to the binomial distribution. Here, n = 40, p = .62, 11 = 24.8, a2 = 9.424, a = 3.0699. Thus,

F((20.5 - 24.8)/3.0699) = F( -1.40) = .0808

5.40 (a) We will use the normal approximation to the binomial with n = 1,000, p = .5, 11 = 500, a2 = 250, a = 15.81. The proportion from .49 to .51 means the actual number is from 490 to 510. Thus,

F((510.5 - 500)/15.81) - F((489.5 - 500)/15.81)

= F(.664) - F( -.664) - = .7467 - .2533 = .4934

(b) Similar to part (a), with parameters n = 10,000, p = .5, 11 = 5,000, a2 = 2500, a = 50. Thus,

F((5100.5 - 5000)/50) - F((4899.5 - 5000)/50)

- F(2.01) - F( -2.01) = .9778 - .0222 = .9556

139

5.41 Let f(x) be the standard normal density. Then F(-z) = J~:Of(x)dx. Using the change of variable, s = -x, and the fact that f(x) = f( -x), we have

F( -z) = - 1~ f( -s)ds = 100 f(s)ds = 1 - 1:00 f(s)ds = 1 - F(z)

5.42 To find the mean of the normal density, we need to find

Using the change of variable u = (x - f.-l)/a yields an odd integrand, u exp( -u2 /2) so the interal is O. Thus, the mean is u.

5.43 We need to find

since the integrand is an even function. Using the change of variable s = (x-f.-l)/a, the variance is equal to

2a2 1000

J7C S2 exp[-s2 /2]ds

V27r 0

Integrating by parts with u = sand dv = sexp[-s2/2]/(27r)1/2ds shows that the variance is equal to

The first term is zero and the second is 1/2 since this is an integration of half of

140

Chapter 5 PROBABILITY DENSITIES

the standard normal density. Thus the variance is (J2.

5.44 (a) Normal with mean = 11.3000 and standard deviation = 5.70000

x

PC X <= x) 0.3112

8.4930

(b) Normal with mean = 11.3000 and standard deviation = 5.70000

x

PC X <= x) 0.7989

16.0740

5.45 The uniform density is:

f(x) = { 01/(,8 - a) a < x < ,8 elsewhere

Thus, the distribution function is

{I x?,8

F(x) = O(X - a)/(,8 - a) a < x < ,8 xS;a

5.46 (a) P(.010 S; error S; .015) = (.015 - .010)/.050 = .1 (b) P( -.012:S error :S .012) = (.012 + .012)/.050 = .48

5.47 Suppose Mr. Harris bids (1 + x)c. Then his expected profit is:

OP(low bid < (1 + x)c) + xcP(low bid ? (1 + x)c)

12c 3

= xc -ds = 3xc[2c - (1 + x)c]/4c = 3c(x - x2)/4

(l+x)c 4c

Thus, his profit is maximum when x = 1/2. So his bid is 3/2 times his cost. Thus, he adds 50 percent to his cost estimate.

141

5.48

= ~ {'Xl exp[-(lnx _ a)2/(2tJ2)]dx y 27ftJ 10

Using the change of variable y = In x gives:

5.49 10/ I, is distributed log-normal with a = 2, tJ2 = .01, tJ = .1. Thus,

P(7::; Io/Ii ::; 7.5) - F((ln(7.5) - 2)/.1) - F((ln(7) - 2)/.1) - F(.149) - F( -.54)

- .5592 - .2946 = .2646

5.50 Using the formulas for J-l and (J2 for the log-normal distribution with a = -1 and

142

Chapter 5 PROBABILITY DENSITIES

{3 = 2, gives:

f1 exp]c + {32/2] = exp[+I + 4/2] = e1 = e = 2.718

a2 e2a+,62 [e,62 - 1] = e-2+4[e4 - 1] = e2[e4 - 1] = e6 - e2 = 396

Thus, a = 19.9

5.51 (a) P(between 3.2 and 8.4) = F((ln(8.4) + 1)/2) - F((ln(3.2) + 1)/2) = F(1.564) - F(1.0816) = .9406 - .8599 = .0807

(b) P(greater than 5) = 1 - F((ln(5) + 1)/2) = 1 - F(1.305) = 1 - .904 = .0960

5.52 Using the formulas for f1 and a2 for the gamma distribution with a = 2 and {3 = 2, gives f1 = a{3 = 4 and, a2 = a{32 = 8. Thus, a = 2.8284.

5.53 When a = 2 and {3 = 2, f(2) = 1. So

{ xe-x/2/4 x » 0

j(x) = 0 x S; 0

Thus,

104 1104

P(X < 4) = j(x)dx = - xe-x/2dx

o 4 0

Integrating by parts gives

1 1 104

__ xe-x/2 14 + - e-x/2dx

2 0 2 0 -

2 -2 -x/2 14 e - e 0

1 - 3e-2 = .5940

5.54 The density of the gamma distribution with a = 3, {3 = 2 is

143

Thus,

1 j,=

P(power inadequate) = - x2e-x/2dx

16 12

Integrating by parts gives

Integrating the second term by parts gives

5.55 (a) The probability that the supports will survive, if f.1 = 3.0 and (j2 = .09, is

P (su pports will survive)

1 - F (In(33) - 3.0) = 1 - F(1.655) .30

1- - .9508 = .0492

(b) If f.1 = 4.0 and (j2 = .36, then

P (supports will survive)

_ 1 _ F (In(33) - 4.0) .60

F(.84) = .7995

1 - F( -.84)

5.56

Using the change of variable y = x] f3 gives

144

Chapter 5 PROBABILITY DENSITIES

Thus,

5.57 We can ignore the constant in the density since it is always positive. Thus, we need to maximize f(x) = xa-1e-x/{3. Taking the derivative

Setting the derivative equal to zero gives the solution x = fJ( a-I). For a > 1, the derivative is positive for x < fJ( a-I) and negative for x > fJ( a-I). Thus, fJ(a - 1) is a maximum. Note that x = 0 is a point of inflection when a > 2. When a = 1, f(x) = e-x/(3 which has a maximum in the interval [0,00] at x = O. When 0 < a < 1, the derivative does not vanish on (0,00) and f(x) is unbounded as x decreases to O.

5.58 The exponential density is

f(x) = { e-x/{3/fJ x » O,fJ > 0

o elsewhere

and the exponential distribution is

F(x) = fox f(s)ds = 1 - e-x/{3 for x > O.

(a) P(X :::; 20) = 1 - e-20/50 = 1 - e-2/5 = .3297 (b) P(X 2 60) = 1 - (1 - e-60/50) = e-6/5 = .3012

5.59 Since the number of breakdowns is a Poisson random variable with parameter ). = .3, the interval between breakdowns is an exponential random variable with parameter). = .3.

145

(a) The probability that the interval is less than 1 week is 1 - e-(·3)1 = .259 or 25.9 percent.

(b) The probability that the interval is greater than 5 weeks is e-(·3)5 = .223 or 22.3 percent.

5.60 Assuming that the time between calls is distributed as an exponential random variable with ). = .6, the probability that there are no arrivals in an interval of length t is «:" = e-·6t.

5.61 Let N be a random variable having the Poisson distribution with parameter ot:

Then P(N = 0) = (at)Oe-at /O! = e-at. Thus, P(waiting time is > t) = e-at and P(waiting time is :::; t) = 1- e-at.

5.62 The density is given by the derivative of P ( waiting time is :::; t ). Thus, the density is ae-at.

5.63 The beta density is

f( ) = r(a + (3) a-1(1._ ),8-1

x r(a)r(f3)x x

for 0 < x < 1, a > 0, and f3 > O. For a = 3 and f3 = 3

Thus,

fo1 f(x)dx - 30 fo\x2-2x3+x4)dx 30(x3/3-x4/2+x5/5) I~

- 30(1/3 - 1/2 + 1/5) = 1

as required.

146

Chapter 5 PROBABILITY DENSITIES

5.64 When a = 2 and (3 = 9, the beta density if

T'(Ll ) 8 10! 8 8

r(2)r(9) x(1 - 1:) = 1 !8! T(1 - x) = 90x(1 - x)

Thus, the required probability is given by

90 l] (:r - 1 + 1)(1 - x)8dx

90[(1 - 10/10 - (1 - x)9/9) lin 90[(.9]0 - 1)/10 - (.99 - 1)/9] = .2639

5.65 (a) The mean of the beta distribution is given by f..t = a/(a + (3). Thus, in the case where 0 = 1 and (3 = 4, IJ = 1/(} + 4) = 1/5 = .2

(b) When a = 1 and (3 = 4, the beta density is

Thus, the required probability is given by

/1 1

4 (1 - lld:r = -(1 - X)4 125 = (.75)4 = .3164

.25 .

5.66 Since the coefficient r(a+(3)/r(o:)f((3) is always> 0, we need to find the maximum of f(x) = xO"-I(l - x)i3-I. When 0> J, ;3> 1, taking the derivative gives

/ (:r) - XO"-2(l- 1;)13-2((0 - 1)(1- x) - ((3 - I)X) XO"-2(l - :r)f:J-2((o - 1) - (0: + (3 - 2)x)

Thus, the derivative is 0 when 1: = (0 - ]) / (a + (3 - 2). It also equals 0 at x = 0 and x = 1 if Q > 2 and (3 > 2 respectivciv, Since f(O) = 0, f(l) = 0, f(x) ~ 0, and /(x) is continuous, x = (0 -1)/(0"'; (3 - 2) is a maximum.

147

5.67 Let X be Weibull random variable with a = .1, j3 = .5 representing the battery lifetime. Then the density is f(x) = (.1)(.5)x-·5e-·lx·5 for x > O. Thus,

t'" .5

P(X:::; 100) = 10 (.I)(.5)x-·5e-·lx dx

Using the change of variable y = x·5 gives:

rIO

P(X :::; 100) = .1 10 e-·lYdy = _e-·ly I~o = 1 - e-l = .6321

5.68 (a) The expected value of the Weibull random variable is given by

When a = 1/5 and j3 = 1/3,

/1 = (1/5)-3r(4) = 125·6 = 750

(b) The probability is given by

1 1 1300 _Xl/3 13001/3 1

- • - Xl/3-l exp(--)dx = -e-y/5dy

5 3 0 5 0 5

-3001/3

= 1 - exp( ) = 1 - e-1.3389 = .7379

5

5.69 The probability is

100 .~

(.025)(.500)x-·5e-(·025)X dx

4,000

JOO .025e-·025Ydy

yf4,OOO

_ e-·025v'4,000 = .2057

148

Chapter 5 PROBABILITY DENSITIES

5.70

Let y = ax(3. Then dy = a(3x(3-ldx. Thus,

Thus, a2 = a-2/(3 [I'(I + 2/(3) - (I'(I + 1/(3))2]

5.71 (a) The joint probability distribution of Xl and X2 is

where Xl = 0,1,2, X2 = 0,1, and 0 ::; Xl + X2 ::; 2. The joint probability distribution 1 (Xl, X2) can be summarized in the following table:

X2
I(XI, X2) 0 1 Total
0 .1 .2 .3
Xl 1 .4 .2 .6
2 .1 0 .1
Total .6 .4 1 (b) Let A be the event that Xl + X2 = 0 or 1, then

P(A) = 1(0,0) + 1(0, 1) + 1(1,0) = .1 +.2 +.4 = .7

149

(c) By (a), the marginal distribution of Xl is

11(0) - 1(0,0) + 1(0, 1) =.1 +.2 =.3
11(1) - 1(1,0) + 1(1,1) =.4 +.2 =.6
11(2) - 1(2,0) + 1(2, 1) = .1 + 0 = .1 (d) Since

12(0) = 1(0,0) + 1(1,0) + 1(2, 0) = .1 + .4 + .1 = .6,

the conditional probability distribution of Xl given X2 = 0 is

fr (010) 1(0,0) .1 1
- - - -
12(0) .6 6
I, (110) 1(1,0) .4 4
- - - -
12(0) .6 6
I, (210) 1(2,0) .1 1
- - -
12(0) .6 6 5.72 (a) The independent random variables Xl and X2 have the same probability distribution b(x; 2, .3). Hence the joint probability distribution of Xl and X2

is

where Xl = 0,1,2, and X2 = 0,1,2.

(b)

P(X1 < X2) - 1(0,1) + 1(0,2) + 1(1, 2)

150

Chapter 5 PROBABILITY DENSITIES

.2058 + .0441 + .0378 = .2877

5.73 (a) P(Xl < 1, X2 < 1) = F(l, 1)

(b) The probability that the sum is less than 1 is given by:

5.74

5.75 The joint distribution function is given by:

for 0 < Xl < 1 and 0 < X2 < 2. Thus, the distribution function is

151

o xixV4 xV4 xi

1

Xl :S 0 or X2 :S 0

o < Xl < 1 and 0 < X2 < 2 Xl ~ 1 and 0 < X2 < 2

o < Xl < 1 and X2 ~ 2

Xl ~ 1 and X2 ~ 2

The distribution function of Xl is

for 0 < Xl < 1.

Thus,

Xl :S 0

0< Xl < 1 Xl ~ 1

Similarly,

for 0 < X2 < 2.

Thus,

It is easy to see that FI(xd . F2(X2) = F(XI,X2)' Thus, the random variables are independent.

5.76 P(.2 < X < .5, .4 < Y < .6)

_ /.5 /.6 ~(x + y2)dydx

.2 .4 5

_ ~ /.5 (.2x + .152/3)dx 5 .2

152

Chapter 5 PROBABILITY DENSITIES

.04345

5.77 The joint distribution function is given by

t' fY 6 3x2y 2xy3

F(x, y) = io io S(u + v2)dvdu = -5- + -5-

for 0 < x < 1, 0 < y < 1

Thus, the joint distribution is

F(x,y) =

o

(3/5)x2y + (2/5)xy3 (3/5)y + (2/5)y3 (3/5)x2 + (2/5)x

1

x:::; 0 or y < 0

o < x < 1, 0 < y < 1 x 2: 1, 0 < y < 1

o < x < 1, y 2: 1

x 2: 1, y 2: 1

The probability of the region in the preceeding exercise is given by

F(.5, .6) - F(.2, .6) - F(.5,.4) + F(.2, .4). - .1332 - .03168 - .0728 + .01472 .04344

5.78 The marginal density for X is given by

The marginal density for Y is given by

(a) Thus,

P(X> .8) = J8\6X/5 + 2/5)dx = (3x2/5 + 2x/5) 1~8 = .296

153

(b)

P(Y < .5) = fo'5 (3/5 + 6y2/5)dy = (3/5)(.5) + (2/5)(.5)3 = .35

5.79 (a) By definition

I ( j ) = I(x, y) = { (x + y2)/(~ + y2) for 0 < y < 1, 0 < x < 1

1 X Y f2(y) 0

elsewhere.

(b) Thus,

_ x + .52 _ { 4x/3 + 1/3 for 0 < x < 1

11 (xj.5) - 1 52-

:2 + . 0 elsewhere.

( c) The mean is given by

5.80 (a) The joint density is:

o < Xl < 1, 0 < X2 < 1 elsewhere

In the example, it was shown that the conditional density of the first random variable given that the second takes on the value X2 is

(b) The marginal density of Xl is

154

Chapter 5 PROBABILITY DENSITIES

Thus,

for 0 < Xl < 1, 0 < X2 < 1.

5.81 (a) To find k, we must integrate the density and set it equal to 1. Thus,

= 101 k(2x + 2)dx = 3k = 1

Thus, k = 1/3.

(b) P( X < Y, Z > 1)

! rl r2joo(x+y)e-ZdzdydX = _!_ t' r2(x+y)dydx

3 io ix 1 3e io ix

_ _!_ rl (2x + 2 - 3x2/2)dx = 5/(6e) = .3066 3e io

5.82 (a)

f2 foo !(x + y)e-Zdzdy = 2(x + 1)/3 io io 3

_ fl foo !(x + y)e-Zdzdx = (y + 1/2)/3

io io 3

rl r2 !(x + y)e-Zdydx = e-z io io 3

Thus,

- !(y + 1/2) . ~(x + 1) . «=

3 3

2

g(xy + y + x/2 + 1/2)e-Z

155

1

=I 3(x + y)e-Z = f(x, y, z)

Thus, the three random variables are not independent.

(b)

f(x,y) f(x, z) f(y, z)

Since

1 2

f(x, y) = 3(x + y) =I ft(x)f2(y) = g(xy + y + x/2 + 1/2),

X and Yare not independent. Since

f(x, z) = ~(x+ 1)e-Z = fl(X)h(z),

X and Z are independent. Since

1

f(y, z) = 3(y + 1/2)e-Z = f2(y)h(z),

Y and Z are independent.

5.83 (a) Notice that f(XI, X2) can be factored into

Thus, Xl and X2 are independent normal random variables. Thus,

P(-8 < Xl < 14, -9 < X2 < 3) = P(-8 < x, < 14)P(-9 < X2 < 3)

= (F((14 - 2)/10) - F(( -8 - 2)/10)) (F((3 + 2)/10) - F(( -9 + 2)/10))

156

Chapter 5 PROBABILITY DENSITIES

= (F(1.2) - F( -1))(F(.5) - F( -.7)) = (.8849 - .1587)(.6915 - .2420)

= (.7262)(.4495) = .3264

(b) When /-il = /-i2 = 0 and (J = 3, the density is

Let R be the region between the two circles. We need to find

Changing to polar coordinates gives

5.84 Let R be the region for acceptable holes. Then

. 11

P(hole WIll acceptable) = - exp( -(xi + x~)/8)dxIdx2

R 87l'

Changing to polar coordinates gives

lr=SIo27r 1 2 loS 1 2 2 S

_e-r /Srd()dr = _e-r /Srdr = _e-r /S 10 = 1 - e-S

r=O 0 87l' 0 4

5.85 The expected value of g(XI' X2) is

j_: g(X1' x2)f(X1, X2)dx1dx2 - 101 102 (Xl + X2)X1X2dx2dx1 - 101 (2xi + 8Xd3)dx1 = 2

157

5.86 The expected value of g(X, Y) is given by

11 11 6 611 x3 x2 6 1 1 1 -x2Y(X+y2)dydx = - (- + -)dx = -(- + -) =-

o 0 5 5 0 2 4 5 8 12 4

5.87 The area of the rectangle is X . Y. Thus the mean of the area distribution is given

by

J,L+a/21w +b/2 xy J,L+a/2 X

-dydx = -Wdx = LW

L-a/2 W -b/2 ab L-a/2 a

The variance is given by

J,L+a/21w +b/2 x2y2

-dydx - (W L)2 L-a/2 W -b/2 ab

(W2 + b2/12)(L2 + a2/12) - (W L)2 112 ((aW)2 + (bL)2 + (ab)2/12)

5.88

Thus,

5.89 (a) E(XI + X2) = E(Xd + E(X2) = 1 + (-1) = O.

(b) Var(Xl + X2) = Var(X1) + Var(X2) = 5 + 5 = 10.

5.90 (a) E(XI - X2) = E(Xd + (-1)E(X2) = (-2) + (-1)5 = -7.

(b) Var(Xl - X2) = Var(X1) + (-1)2Var(X2) = 2 + (-1)2(5) = 7.

5.91 (a) E(XI + 2X2 - 3) = E(XI + 2X2) - 3 = E(X1) + 2E(X2) - 3 = 1 + 2( -2) - 3

= -6.

(b) Var(Xl + 2X2 - 3) = Var(Xl + 2X2) = Var(X1) + 22Var(X2) = 5 + 22(5) = 25.

158

Chapter 5 PROBABILITY DENSITIES

5.92 (a) When checking a single chip, the time saving is Y = Xl - X2. Hence the expected time saving is

(b) E(200Y) = 200E(Y) = 200(11) = 2200 milliseconds = 2.2 seconds

(c) Var(Y) = Var(XI - X2) = Var(XI) + (-1)2Var(X2) = 16 + (-1)29 = 25 Var(200Y) = 2002Var(Y) = 200252 = (1000)2.

Hence the standard deviations of Y and 200Y are 5 and 1000 milliseconds respectively.

5.93 (a) E(XI + X2 + ... + X20) = E(XI) + E(X2) + ... + E(X20) = 20(10) = 200.

(b) Var(XI + X2 + ... + X20) = Var(XI) + Var(X2) + ... + Var(X20) = 20(3) = 60.

5.94 (a) For any seven observations the normal-scores are those zi, i=1, ... ,7 that satisfy F(Zi) = i/(7 + 1). Thus,

F(Zl) = .125 = F( -1.15), F(Z2) = .25 = F( -.67),

F(Z3) = .375 = F( -.32), F(Z5) = .625 = F(.32), F(Z7) = .875 = F(1.15)

F(Z4) = .500 = F(O), F(Z6) = .750 = F(.67),

So, -1.15, -0.67, -0.32,0,0.32,0.67,1.15 are the normal-scores of any seven observations.

(b) The normal-scores plot using the seven observations 16, 10, 18, 27, 29, 19 and 17 is given in Figure 5.1.

159

•
LO •
C\J
• •
LO • •
,.-
•
LO
0
-1.0 0.0 0.5 1.0
normal-scores Figure 5.1: Normal-scores plot for Exercise 5.94b

5.95 (a) For any eleven observations the normal scores z., i=l, ... , 11, satisfy F(Zi) = i/12, thus using Table 3 the normal-scores are:

-1.38, -0.97, -0.67, -0.43, -0.21,0,0.21,0.43,0.67,0.97,1.38

(b) The observations on the times (sec.) between neutrinos are: .107, .196, .021, .283, .179, .854, .58, .19, 7.3, 1.18, 2.0. The normal-scores plot is given in Figure 5.2.

160

Chapter 5 PROBABILITY DENSITIES

en
0
c:: *
·c
-
::l (0
Q)
c::
c::
Q) '<t
Q)
;;
-
Q) C\J *
.0
en *
Q) *
*
E 0 * * * * * *
~
-1.5 -0.5 0.5 1.5 Normal-scores

Figure 5.2: Normal-scores plot for Exercise 5.95b

5.96 (a) The normal-scores plot of the aluminum alloy strength data is given in Figure 5.3.

>.
.2
tti •
E '<t ••
::l I'-
c::
·E C\J
I'-
::l
ctl 0
- I'-
0
.s:::. co
- (0
Ol
c::
Q) •
....
US
-2 -1 0 1 2 Normal-scores

Figure 5.3: Normal-scores plot for Exercise 5.96a

(b) The normal-scores plot of the decay data is given in Figure 5.4.

161

•
CD ~ •
E •
C\J
.+:; ••
>.
co
o 0 ".,..
CD T-"
0
0 .... ___..-/
a
-2 -1 0 1 2
Normal-scores Figure 5.4: Normal-scores plots for Exercise 5.96b

5.97 (a) The normal-scores plots ofthe logarithmic, square root and fourth root transformations for the decay time data are given in Figures 5.5, 5.6 and 5.7, respectively.

,.... •
••
m ••
E 0 ,F
:;::;
>-
ctl ,....
0 I ,
Q) "
a N
CJ) I _...,;
.Q
C') •••
I •
-2 -1 0 1 2
Normal-scores Figure 5.5: Normal-scores plot of the log (decay time) data. Exercise 5.97a

162

Chapter 5 PROBABILITY DENSITIES

CD
E •
:p io
>- •
ell ,.- •
U ••
CD
"0 0
- r
0 ,.-
-
0 ... _/
0
..... io
CD 0
.....
ell
::J
cr
oo
-2 -1 0 1 2
Normal-scores Figure 5.6: Normal-scores plot of square root of the decay time data. Exercise 5.97a

CD •
E •
:p •
>- ••
ell C!
U ".".,.
CD ,.-
"0
- /
0
-
0
e co
..c: 0 ... ..".,.--
-
~ •

-2 -1 0 1 2
Normal-scores Figure 5.7: Normal-scores plot of fourth root of the decay time data. Exercise 5.97a

163

(b) The normal-scores plots of the logarithmic, square root and fourth root transformations for the inter arrival time data are given in Figures 5.8,5.9 and 5.10, respectively.

- ~ •
Q) ~
E •
:;:; 0 ,.. ..
(ij ~ ."
>
.;:: /
.... 0>
ctS /"
....
Q)
- eo
C
C; ......
.Q f'-. •
-2 -1 0 1 2
Normal-scores Figure 5.8: Normal-scores plot of the log(interarrival time) data. Exercise 5.97b

Q)
E
:;:;
(ij •
0
> 1.0
.;:: C\I
....
f!! •
Q) •
- r·
c 0
- 1.0 •
0 ~ ."
- ... ./'
0
E
~ 0
1.0
ctS •
:::J
0-
CJ) -2 -1 0 1 2
Normal-scores Figure 5.9: Normal-scores plot of square root of the interarrival time data. Exercise 5.97b

164

Chapter 5 PROBABILITY DENSITIES

Q)
E
:;:::: <.0 •
ctj ~
> •
. ~ •
.....
ctj C\l ,....
.....
Q) •
.... ~ ."
c /'
-
0 /
.... en
0
0
..... <.0 ... ..,.
..c. •
....
I
~
-2 -1 0 1 2
Normal-scores Figure 5.10: Normal-scores plot of fourth root of the interarrival time data. Exercise 5.97b

5.98 The simulated values of the Weibull random variable with 0: = .05 and (3 = 2.0

are:

x = (-20ln(1- .26))~ = 2.45400, x = (-201n(1- .12))~ = 1.59896.

x = (-20 In(l - .77)) ~ = 5.42158,

5.99 The time it takes a person to learn how to operate a certain machine is normal random variable with mean I-L = 5.8 and standard deviation a = 1.2 and it takes two persons to operate the machine. Using Minitab , we generate two columns of normal variates having mean 5.8 and standard deviation 1.2. Each row represents a pair of workers.

ROW C1 C2

1 6.16327 5.71118 2 5.37122 5.15066 3 5.21900 6.44433 4 4.73400 6.35715

165

The simulated times it takes four pairs to learn how to operate the machine are

6.16327, 5.37122, 6.44433 and 6.35715.

5.100 We generate the five exponential variates, with mean (3 = 2 in Cl , Each row represents the durability of paint on one house.

ROW C1

1 3.10639 2 0.96682 3 3.24836 4 2.83178 5 2.27078

(a) The time of the first failure is .96682

(b) The time of the fifth failure is 3.24836

5.101 (a) The density is

{ .3e-·3x x > 0

f(x) = 0 elsewhere

Thus, the corresponding distribution function is

F(x) = { J;ot .3e-·3sds = _e-·3s I~ = 1 - e-·3x for x > 0 elsewhere

(b) We solve u = F(x) for x. Since u = F(x) = 1 - e-·3x, so e-·3x = 1 - u or -.3x = lnt I - u). The solution is then x = -In(l - u)j.3

5.102 (a) The density is

{ Q{3xf3-1e-axf3 for x > 0

f(x) = 0 elsewhere

Thus, the corresponding distribution function is

166

Chapter 5 PROBABILITY DENSITIES

F x = { J~ 0'./3si3-1e-nsiJ ds = _e-nsiJ I~ = 1 - «=" for x > 0

( ) 0 elsew here

(b) We solve u = F(x) for x. Since u = F(x) = 1 - e-nxiJ, so e-nxiJ = 1 - u or -O'.xi3 = In(1 - u). The solution is then x = (-(1/0'.) In(1 - U))l/i3.

5.103 Let Zl, Z2 be independent standard normal random variables. Under a change of polar coordinates, Zl = Rcos(8), Z2 = Rsin(8), the joint density of Rand 8 is

o < (J < 27r, r > 0

elsewhere

(a) The marginal distribution of 8 is

1000 1 1000 2

h((J) = f(r, (J)dr = - re " /2dr

o 27r 0

Hence 8 has uniform distribution on (0, 27r). The marginal distribution of R is

1 1027r 2 1 2 2 2

II (r) = - -e: /2d() = -re-r /2() I 7r = -e:: /2 for r > 0

27r 0 27r 0

Thus, R has Weibull distribution with 0'. = 1/2 and /3 = 2. Since

Rand 8 are independent.

(b) Let U1 = 8/27r and U2 = 1 - e-R2 /2, then U1 and U2 are independent since

167

Rand e are independent. The distribution function of U1 is

Thus, U1 has uniform distribution on (0, 1). The distribution function of U2

is

P(U2 < U2) = P(l - e-R2/2 < U2)

P(R2 < -2ln(1 - U2)) = P(R < (-2ln(1 - U2))1/2)

Thus, U2 has a uniform distribution on (0, 1).

(c) Since 1 - U2 = e-R2/2 and U1 = 8 /27f, we have

z, = Rcos8 = V-2lne-R2/2cos8 = V-2ln(1- U2)cos(27fUd

Note that 1 - U2 also has a uniform distribution on (0, 1). Hence ln U2 can be used in place of In(l - U2) in above equations, when we use independent uniform random variables U1 and U2 to generate independent standard normal variables Zl and Z2. This completes the proof.

5.104 The eight exponential variates we generated, with (3 = .2, are:

0.053320 0.025485 0.071669 0.377028 0.208014 0.026851 0.433405 0.747732

5.105 (a) Histogram of the time of the 100 first failures is

168

Chapter 5 PROBABILITY DENSITIES

0 -
-.:t
0
(1)
0 1--,--
C\I
0 ~
..-
0 ~ ~

0.0 0.5 1.0 1.5 2.0
x Figure 5.11: Histogram of the 100 first failure times. Exercise 5.105a.

(b) Histogram of the time of the 100 fifth failures is given in Figure 5.12.

r--
0
(1)
r--
0
C\I
- I--
0
..- n
0

0 5 10 15
x Figure 5.12: Histogram of the 100 fifth failure times. Exercise 5.105b.

169

5.106 Eight generated values for the normal random variable with J1 = 123 and (J = 23.5

are:

139.529,159.167,129.411,143.942,176.539,88.797, 89.581,196.666

5.107 (a) The histogram of the 400 learning times for the pairs of operators is

0
C\J r-
T'"
-
0
co -
0 -
v ~
0 h

4 5 6 7 8 9
x (b) The histogram of the 100 values representing the time to train four pairs of operators is given in Figure 5.13.

o C\J

o C')

o

o

6

7

8

9

x

Figure 5.13: Histogram for Exercise 107b.

170

Chapter 5 PROBABILITY DENSITIES

5.108 Let X be a random variable with density function

{ k(l - x2) for 0 < x < 1 f(x) =

o elsewhere

To find k we need to integrate f(x) from 0 to 1 and set it equal to 1. Since,

101 k(l - x2)dx = k(x - x3 /3) I~ = k(l - 1/3) = k(2/3) = 1,

we have k = 3/2.

( a)

3 J02 3

P(.l < X < .2) = - (1 - x2)dx = -(x - x3/3) 1021

2 01 2 0

3 3.293 .293

= "2(.2 - .008/3 - .1 + .001/3) = "2-3- = -2- = .1465

(b) P(X> .5) = (3/2)(x - x3 /3) 1~5 = (3/2)(1 - 1/3 - .5 + (.53)/3) = 5/16

5.109 The distribution function is given by:

1x loX 3 3

F(x) = f(s)ds = -(1 - s2)ds = -(x - x3/3)

-00 0 2 2

Thus,

(a) P(X < .3) = F(.3) = (3/2)(.3 - .33/3) = .4365

(b) P(.4 < X < .6) = F(.6) - F(.4) = (3/2)(.6 - .63/3) - (3/2)(.4 - .43/3) = .792 - .568 = .224.

5.110 (a) P(error will be between -0.03 and 0.04)

1004 1002

= 25dx = 25dx = 1

-003 -002

171

(b) P(error will be between -0.005 and 0.005) = 25(.01) = .25

5.111

lol 3lo1 2 3 (X2 X4) 1 3

f-l = xf(x)dx = - x(l - x )dx = - - - - 10 = -

o 20 22 4 8

Next,

3 1 3 (3 5)

I 2 2 X X 1

f-l2 = - r x (1 - x )dx = - - - - 10 = .2

2 Jo 2 3 5

so the variance is equal to

2 I 2 (3)2

a = f-l2 - u =.2 - 8 = .0594

5.112 Let Z be a random variable with the standard normal distribution.

(a) P(O < Z < 2.5) = F(2.5) - .5 = .9938 - .5 = .4938

(b) P(1.22 < Z < 2.35) = F(2.35) - F(1.22) = .9906 - .8888 = .1018 (c) P( -1.33 < Z < -.33) - F( -.33) - F( -1.33)

= .3707 - .0918 = .2789

(d) P( -1.60 < Z < 1.80) = F(1.80) + F( -1.60) = .9641 - .0548 = .9093

-3 -2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.112 (a) z

5.112 (b) z

172

Chapter 5 PROBABILITY DENSITIES

0.2789

-3 -2 -1 0 1 2 3

-3 -2 -1 0 1 2 3

5.112 (c) z

5.112 (d) z

5.113 Let X be a normal random variable with JJ = 4.76 and (J = .04

(a) P(X < 4.66) = F((4.66 - 4.76)/.04) = F(-2.5) = .0062

0.0062

4.6

4.7

4.8

4.9

-3 -2 -1 0 1 2 3

5.113 (a) x

z

(b) P(X > 4.8) = 1 - F((4.8 - 4.76);'04) = 1 - F(l) = 1 - .8413 = .1587

173

(c) P(4.7 < X < 4.82) = F(4.82 - 4.76)/.04) - F((4.7 - 4.76)/.04) = F(1.5) - F( -1.5) = .9332 - .0668 = .8664

4.6

4.7 4.8

4.9

-3 -2 -1 0 1 2 3

5.113 (b) x

z

4.6

4.7 4.8

4.9

-3 -2 -1 0 1 2 3

5.113(c) X

z

5.114 (a) F(Z.lO) = .90 = F(1.28), thus Z.lO = 1.28 (b) F(Z.OOl) = .999 = F(3.09), thus Z.OOl = 3.09

5.115 Ql = -aZ.25 + J-l = -27(.675) + 102 = 83.775 Q2 = aZ.5 + J-l = J-l = 102

Q3 = aZ.25 + J-l = 27(.675) + 102 = 120.225

5.116 The density function is

174

Chapter 5 PROBABILITY DENSITIES

f(x) = { OArrpX-1 exp [-(lnx - a)2/2rB2] x> 0, e > 0 elsewhere

(a)

P(X > 200)

1200 1 -1 (I )2

1 - P(X < 200) = 1 - --x-1e2i37 nx-adx

o v'2iirB

jln(200) 1 ()2 In(200) - a

1 - --e- Y-Q dy = 1 - F( )

-00 v'2iirB rB

1 - F( In(200) - 8.85) = 1 - F( -3.448)

1.03

F(3.448) = .9997

(b)

P(X < 300) = F( In(300) - 8.85) = F( -3.05) = .0011 1.03

5.117 The density function is

{ .25e-·25x x > 0 f(x) =

o elsewhere

(a) P(time to observe a particle is more than 200 microseconds) = _e-·25x I~ = e-·05 = .951

(b) P(time to observe a particle is less than 10 microseconds) = 1 - e-·0025 = 1 - .9975 = .0025

5.118 The normal-scores plot of the suspended solids data is given in Figure 5.14.

5.119 The normal-scores plot of the velocity of light data is given in Figure 5.15.

175

• •
en
~ •
g 0
io
-0
CI>
-0
C 0
CI> C") • • •
0.
en •
::l • •
Cf) •
0 •
-1.5 -0.5 0.5 1.5
Normal-scores Figure 5.14: Normal-scores plot of the suspended solids data. Exercise 5.118 .

•
-
L: 0
.2> ""'" •
-
0 0
>. • • •
- C")
'0 • •
0 •
a:; 0
> C\I • •
•
-1.5 -0.5 0.5 1.5
Normal-scores Figure 5.15: Normal-scores plot of the velocity of the light data. Exercise 5.119.

176

Chapter 5 PROBABILITY DENSITIES

5.120 Let X be the gross sales volume, having the gamma distribution with 0:' = 100y'ri and (3 = ~ If the sales costs are 5,000 dollars per salesman then the expected profit is E(P) = EX - 5n = 0:'(3 - 5n = 50y'ri - 5n which is maximized at n = 25.

5.121 Let X be the strength of a support beam, having the Weibull distribution with 0:' = .02 and (3 = 3.0.

100 3

P(X > 4.5) = (.02)3x2e-·02x dx 4.5

Using the change of variable u = x3, we have

100 3

P(X > 4.5) = .02e-·02udu = e-·02(4.5) = .1616

(4.5)3

5.122 Let (X, Y) have the joint density function

{ .04e-·2x-.2y f(x, y) =

o

for x > 0, y > 0 elsewhere

(a)

II (x) = 1000 .04e-·2x-.2Ydy = .2e-·2x for x > 0

h(Y) = 1000 .04e-·2x-.2Ydx = .2e-·2y for y > °

(b)

1000 1000 (x + y)(.04)e-·2x-.2Ydxdy

_ 100 100 x(.04)e-·2X-.2Ydydx + 100 100 y(.04)e-·2x-.2Ydxdy

100 x(.2)e-·2X (1000 .2e-·2Ydy) dx

+ 100 y(.2)e-·2Y (100 .2e-·2Xdx) dy

177

100 100 1 1

= x(.2)e-·2xdx + y(.2)e-·2Ydy = - + - = 10.

o 0 .2 .2

(c) Here E(X) = 1/.2 = E(Y), so E(X + Y) = E(X) + E(Y) = 10.

5.123 (a) The independent random variables Xl and X2 have the same probability distribution b(x; 2, .7). Hence the joint probability distribution of Xl and X2

is

where Xl = 0,1,2, and X2 = 0,1,2.

(b)

1(0,1) + 1(0,2) + 1(1,2)

- .0378 + .0441 + .2058 = .2877

5.124 (a) E(3X1 + 5X2 + 2) = E(3X1 + 5X2) + 2 = 3E(X1) + 5E(X2) + 2 = 3(-5) + 5(1) + 2 = -8.

178

5.125 (a) E(XI + X2 + ... + X30) = E(X1) + E(X2) + ... + E(X30) = 30( -5) = -150.

(b) Var(X1 + X2 + ... + X30) = Var(Xr) + Var(X2) + ... + Var(X30) = 30(2) = 60.