# iv

INTRODUCTORY NOTE

CHAPTER 2

This solutions manual complements a study of probability theory, based on the new textbook Probability for Risk Management, by Matthew Hassett and Donald Stewart.

The solutions given here are brief, and are intended to give an outline of how to proceed. If the reader understands the examples in the text, these solutions should be adequate,

In most cases there will be more than one way to work a problem. The reader may expect to encounter solutions which differ from those given here.

The red suits are hearts (H) and diamonds (D), and the face cards kings (K), queens CO) and jacks (1). Thus the outcomes are KH, OH, JH, KD, OD and 10,

2-2.

(a) The loss can be any positive rational number.

(b) The loss is any rational number in (1,000, 1,000,000).

/

2-3. (a) S consists of the positive integers from I to 25.

(b) E consists of the odd integers from 1 to 25,

2-4, S consists of all ordered pairs (r,g) where r = 1,2,3,4,5, or 6 and g = 1,2,3,4,5, or 6.

2-5. Count the number of ordered pairs with the desired sum in the list in the answer For Exercise 2-4, For example, the only two pairs which sum to II are (5,6) and (6,5), so the answer to part (c) is 2.

2-6. S consist of all sequences xyz where each of x, y and z is either BorG,

2-7. The outcomes in ~E are the even (not odd) integers between and 25.

2

Chapter]

Chapter 2

3

2-8.

The outcomes in AnB are the club face cards.

To verify (2.2) draw a Venn diagram for Au(BnC) and a second one showing (AvB) and (AuC), and observe that first is the intersection of the second and the third.

2-9. AuB is all rational numbers in (1,000, 500,000).

Au(BnC)

AnB is all rational numbers in (50,000, 100,000).

2-10. AnB requires an outcome where the coin shows a head and the number on the die is greater than 2. It consists of the ordered pairs (H,3). (H,4), (H,5) and (H,6).

A

2-11. E = {(l,5),(2,4),(3,3),(4,2),(5,1)}

F = {( 1,1 ),(2.2),(3,3),(4,4),(5,5),(6,6)}

EuF = {( I ,5),(2,4),(3,3),(4,2),(5,1 ),e 1,1 ),(2,2),( 4,4),(5,5),(6,6)}

(AuB)n(AuC)

2-14. To verify (2.3) draw a Venn diagram for -(AuB) and a second one showing -A and -B, and verify that the first is the intersection of the second and the third.

EnF = {(3,3)}

-(AvB)

2-12. E = {GGG,GGB.GBG,GBB} F = {GBG,GBB.BBG,BBB}

EuF = {GGG,GGB,GBG,GBB.BBG,BBB}

Enf = {GBG,GBB}

-An-B

If -, "\

Iv

\11 t>- L

-rr- 2-13. To verify (2.1) draw a Venn diagram for An(BuC) and a second one showing (AnB) and (AnC) and observe that the first is the union of the second and the third.

To verify (2.4) draw a Venn diagram for -(AnB) and a second one showing -A and -B, and veri:fy that the first is the union of the second and the third.

-(AnB)

An(BuC)

(AnB)u(AnC)

-,

-Au-B

"V ""

V

~

IA _...., 4

Chapter 2

Chapter 2

2-15. See verbal statements in Answers to the Exercises.

2-20. Using the given information we get the following Venn diagram.

2-16. Let H be the set of those with health insurance and L be the set of those with life insurance. Then (LuH) ~ 38 + 29 - 21 = 46.

s

2-17. Let L be the set of those with the company more than 10 years and D be the set of those with college degrees. The given data can be used to fill in the Venn diagram below. Then n(LuD) = 134 - 23 = III. Then nrD) = n(LuD) - 19 = III - 19 = 92.

The total number of clients is 61.

LC]_)D

19 65

23

2-21. Using the given information we get the following Venn diagram. 'd5"

2- I 8 Let S be the set of those who own stocks and B be the set of those who own bonds. So n(SnB) = 67 + 52 - 94 = 25.

L

2-19. The given data can be used to fill in the Venn diagram below.

The number of students taking exactly one of these courses is 54 + 38 = 92.

(a) The number with no policies is 85 -74 = II. (b) The number with only health policies is 17.

(c) The number with exactly one pol icy is 21 + 17 + 6 = 44. (d) The number with life or health but not auto is

21+12+17=50.

H~B ~

2-22. Number of outcomes is 2·6 = 12 .

. 2-23. Number of ways to select car is 4· 15·6 = 360.

2-24. Number of ways to select classes is 7 . 8 . 4 . 7 = 1568.

5

6

Chapter 2

Chapter 2

2-25. Number of outcomes is 26·2 + 26·6 = 208.

2-33. First choose a pair of adjacent seats, which can be done in 7 ways, and the two people can be seated in those seats in 2 ways. Then arrange the 6 remaining people in remaining 6 chairs. This can be done in 6! ways, so the number of seating arrangements is 7· 2 . 6! = 10,080.

2-26. Each subset corresponds with exactly one sequence. For

example the subset {XI,X2} corresponds with the sequence (I, I ,0,0, ... ,0). There are 2" sequences, so there are 2" subsets.

2-34. The number of possible committees is C(30,3) = 4060. (Order is irrelevant.)

2-27. If repetitions are allowed, the number of words is 64 = 1296.

Ifrepetitions are not allowed, the number is P(6,4) = 360.

2-35. The number of hands IS C(52,5) irrelevant.)

2,598,960. (Order is

2-28. If repetitions are allowed there are 8· 106 possible numbers.

2-36. (a) Number of hands with all hearts is C(13,5) = 1287.

If repetitions are not allowed there are

8 ·9· 8 . 7·6 . 5 ·4 = 483.840 numbers.

(b) Number of hands with all same suit is 4C(13,5) = 5148. (Choose one of the 4 suits and then choose 5 cards.)

2-29. The number of seating arrangements (in order) is P( 12,7) = 3,991,280.

(c) Number of hands with (A.A.KKJ) is C(4,2)C(4,2) . 4 = 144. (Choose 2 aces and then 2 kings and then I jack.)

2-30. The number of ran kings (in order) is P(1O,4) = 5040.

2-37. Number of ways to pick cast is C(15,4)C(13,5) = 1,756,755. (Choose 4 boys and then 5 girls.)

2-31. The number of ways to fill the offices is P(30,3) = 24,360.

2-38. Number of ways to select balls is C(29,5)· 42 = 80,089,128.

2-32. The number of seating arrangements is P(4,4)P(6,6) ~ 17,280. (Arrange the left-handers in the first 4 seats, and then arrange the right-banders in the remaining 6 seats.)

2-39. Number of distinguishable arrangements of MISSISSIPPI is 11 !/(4!4!2!) = 34,650. (There are II slots to fill with the II letters. First select 4 slots for the 4 I's, and then 4 slots for the 4 S's, and 2 slots for the 2 P's, leaving one slot left for the M.)

7

2-40. Number of ways to assign the actuaries is 12!/(5!3!41) = 27,720, since this constitutes a partition of the group.

2-41. Number of ways to assign the analysts is 91/{3!)J = 1680, since this constitutes a partition of the group.

2-42. In this case the 3 teams are not distinguishable by task. If the 9 analysts are divided into 3 three-man teams, A. B and C, and have distinguishable tasks, these teams can be assigned to the 3 tasks in 3! = 6 ways. But if the tasks are all the same, those 6 partitions constitute a single division into teams. Hence the divisions into teams is the number of partitions divided by 6 which is 1680/6 = 280.

2-43. (2s - t/ = (2S)4 + 4{2s/( -t) + 6(25/( _t)2 + 4(2s){ _t)3 + (_t)4

2-44. Term is C(8,3)(2u)5(-3v/, and coefficient is -48,384.

2-45. An x"·kl is obtained by selecting k of the (x+y) factors from which to take y and taking x from the remaining n-k factors. This can be done in C(n,k) ways.

2-46. The number of subsets of size k from a set of size n is C(n,k).

Then 2" = (I + 1)" = IC(n,k) = the total number of subsets.

CHAPTER 3

There are 8 total outcomes. 3 of which give 2 heads and 1 tail, HHT, HTH or THH. Hence P(2 heads and I tail) = 3/8.

3-2.

Peat least I head) = 1 - P(no heads) = 1 - 1/8 = 7/8

3-3.

(a) P(ball is red) = 3/16

(b) P(ball is not green) = P(ball is red or blue) = (6 + 3)/16 = 9116

3-4.

Number of employees with degree in either mathematics or economics is 21+33-7 = 47. The probability tbat the person chosen has one of these is 47/68.

3-5.

There are 36 outcomes, of which 6 give a 7, 2 an 11. and 6 give a number less that 5.

(a) P(7) = 6/36

(b) P(II)=2/36

(c) P(less that 5) = 6/36

3-6.

Number of clients with either life or auto insurance 15

45+32 - 16 = 61. The probability a client has neither is 17178.

3-7.

(a) P(a1l3 red) = C(4,3)1C(10,3) = 4/120 = 1130

(b) pel red and 2 green) = C(4,1)C(6,2)1120 = 60fl20= 1/2

(c) P(al! red or all green) = [C(4,3)+C(6,3)]/]20 ee 24/120 = 1/5

10

II

Chapter 3

Chapter 3

3-8.

(a) P(a1l4 good) = C(35,4)/C(40,4) = 52360/91390 = .5279

3-15. (a) P(7) = 1/6, P(~7) = 5/6. Odds for a 7 are 1 :5.

(b) P(ll)= 1118, P(-ll) = 17/18. Odds against an II arc 17:1.

(b) P(2 good and 2 defective) = C(35,2)C(5,2)/C(40,4) = 5950/91390 = .0651

3-9. Total seating arrangements is 10! = 3,628,800. There are two choices, M or W, for the first chair. This determines the 5 chairs for the 5 men and the 5 chairs for the 5 women. The number of seating arrangements alternating men and women is 2(5 !)(5!) = 28,800. The probability is .0079.

3-16. If the odds against Fare a:b, the odds for F arc b:a, and P(F) = b/(b+a).

3-17. Eu-E = Sand Erv-E = 0, so Pf E) + PC -E) = P(S) = I.

3-10. Peat least 2 born on same date)

= 1 - P(alJ born on different days) = I - PC31 ,8)13 I R = .6271

3-18. A = (AnB)u(An-B) and B = (BnA)u(Bn-A), so (AuB) = (An-B)u(AnB)u(Bn-A).

P(AuB) = P(An-B) + P(AnB) + P(Bn-A)

= P(CAn-B) + P(AnB) + P(Bn-A) + [P(AnB) - P(AnB)] = peA) + PCB) - P(AnB)

3-11. peat least 2 born on same day of week)

= 1 - P(all born on different days) = I - P(7,4)174 = .6501

3-19. Ptjunior or senior) = .244 + .215 = .459

3-12. P(balls of both colors) = I - P(all red or all blue) = 1 - [C(5,4) + C(6,4)]/C(II,4) = 1 -20/330 = 31/33

3-20. P(no claim) = 1- Ptliabiliry or comprehensive)

= 1-(.22 + .37 - .13)= .54

3-13. Number of hands is 2,598,960 (Exercise 2-35). Number of ways to pick the 2 ranks is P(13,2). Number of full houses is P( 13 ,2)C( 4,3)C( 4,2) = 3744, so Ptful! house) = .0014.

3-21. (a) P(C or better) = .131 +.278 + .312 = .721 (b) P(D or E) = .089 + .094 = .183

3-14. Number of ways to pick 2 suits for pairs is C(13,2). (Order doesn't matter here.) Number of hands is C(13,2)C(4,2)C(4,2)·44 = 123,552, so P(two pair) = .0475.

3-22. P(flu shot or tuberculosis test) = 1 - .21 = .79

= P(flu shot) + P(tuberculosis test) - P(both)

= .37 + .58 - P(both). P(both) = .95 - .79 = .16

3-23. P(AiCorbetter)=.131/.721=.1817

12

Chapter 3

Chapter 3

13

3-24. P(missed I day I missed work) = .237/.365 = .6493

3-32. peA) = 2/3, PCB) = 112 and PCc) = 112.

3-25. P{low claim I claim filed) = .169/.208 = .8125

P{AnB) = lI2 '" P(A)P(B), so pair is dependent.

P(AnC) = 1/3 = P(A)P(C), so pair is independent.

3-26. (a) P(both hearts) = P(I" heart)P(2'd heart Il't heart) = (114)(12/51) = .0588

(b) P(neither a heart)

= P( I st not a heart)P(2,d not a heart II" not a heart) = (3/4)(38/51) = .5588

(c) P(exactly one heart) = 1- .0588 - .5588 = .3824

P(BnC) = 1/3 '" P(B)P(C), so pair is dependent.

3-33. P{color blind) = 9/300, and P(color blind I male) = 71130.

Male and color blind are dependent.

3-27. P(exactly 2 heads I at least I head) = (3/8)/(7/8) = 317

3-34. (a) P(pass both classes) = (.75)(.84) = .63

3-28. P(exactly 2 heads II" is a head)

= P(exactly 2 heads and I" is a head)/P( I" is a head) = (2/8)/(4/8) = 112

(b) P(fail both classes) = (.25)(.16) = .04

P(pass exactly one class) = I - (.63+.04) = 33

3-35. P(none break down) = .953 = .8574

3-29. Let E be the event exactly 2 cards are hearts and F be the event all three are hearts. EuF is the event at least 2 are hearts. Then neE) = C( [3,2)(39) = 3042 and n(F) = C( 13,3) = 286. so P(F I EuF) = 286/(286+3042) = .0859.

3-36. peat least one fails) = 1 - P(neither fails) = 1- (.83)(.88) = .2696

3-37. peE) = 1/2 and P(F) = 3/8. P(EnF) = 114. (EnF = {HHT,HTH}) P(EnF) '" P(E)P(F), so events are dependent.

3-30. Let E be the event the card is a spade. peA) = PCB) = P(C) = 112.

P(AnB) = P(AnC) = P(BnC) = peS) = 1/4 = P(A)P(B) = P(A)P(C) = P(B)P(C), so each pair is independent.

3-38. (a) P(injury) = (.6)(.35) + (.4)(.2) = .29

(b) peassembly line worker I injury) = .08/.29 = .2759

P(AnBnC) = peS) = 114", P(A)P(B)P(C)

3-39. P(nickel) = (112)(4/15) + (1/2)(1/3) = 3/10 Ptjar II I nickel) = (J/6)/(3110) = 5/9

3-31. P(AlnA2) = 42/522, and P(A2) = 4/52 = P(AI)

P(A2IAl) = P(AlnA2)/P(AI) = 4/52 = P(A2), so events are independent.

14

Chapter 3

3-40.

P(O claims) = (.6)(.8) + (.4) (.5) = .68 P(2 claims) = (.6)(.05) + (.4)(.02) = .11

(a) P{low risk I 0 claims) = AS/.68 = .7059 (b) P(high risk 12 claims) = .08/.11 = .7273

3-41. P{defective) = (.4)(.01) + (.35)(.02) + (.25)(.04) = .021 P(machine A I defective) = .004/.021 = .1905

3-42. Define the events; D: has disease, -D: doesn't have disease.

Y: tests positive, and N: tests negative. .

P(Y) = (.2){.95) + (.8)(.3) = .43, and PeN) = (.2){.05) + (.8)(.7) = .57

(a) P(-D I Y) = .24/.43 = .5581

(b) P(D IN) = .01/.57 = .0175

3-43. Let A be event first card is a heart, and B that second is a heart.

P(AnB) = (1/4)(12/51), and PC -AnB) = (3/4)(13151).

P(second card is a heart) = (12 +39)/(4·51) = 1/4

3-44. P(severe injury) = (.7)(.10) + (.3)(.15) = . [ 15 P(plant A I severe injury) = .07/.115 = .6087

3-45. P(injury) = I - P(no injury) = I - [(.7)(.6) + (.3)(.5)] =.43 P(plant B and minor injury I injury) = (.3)(.35)1.43 = .2442

CHAPTER 4

4-[. Using the tree on page 26 we get the following table.

Number ofheads(x) ° I 2 I J

Number of outcomes I J 3 I J

p(x) II8 3/8 3/8 1118 4-2. P(X = 0) = P(first card is the ace) = 1110,

P(X = I) = P( second card is the ace I first was not the ace) = 9110 . 1/9 = 1/10,

P(X = 2) = P(third card is the ace I neither the first nor the second was the ace) = 9110·8/9·118 = 1110, etc.

4-3. P(X=x)=(5/6)'(l/6), x=0,[,2 •...

F(x) = 116 + 1/6· 5/6 + + (1/6)(5/6)'

= (1/6)(1 + 5/6 + + (5/6)') = (1/6)[1 - (5/6)'+']/( I - 5/6)

= 1-(S/6y+',x=0, 1,2, ...

4-4.

x Number of Outcomes p(x) F(x)

2 1 1136 1/36

3 2 2/36 3/36

4 3 3/36 6/36

5 4 4/36 10/36

6 5 5136 15/36

7 6 6/36 21136

8 S 5/36 26/36

9 4 4136 30/36

10 3 3136 33/36

II 2 2136 35/36

12 I 1136 36/36 16

Chapter 4

Chapter 4

17

4-5. E(X) = Lxp(x) = (2 + 3·2 + 4·3 + 504 + 6·5 + 7·6 + 8·5 + 9-4

+ 10·3 + 11·2 + 12)/36 = 7

4-13. V(X) = (10,000-114)\.0041) +(1000-114i(.073)

+ (0 - 114i<,9229) = 470.004 For 5 units, alX = 5ax = 3427.84

4-6. E(X) = (4·15 + 3·33 + 2·51 +6·1 + 3·0)/108 = 2.47

4-7. E(X) = 0(.9929) + $1000(.073) + $10,000(,0041) = $114

Average cost per claim = (730($1000) + 41($10,000»110,000 = $114

4-14. IfY=X+b,E(Y)= )l.y=E(X)+b=).!x+b.

Y-).!y=X+b-().!x+b) =X-).!x V(Y) = L(y - ).!yi = L(x -).!xi = VeX)

4-8. IfY is the claim amount for one unit, then X = lOY + 50.

From Exercise 4-7 we have E(Y) = $114.

E(X) = E( 1 OY + 50) = 10E(Y) + 50 = $1140 + 50 = $1190

4-15. (a) Lower bound is 1- (1I2i = .75.

(b) Two standard deviations equals 2(5 .8333/,2 = 4.8305.

P(7 - 4.8305 < X < 7 + 4.8305) = P(3 <=: X <; II) = 34/36 = .9444

4-9. E(aX + b) = I(ax+b)py(ax+b)

= aIxpx(x) + bIpx(x) = aE(X) + b

4-16. ).1 =(0·11,425 + I ·3100 + 2·385 + 3·90)/15.000 = .276 a2 = [11,425(G-.276)2 + 3100(1-.276i + 385(2:_.276)2

+ 90(3-.276il/15,OOO = .0287157

IT = .53587

~

4-10. E(X) = Lx(5/6)' (1/6)

4-17. x =(0·3+1·5+2·6+3·9+4·11 +5·7+6·5+7·3+8·1)/50

= 3.64

s = [Lf-(n - 3.64/149}"2 = 1.9667

~

= (5/6)(116) L x(5/6),-'

x=1

= (516)(1/6)/(1 -5/6)" = (5/6)/(1/6) = 5

4-11. Three outcomes yield I head and three outcomes yield 2 heads.

Thus either 1 or 2 is a mode.

4-12. VeX) = 2:(x-7/p(x) = (25 + 2(16) + 3(9)+ 4(4) + 5(1)+ 6(0)

+ 5(1) + 4(4) + 3(9) + 2(16) + 25)/36 = 210/36

Chapter 5

19

5-6.

(a) P(4 have B+ blood) = C(20,4)(.lO)\90)16 = .0898

(b) peat most 3 have B+ blood) = (.90/° + C(20, I )(. I 0)(.90)19 + C(20,2)(.1 0/C·90)IS

+ C(20,3)(.10),(.90)17 = .8670

CHAPTER 5

5-1. (a) P(X = 5) = C( 10,5)( lI2)1( l!2)5 = .2461 (p = q = 1;2)

(b) r(X 2" 8) = C(I 0,8)( !l2)IO + C(l 0,9)( 112)10 + C( [0,1 0)( 1/2)10 =.05467

5-7. Let X be the number of pints ofB+ blood donated.

E(X) = 50,000(.10) = 5000; VeX) = 50,000(.10)(.90) = 4500

5-2. (a) r(X = 2) = C(10,2)( 1/6)\5/6)8 = .2907

(b) P(X 2" 2) = 1 ~ P(X < 2) = I ~ [(5/6)10 + 10(1l6)(5/6)9] = .5155

5-8. (a) P(2 aces) = C(12,2)(l/13)2(l21l3)10 = .1754 (b) P(3 hearts) = C(12,3)(1l4)\314)9 = .2581

(e) Punore than 1 heart) = 1 ~ [(314)12 + 12(1/4)(3/4)"] = .8416

5-3. P(three file claims) = C(12,3)(.023/(.977)9 = .00217

5-9. P(no more than 3) = (.95)15 + C( 15, I )(.05)(.95)14

+ C(l5,2)(.05)"(.95)13 + C( 15.3)(.05)'(.95f2 = .9945

5-4. (a) P(exactly 2 defective) = C(50,2)(.02)2(.98)4& = .1858

5- J O. F or a binomial random variable with n = 2 and peS) = p we have

(b) ~ = 1000(.02) = 20; 0'2 = 1000(.02)(.98) = 19.6

2

5-5. Let X be the random variable for the number of wins, and Y = 100 ~ X be the random variable for the number of losses. The expected amount of gain = E(IOX) = 10(100)(.493) = 493 and, the expected amount of loss = E(lOY) = 10(100)(.507) = 507. The expected end result is 493 - 507 = -14 (a loss of$14).

(a) E(X)= O(l-p/+ 1(2p)(I-p)+2pl=2p

(b) V(X)=(0-2Pi(l-pi+(1-2Pi(2p)(I-p)+(2-2Pip2 = 2p(l-p)[2p(l-p) + (l-2p)2 + 2p{l-p)] = 2p(1 - p)

5-11. P(2 aces) = C(2,2)C(8,3)/C(10,5) = 2/9

5-12. P(2 with AIDS) = C(4,2)C(12,4)/C( 16,6) = .3709

20

Chapter S

5-13.

(a) P(4 of each) = C( 1O,4)C(6,4)/C(16,8) = .2448

(b) If X is number of left-handed batters, E(X) = 8(6/16) = 3

5-14. Let X be the number of Republicans on committee. (a) E(X) = 15(54/100) = 8.1

VeX) = 15(54/100)(46([00)(85/99) = 3.199

5-15. E(X) = 13(1/4) = 3.25; VeX) = 13(1/4)(3/4)(39/51) = 1.864

5-16. (a) P(X= 1)=.6e-6=.3293

(c) P(X>I)=1-e-6-.6e-6=.1219

5-17. (a) P(X = 0) = e-1 s = .2231

(b) P(X= 1)= J.5e-15=.3347 (d) P(X = 2) = 1.51e-15/2 = .2510

5-18. If X is the number of claims per year, SOOOX is the claim amount. E(5000X) = 5000E(X) = 5000(.38) = 1900

5-19. (a) The rate of claims Ber policy per year is A = 12,200/50,000 =.244. (b) P(X = 0 or 1) = e- 44(1 + .244) = .9747

(c) E(1000X) = 1000E(X) = 1000(.244) = 244

5-20. Let X and Y be the number of claims filed by two policyholders. (a) P(X =1 and Y = 1) = (OAe-4f = .0719

P(X = 0 or Y = 0) = P(X = 0) + P{Y = 0) - P(X = 0 and Y = 0) = 2e-4 - e-& = .8913.

ChapterS

21

5-21.

kk-le-k k kk-Ie-< ee-k

P(X=k-l)= --=_._-=-- =P(X=k)

(k-l)! k (k-l)! kl

k"e-k k k"e-k

For n z k, P(X = n)= -->-_ .... _-

nl n + 1 n!

k"+le-k

= ---= P(X=n+1) (n + I)!

Hence for n > k, the probabilities get smaller as n gets larger.

k"e-k k k"e-k

Forn<k-l,P(X=n)= --<-_._-

n! n + 1 nl

kfl+ie-k

= ---= P(X=n + I) (n + I)!

Hence for n < k - 1, the probabilities get larger as n gets larger. Therefore the largest probability occurs when n = k - 1 or k.

5-22. In Section 5.3A it was shown that E(X) = A for the Poisson random variable.

VeX) = E[(X - Ail = E(X2 - nx + A2)

= f (k2_2Ak+A2)A,ke!-A

k=O k.

= f k2 A,ke!-;, - 211. fk A,k~!-< + A2 f A,k:!-A

k=O k. bO . kdJ'

The second term is -2AE(X) = _2/02, and the last term is A2L:p(X) = A2( 1).

00 2 A,ke-)· _ ~ A,~e-.l

Lk J;:!- L."k (k-l)!

hO bl

If we let n = k - 1. then k = n + 1 and this sum becomes

~ (n + [) A,n+le-'< = A( fn A,"e-'; + f )~"e-A )

~ nl "=0 n! "~o nl

= A[E(X) + 1] = },2 + A.

Hence VeX) = A2 + A - 2102 + A2 = }"

22

Chapter 5

23

Chapter 5

5-23. If the first II occurs on the eighth roll, there are 7 failures first, and p = 1/1 8 and q = 17118.

5-28. (a) If the fourth 6 appears on the twentieth roll, there are 16 initial fai lures.

P(X = 7)=(lIlIS)7(l/IS) = .0372

P(X = 16) = C(J9J)(S/6)16(U6)4 = .0404

S-24. (a) The number of initial failures is 4. and p = II., q = :y,.

P(X = 4) = (3/4)4(1/4) = .0791

(b) If X is the number of failures before the fourth 6, then

E(X) = rq/p = 4(5/6)/( 1I6) = 20.

(b) The number of initial failures is 9, and p = 1/13, q = 12/13. The expected total number ofrolls is 20.,. 4 = 24.

peX = 9) = (12113)9(1113) = .0374

5-25. Probability of an ace is p = 1/13. so q = 12113.

5-29. Let X be the number of failures before the fifth 6.

E(X) = q/p = (12/13)/(1/13) = 12

V (X) = q/p2 = (I 2113 )/( JIl3)2 = 156

E(X) = rq/p = 5(516)1(1/6) = 25 VeX) = rq/p2 = 5(5/6)1(1/6i = ISO

5-26. (a) If the fifth patient is the first one with the disease, then there are 4 initial failures.

5-30. If the fifth success occurs on the sixteenth call. there are 11 initial failures.

rex = II) = C(IS,4)(.8)ll(.2)5 = .0375

,i P(X=4)=(.85)\,15) =.0783

(b) For the second part there are 9 initial failures.

5-31. Number of successes needed is 20001250 = 8. If X is the number of failures before the eighth success, then

r(x = 9) = (.85)\. [5) = .0347

E(X) = 8(.8)/(.2) = 32.

5-27. (a) If the third heart appears on the tenth draw, there are 7 failures.

The total expected number of calls is 32 + 8 = 40.

P(X = 7) = C(7+3-I,J-l)qV = C(9.2)(3/4)7(1/4/ = .0751

(b) If X is the number of non-hearts drawn before the fifth heart

then '

EeX) = rq/p = 5(3/4)1(1/4) = 15.

24

ChapterS

25

Chapter 5

5-32. (a) There would be 12 failures before the third success.

5-35. Let X be the discrete uniform random variable with p(x)= lin for x = 1,2,3, ... .n.

P(X = 12) = C(l4,2)(.85)1l(.] 5)3 = .0437

] + 2 + ... + n 1 n(n + I) n + 1 _

E(X) = Lxp(x) = - ~.-~=--].l

n n 2 2

(b) If X is the number patients without the disease tested before the sixth with, then

VeX} = E[(X - fiil = E(X' - 2fiX _._ fi'} = L.(k1 - 2).lk + ~/)p(k) = a2p(k) - 2)lL.kp(k) + 1-l'L:p(k)

E(X) = 6(.85)1(.15) = 34.

5-33.

The second term is -2IleX) = _2)l2, and the last term is ~l.

. 2 (n + 1)2

sum of these two IS -).l = ~-4- .

The

The outcome X is a discrete uniform random variable with p(x) = 1/6, x = 1,2,3,4,5,6.

E(X) = (1 + 2 + 3 + 4 + 5 + 6)/6 = 21/6 = 3.5

VeX) = [(1-3.5i + (2-3.5)' + (3-3.5)2 + (4-3.5i + (5-3.5i +(6-3.5)']/6 = (6.25 +2.25 + .25 + .25 + 2.25 + 6.25)/6 = 17.5/6=35/12

.!.. n(n + 1)(2n+ 1)

n 6

2n2 +3n+1

6

5-34.

Let Y be the number on the ball chosen. Y is a discrete uniform random variable, with n = 25 and pen) = 1125, n = 1,2, ... ,25. and X = 1000Y.

2n 2 + 3 n + 1 n 2 + 2n + 1

VeX) = 6 4

801 +120+4-602 -12n-6 n2_1

24 12

E(X) = 1 OOOE(Y) = 1000(26)12 = $13,000

VeX) = 1 0002VCY) = 10002(252 - 1)/12 = 10002(52) = ri

c = $7,211.1 0

Chapter 6

27

6-3. Using the utility function u(x) = In(x+ 1) we have the following:

Method 1

CHAPTER 6

Wealth w 0 10,000

u(w) = In(w+ I) 0 In(l 0,00 I)

pew) .10 .90 6-1.

Method 2

Wealth w 0 9,025

u(w) - In(w+ I) 0 In(9,026)

pew) .02 .98 o

.69

.23

(a) E(X) = 0(.69) + 1(.23) + 2(.07) + 3(.0 I) = .40 E{500X + 50) = 500E(X) + 50 = 250

(b) E(X2) = 0(.69) + 1(.23) + 4(.07) + 9(.01) = .60 (c) E(X3) = 0(.69) + 1(.23) + 8(.07) + 27(.01) = 1.06

E[u(W])] = .10(0) + .90ln(l0,001) = 8.289 E[U(W2)] = .02(0) + .98In(9,026) = 8.926

Method 2 gives a higher expected utility.

6-2.

Th f II

t bl I t

th oIling of a pair of dice.

e a owm a ere a es to er

X p(x) xp(x) x·p(x)

2 ](36 2/36 4/36

3 2/36 6/36 18/36

4 3136 12/36 48/36

5 4136 20/36 100/36

6 5/36 30/36 180/36

7 6/36 42/36 294/36

8 5/36 40/36 320136

9 4/36 36/36 324/36

10 3/36 30/36 300/36

11 2/36 22136 242/36

12 1136 12136 144/36 6-4. This is primarily an exercise for the reader to observe the precision of his or her particular computer or calculator.

E(X) = 1,000,000,000 = I-t

V(X)=E[(X-I-t)l] =(1/3)(.01 + 0 + .01) = .0066

The value you get using VeX) = E(X2) - E(X)' will vary from calculator to cal culator.

6-5. Let X be the binomial random variable with n trials and peS) = p.

(q + p)" = 2:C(n,x)p' g"'x

E(X) = kXp(X) = 252/36 = 7

E(X2) = 2:x2p(X) = 1974/36 = 54.8333

VeX) = EeX2) - E(xi = 54.8333 - 49 = 5.8333

Ifwe replace p with pel, we get Mx(t) = (q + pel)".

28

Chapter 6

Chapter 6

6-6. For the Poisson random variable with rate I., Mx(t) = e.t(e'-l) .

6.8. For the negative binomial random variable with Pt S) = p and X the number of failures before r successes.

M (t) _ p'

x - (I-qe')'

M;'; (t) = Ae'e,Ce'-l) , M~ (0) = 1.(1 )(1) = A = E(X)

r 1

M' (t)- rp qe

x - (I-qe,)'+l

VeX) = E(X2) - E(xi = t.

EeX) =

M' (0)- rp'q

x - (I_q)'+l

rp' q rq pt+l =p

• rp'qel(l_qet)+rCr+l)p'qlell

M (t) - -'---'--'-----'----'----::~:--'-'~~

x - (1- qe 1 )'+2

EeXl) = M" (0)= rp'q(l-q)+r(r+ l)p'q2

x (1- q)'+l

_ rpq + r(r + I)ql pl

V(X)=E(X2)-E(X/= rpq+r(r+l)q2 r2q2

p2 p2

rpq + rq 2 rq(p + q) rq

= pl = p ' = PT

ji. 'I

~',

11

Iii

I •.• ',jI'

l

For the geometric random variable with Pt S) = p, Mx(t) = _p_, and M~(t)= _p(_q~l;.

1- qe ' (1- qe )

E(X) = M~ (0) = __E'!____ = pq = _g_ (1- q)2 pl P

M" (t) = pge' (1- qet) + 2pq1e11

x (1- gel)3

ECX1) = M" (0) = pq(l- q) + 2pq2

X (I_g)]

6-7.

= pq + 2q2 p2

VeX) = ECX2) _ E(xi = pq + 2q2 p2

6-9. For the discrete uniform random variable, p(x) = lin, for x = L 2, ... , n.

" I 0 e'(l-e"')

(a) Mx(t)=Le"pCt)=-l>X1= , ,t"'O

x~l n x~1 n(l - e )

29

30

n

M;(t)~ 2>2C"

E(X2) = M;(O) ~ 1 ~,>2 = 1. nCn + 1)(2n + 1)

n x~1 n 6

= (n + I )(2n + I) 6

VeX) = E(X2) _ E(Xi ~ (n + 1)(2n + I) _ (n + 1)2

6 4

6-10. Mx(t) = 2:e"p(x) = .42 + .30e' + .17e2' + .ue"

M~ (I) = .30e' + .17(2)e2t + .11(3)e" E(X) = M~ (0) = .30 + .34 + .33 ~ .97

M;(t) = .30e' +. [7(4)e2t + .11(9)e" E(X2) ~ M~ (0) = .30 + .68 + .99 = 1.97

6-11. Mx(at) ~ E(e"x) = L:e"'Xp(x)

Chapter ti

r

Chapter 6

31

6-12. For the binomial random variable with n trials and peS) ~ p, Mx(t) ~ (q+pe')",

Ifn = 8 and p ~ .6, then Mx(t) = (.4 + .6e't

IfY = 3X + 4, then My(t) ~ e~'(.4 + .6cJ1i

6-13. The moment generating function for the negative: binomial distribution is MxCt)=l(-p-, J'

1- qe

Therefore (~)5 is the moment generating of the negative 1-.3e

binomial distribution with r = 5 and p = .70.

6-14. Successive applications of the linear congruence y = 9x + 11 (mod 16) yields the following table.

k Xk 9Xk + 11 9x, + II

(mod 16)

I 6 65 1

2 I 20 4

3 4 47 i 15 :

4 15 146 2

5 2 29 13

6 13 128 0

7 0 II II

8 [] lID 14

9 14 137 9

[0 9 92 .-

12

11 12 119 7

12 7 74 10

13 10 101 5

14 5 56 8 -----j

15 8 83 i 3

16 3 38 6 _ __j 32

Chapter 6

The following table is for the simulations in Exercises 6-15 and 6-16. In Exercise 6-15, if the random number x < .40, the result was a success. Otherwise it was a failure. In Exercise 6-16, if the random number

x < .50, the result was a head. Otherwise it was a tail. CllAPTER 7

**Trial Random Number S orF HarT
**

1 .5619 F T

2 .4500 F H

3 .3566 S H

4 .5844 F T

5 .8638 F T

6 .9983 F T

7 .0225 I S H

8 .8026 F T

9 .3516 S H

10 .4584 F H

11 .7855 F T

12 .9955 F T

13 .6558 F T

14 .1280 S H

15 .3908 S H

16 .3729 S H

17 .1326 S H

18 .9246 F T

19 .6867 F T

20 .9638 F T 7-1. The function f(x) = 1.5x + .25, for 0:0; x s I, and ° elsewhere,

(a) Clearly f(x) ::: 0 for all x, so we only need to show that the area under the curve is 1.

[f(X)dx= f~(l·5x+.25)dx=(.75x2 +.25x) I~ =.75+.25=1

Hence f(x) is a probability density function.

(b) F(x) = P(X:O; x) = D1.5t + .25)d/ = (.75x2 + .25x) for O:o;xs; l. F(x)=Oforx<O,andF(x)= 1 forx> 1.

(c) P(O S; X S; 112) = F(.5) - FeD) = [.75(.5/ + .25(.5)J - [0 + 0] = .3125

P( 114:0; X S; 3/4) = F(.75) - F(.25)

= [.75(.75)2+.25(.75)]-[.75(.25»)+.25(.25)] = .609375 ~ .109375 = .50

7-2.

(a) We need to find a so that the area under the curve is 1.

foo f(x)dx = f~ are -2x ~ C -;, )dx = a[ -( 1 12)e -2, + (113 )e'X ]I~

-~ 0 0

= a(O) - a(-lI2 + 1/3) = a/6

For this to be a probability density function a/6 = 1, and a = 6.

6-15. In these 20 trials there are 7 successes.

6-16. The first trial (first set of 5 numbers) yields 2 heads, the second trial yields 3 heads, the third trial yields 2 heads, and the fourth trial yields 2 heads.

(b) P(X S; 1) = 6 f~( e -2x ~ e ··3x )dx = (- 3e -2, + 2e'}') I:

= (_3e-2 + 2e})- (-3 + 2) = ~.3064 + 1 = .6936

34

Chapter 7

7-3.

P(.! 0 :5: X :5: .60) = ff(X)dx = t25xdx + Ll.5625{l- x)dx

2 j

= l2.5x21- -.78125(I-dl

.ll

= .375 + .375 = .75

7-4. (a) f ~ f(x)dx = r"'_a -2 = atan-'(x)l~ = a(11/2)

-~ 10 1 + X D

For this to be a probability density function, a(l1l2) must equal I, so a = 21n.

(b) P(X s 1) = (217t) fl___1__,..dx = (2/rr)[tan·1(l) - tan·1(0)] ijl+x-

= (2/11)(11/4 - 0) = 112

7-5. For the density function in Exercise 7-1, F(x) = .75xl + .25x.

To find xp we need to solve F(x) = p.

For p = .25 we have .75x2 + .25x = .25, or 3Xl + x = 1. The positive solution of this equation is x = ,4343 = X25.

For p = 50 we have .75x2 + .25x = .5, or 3x2 + x = 2. The positive solution of this equation is x = 2/3 = X;o·

For P = .75 we have .75x2 + .25x = .75, or 3X2 + X = 3. The positive solution of the equation is x = .8471 = x".

7-6. If f{x) = e' for 0 s x s In2, then

F(x) = f: edt = eX - l.

(a) Solving F(x) = p when p = .50 we get eX _ I = .50 , or e" = 1.5.

The solution of this equation is x = In1.5 = ,4055 = X50· Solving F(x) = p when p = .90 we get

eX _ 1 = .90 ,or eX = 1.9.

The solution of this equation is x = In 1.9 = .6419 = X90·

Chapter 7

35

(b)

The mode of this distribution is the value of x for which the density function, f(x), is a maximum. Since e" is increasing, the maximum occurs at the right hand endpoint, x= In2 .

7-7. For the density function in Exercise 7-3, if 0 :5: x ~.2, then F(x) = foX25tdt = 12.5x 2, and F(.2) = .50.

Hence the median is .20.

FOL2 < x ~ I, F(x) = .5 + 1.5625 n 1- t)dt = .5 + 1.5625(t - t212)1:

= .5 + LS625(x - x2/2 - .18)

= .5 + 1.5625x - .78125x1 - .28125.

Solving F(x) = .8 we get

.78125x2 - 1.5625x + .58125 = O.

The solution to this equation in (.2, I) is x = ,4940 = x sc.

7-8. For the density function in Exercise 7-1:

E(X)= L~(l.5X+.25)dx= fo{ux2 +.25x)d.x

I

= (5x3 + .l25x2)lo = .625

E(X2) = fo~ 2 (Ux + ,25)dx = Lb .5x' + .25x 2 )dx = [(LS/4)x4 + (.25/3)x3ll~ = ,45833

V(X) = E(X1) - E(X)l = ,45833 - .6252 = .0677

7-9. E(X)= L2X(25x)dX+ J~ 1.5625x(l-x)dx

L225X'dx + 1.5625 J~(X - X2 )dx = (25/3 )X3 I: + 1.5625( xl/2 - X313) 112

- -

= .0666 + .2333 = JO

7-10. For the density function f(x) = .75(1 - X2) if -1 ::; x s 1 and 0 elsewhere,

E(X)= .75f\(l-x2)dx~.75(x2/2-x4/4) II =0.

-I -I

F(x) = (75(1- t1 )dt = .75(t - tl/3) [ = .75x - .25xJ + .50

To find the median we have to solve F(x) = .50.

.75x- .25x3 + .50 = .50 .25(3x - x') = 0

The only solution in [-1, I] is 0, so the median is 0, and the mean and the median are the same.

itv function fl 2",

7-1 L For the density nction x) = 2 lor x z 0,

11"(1 + X )

E(X) = ~ f~~x = (lI7t)ln(l + xl)I~·

11" 01 + x 0

This does not have a finite value, so E(X) does not exist.

r

CHAPTERS

8-1. For the uniform distribution on [a, b], f(x) = I/(b - a).

Ib' b3 3

E{Xl) = ~x =~ = (l/3)(b1 + ab + a2)

,b-a 3(b-a)

VeX) = E(x2) - E(Xi

= (1/3)(b2 + ab + a2) - (1/4)(b2 + 2ab + a2) = (b1 _ 2ab + i)/12 = (b - a)11J2

8-2. 1fT is uniformly distributed on [0, I 00]:

E(I) = (I 00 + 0)/2 = 50

VCI) = (l00 - oill2 = 833.33

8-3. Let I be the time (in minutes from 5:00) that the baby is born; then I is uniformly distributed on [0,60].

Ftt) = PCI ::; t) = t/60

PCl5 ::; T s 25) = F(25) - F(l5) = 25/60 - 15/60 = 116

8-4. Since X is uniform on [-.50, .50],

F{x)= x-(-.50) =x+.50. .50 - (-.50)

(a) P(-.lOsX S .20) = F(.20)-F(-.10) = .70- .40=.30

(b) VeX) = [.50 -(-.50)]2/12 = 1/12

38

8-5. (a) T is uniform on (35,50), so

E(T) = (50 + 35)12 = 42.5 V(T)=(SO-3SiI12= 152/12= 18.75

ChapterS

(b) The time at which 60 percent will be finished, t60, is the solution of F(t) = .60.

F(t) = (t - 35)/15, so we solve (1-35)115 = .60

t = 44.

8-6.

For y in [c,d] and T uniform on [a,b]:

P(Y':: y)=p(TsylcsTsb) = P(c:OS;T:OS;y) P(c'::Tsd) (y-c)j

= leo-a) = y-c

(d-c)j d-c

1(0 -a)

Hence Y has a uniform distribution over [c.d],

8-7.

(a) Since T is uniform on [40,100) we have

E(T) = (100 + 40)/2 = 70 VeT) = (l00 - 40)2/12 = 300

(b) peT> 57)=] - P(T::; 57) = 1 -17/60= .7167

r

;

Chapter B

8-8.

39

Let X and Y be the random variables of the ages at time of death for the 45-year-old and the 50-year-old, respectively. Then X is uniform on [45,100) and Y is uniform on [50,100].

The probability that the 45-year-old dies in the next 20 years is

peX::; 65) = (65 - 45)1(100 - 45) = 20/55 = 4111.

The probability that the 50-year-old dies in the next 20 years is

P(Y .:: 70) = 20/50 = 215.

(a) peX > 65 and Y> 70) = (1 -4111)(1- 2/5) = .3818 (b) peX::; 65 and Y :s 70) = (41l1 )(2/5) = .1455

8-9.

T has an exponential distribution with mean I Ii, = 500, or )._= .002.

P(T::; t) = F(t) = 1 _ e-002' peT > t) = Set) = e -0021

(a) P(T.:: 300) = 1 - e-·OO2(300) = 1 - e-·6 = .4512

(b) Pt'I > 900)= e-·002(900)=e-18 = .1653

8-10.

To find the median of the exponential distribution we solve

F(t} = 1 - e-1J = 50, or e-)·t = .50.

Then -/,t = In(.5) = -ln2. and the median is (lIA)ln2.

8-11. For the exponential distribution with mean 60, /, = 1160.

(a) P(T:os; 50) = J - e-50i60 = .5654

(b) peT> 100) = e-100!60 = .1889

40

Chapter 8

8·12. For the uniform distribution on [a,b],

F(t) = If(b - a) and Set) = (b -t)l(b - a).

The failure rate is

A(t) = f(t) ~_1_ 0- b -t =_1_.

Set) b - a b - a b - t

8-13.

Let T be the random variable of the incubation period. Tis exponential with A = 1/38.

(a) p(r s 25) = F(25) = I - e-2li38 = .4821

(b) P(T > 30) = S(30) = e-JO!38 = .4541

8-14.

Let T have an exponential distribution with parameter /... Then E{T) = IIA.

F(t) = l-e-)·I

F[E(T)] = F( 1IA) = 1 - e-'(Ii1·l = I - el '" .632

8-15. Let X and Y be the random variables for the time between accidents at these intersections. Each is exponentially distributed with parameters 2 and 2.5, respectively. Let SI and S2 be their respective survival functions. No accident in a month at an intersection means X (or Y) > I .

(a) (b)

P(X> lorY> 1)=SI(l)+Sll)-Sll)S2(l)

= e-2 + e-2l _ (e-2)(e-2S) = .2063

r

Chapter 8

8-16. If T has an exponential distribution with A = .15, then F(t) = I - e-15t• To find ~, solve F(t) = p.

If p = .25, 1 - e 151 = .25 e-·15t = .75 -.ISt= In.75 t25 = 1.9179

Ifp=.75,

1- C-15t = .75

e" III = .25

-.ISt = In.25 t,5 = 9.2420

8-17.

nn) = L~x n-Ie-xdx (See Equation 8.8)

let u = xo-I dv = e xdx

du = (n-I )x"·ldx v = --e-x

[(0)= _e-xxn-II: +(n-I)S;Xn-2e-xdx

If n > I, then - e -'x n-I I: = O. (See Equation 8.6) r(n) = (n -1) So\(n-Il-Ie-Xdx

= (n -1)[(n-l)

8-18. 1fT is an exponential random variable with parameter I"~ then P(T;::a + b)

P(T;:: a + b] T;:: a) = P(T;:: a)

= e -A, (a-i--b) ;;::;; e-Ab e-J...a

=P(T;:: b).

41

42

Chapter 8

8-19. For the population in Exercise 8-11, T (the time until death) was exponential with A = 1/60.

(a)

P(T.::: 50 IT:;. 40) = 1- peT:;. 501T:;. 40)

= 1-P(T::'10)= l_e-loi6o= .1535

(b) P(T::' 1001 T;:: 40) = P(T;:: 60) = e""""%o = .3679

8-20. For the gamma distribution f(x) = L xa-1e-/l' rea)

and SO> n -axd _ ro. + I)

oX e x~~.

8-21.

E(X2) = f~x 2f(x)dx = Lj"'xa+le-fI<dx

o f(a) 0

= L.1(a+2) (a + l)l(a+l)

rca) fJu+2 1(a)fJ2

= (ex. + l)a/13"

VeX) = E(x2) - E(Xi = a(a + 1)1132 - u2/131 = a/132

8-22 If the accidents occur at the rate of 2.5 per month, the waiting time between accidents is exponential with 13 = 2.5. The waiting time from the beginning of observation until the third accident, T, has a gamma distribution with a = 3 and 13 = 2.5.

E(T) = a/l3 = 3/2.5 = 1.2 VCT) = a/132 = 3/C2.5i = .48

r

Chapter 8

43

8-23. The waiting time T until the 12th new employee is hired has a gamma distribution with a = 12 and 13 = 8.

E(T) = alJ3 = 12/8 = 1.5 VeT) = a/132 = 12/64 = .1875

8-24. Let T be a gamma distribution with a mean of 18 and a variance of27.

E(T) = a I f3 = f3 = 18/27 = 2/3 VeT) a I f32

a = I3E(T) = (2/3)18 = 12

8-25. Let X be a gamma distribution with u = 2 and 13 = 3.

Then f(x) = ic~) x 2-1e-3x .

(a)

F(x) = 9 J \e -!tdt (integrating by parts) {)

(b) (c)

P(O::'; X::; 3) = F(3) = 1 - e-~(lO) = .9988

P(l <;X";2)=F(2)-F(I)=(l-7e"""")-(1-4e 1) = .1818

8-26. X has a gamma distribution with a = 2 and 13 = 1/3.

E(X) = 6, and EeXl) = a( o. t 1 )/132 = 54 E(C) = E(500X + 50X2)

= J~(500X + SOx 2 )f(x)dx o

= 500 C xf(x)dx + 50 r x 2 f(x)dx = 500ECX) + 50E(X2) = 3270

44

Chapter 8

8·27. (a) Fz(1.56)-Fz(-I.l5)=.9406-.1251=.8155

(b) Fz(2.13)-Fz(0.15)=.9834-.5596= .4238

(e) Fz(1.0)-Fz(-1.0)=.8413-.l587=.6826

(d) 1 - Fz( 1.65) + Fz(- 1.65) = 1- .9505 + .0495 = .0990

8·28. (a) Fz(z) = .8238, so z = 0.93

(b) fz(z) = .0287, so Z =- 1.90

(e) I - Fz(z) = .9115, so Fz(z) = .0885 and z = - 1.35

(d) 1- Fz(z) = .1660, so Fz(z) = .8340 and z = 0.97

(e) By symmetry of the standard normal density function, P(Z ::-. z) = P(Z s - z).

P(IZI ::-. z) = P(Z :0: z) + P(Z s - z) = 2P(Z ::-. z) = 2(l-Fz(z»= .10

Fz(z) = .9500, so z = 1.645

(Note that .9500 falls between two z-values, and we haven't done any interpolation. But this is a very important z-value to know.)

(f) P(IZI s z) = I - P(IZI;:> z) = I - 2(1- Fz(z» = 2Fz(z) - I = .95

Fz(z) = .9750, so z = 1.96

8·29. Let z > 0 and Fz(z) = a = peZ::; z). By symmetry of standard density normal density function:

i. Fz(- z) =P(Z ::;-z) = peZ ::-.z)= I-P(Zs z)= [- a

P(-zs Z::; z) = P(Z::; z) -P(Z-::;-z)=a- (I-a) = 2(1- [

"I I

Chapter 8

45

In subsequent problems involving probabilities of the normal random variable, the z-values will be rounded to 2 decimal places to make use of the z-table in Appendix A. If you are using the TI·83 or other calculator, this will not be necessary. The solutions here will only be those using the z-table.

8·30. X is a normal random variable with 11 = 17.1 and a = 3.2.

(14 - .u 25 - u I

P(14::;Xs25)= ~(T-<;Z::;-O'-)

=(14-17.7 sZ::;:25-17.1\

3.2 3.2 /

= (- .97 <; Z::; 2.47) = .9932 - .1660 = .8272

8.31 Let S be the total claims on the 5000 policies. Then S has a normal distribution with 11 = 5000(495) and 132 = 5000(30,000) and (J = [2,247.44.

P S < 2500000) = p(Z < 2,500,000 - 2,475,000J

( -" - 12,247.44

= P(Z s 2.04) = .9793

8·32. Let X be the length of the rod. X is normally distributed with mean of 7.505 and standard deviation of .01.

(a) P(7.48 ::; X ::; 7.52) = P C.4S ~ ;-505 s Z < 752 ~ ;-505) = P{- 2.5 ::; Z .s 1.5) = .9270

(b) If 4 rods, let Y be the number that meet the specifications.

P(Y::-' 3) = 4(.927)\073) + (.927)4 = .9711

!;

"

46

Chapter B

8-33. Let X be the lifetime ofa light bulb, and 11= 1500 and c = 125.

(a) peX 2 1400) = p( Z > 140~~; 500)

= P(Z 2- .8)= 1 -pez $- .8) = .7881

(b) P(all three bumingafter 1400 hrs) = (.7881/ = .4895

8-34. If X is a number picked from [0,1], x is uniformly distributed with mean .5 and variance 1112. Let S be the sum of 50 such numbers. Then S is approximately normal with Il = 50(.5) = 25

and rr = .JSOIl2 =2.0412.

P(24 < S <27)= p(24-2S s:Z::; 27-25)

- - 2.0412 2.0412

= P(-.49::; Z::; .98) = .5244

8-35. P(XS29.9)=.9192

P(Z s (29.9 -25)/0) = .9[92 = Fz(I.4) 4.9/cr = 1.4, a = 3.5

8-36. Let Y = eX , where X is normal with Il = 5 and rr = .40.

E(Y) = ei'+'f = e5.08 = 160.77

V(Y) = E(y)2( e" -1) = elOl\e·16 - I) = 4,484.96

8-37. Let Y = eX , where X is normal with 11 = 5.2 and cr = .80.

P(lOO s Y s: 500) = P(ln 100 S X::; In500)

= P (In I 00 - 5.2 s X ::; InSOO - 5.2)

.8 .8

= P(- .74 s X s 1.27) = .6684

,

r

Chapter B

8-38. Let Y = eX, where X is normal with 11 = 6.8 and c = .60.

P(Y 2 1750) = I - P(X s In1750) = 1 - P(X s: ln1750)

= I - P(Z s: (in I 750 - 6.8)/.6) = I - P(Z sUI) = .1335

8-39. To find the median ofY we solve the equation P(YSy)=.50

P(X s: lny) = .50

peZ s: (Iny - 11)/0) = .50 (Iny - Il)/cr = 0

lny = Il.

Then y = c'' is the median.

8-40.

Y = 100ex , where X is normal with 11 = .10 an 0' = .03.

(a) P(lOOY2112.50)=I-P(100YSlI2.S)

= I-P(Ys 1.125)

= I - pex::; InU25)

= 1- P{Z S (InU25 -.10)/.03) = 1 - P(Z s .59) = .2776

(b) P(IOOY":: 107.5) = P{Z'5 (lnl.075-.I}/.03)

= peZ s - .92) = .1788

8-41.

Let Y be lognormal with E(Y) = 2,SOO and V(Y) = 1,000,000. Note that V(Y) can be written as

V(Y)=E(Y/(e'" -I).

e" - 1 = 1,000,000/(2,500/ = .16 cr2 = In1.l6

(J = .3853

In(E(Y)) = 11 + 0'2/2 = ln2500

11 = In2500 - (.3853iI2 = 7.7498

47

48

Chapter B

8-42. Rewrite the Pareto density function as f(x) = ap"x-u-I.

(a) F(x) = afJ" J\-a-ldt = af)" _c:_ I'

p (-a) P

= _P"(x-O: - p-«) = ! - (pixt

(b) E(X) = J:f(X)dx=ar fx-"dx

= afJ" x -,,+1 I~ = ~ (0 _ /3-a+l)

-a+! P -(a-I)

=aP/(a-I), (a> 1)

(c) EeX2) = L?f(x)dx =ar r;x.-a+1dx

= ap" x -"+21'" = aft" (0 _ /3-a+2)

-a+2 p -(a-2)

= ap2j(a - 2), (a> 2)

VeX) = E(X2) _ E(xi = a/l' _ ( afJ )2 a-2 a-l

8-43. Let X be a Pareto random variable with a = 3.5 and p = 4.

(a) E(X) = ap/(a - 1) = 4(3.5)/2.5 = 5.6

(b) VeX) = ap2/(a - 2) - [ap!(a - 1)]2 = 3.5(16)/1.5 - 5.62 = 5.9733

(c) To find the median solve F(x) = 1 - (4/x)35 = .50. (4/x)'5 = .50

x = 4/(5)1/35 = 4.8761

(d) P(6 s: X <;: 12) = F(l2) - F(6)

= [I - (4112/5]- [1- (4/6/5] = .97862 - .75808 = .22054

Chapter B

49

8-44. Because of the deductible, only claims for losses of more than 5 (hundreds of dollars) are filed. Hence p = 5. Since the failure rate is 3.5/x, a = 3.5.

(a) Mean loss amount is E(X) = (3.5)(5)/2.5 = 7. ($700)

(b) Expected amount of a single claim is loss - deductible = $200.

(c) V(IOOX) = 1002[(3.5)(25)/1.5 - i] = 93,333.33

8-45.

We are given [(112) = nli2, and we know F(x t I) = xl (x).

(a) 1(3/2) = (1/2)[(1/2) = (1/2)nl/2 (b) [(5/2) = (3/2)[(312) = (3/4)nli2 (c) [(712) = (5/2)[(5/2) = (IS/8)nl12

8-46. For the Weibull distribution F(x) = 1- e·/k" .

If a = 3 and p = 3.S, then F(x) = 1- e-J5.,2 .

(a) P(X <; 0.4) = F(O.4) = l_e-J5(4i' = .2007

(b) P(X>0.8)= I-F(0.8)= e-15(S)' =.1666

8-47. The failure rate for the Weibull distribution is A(x) = ar,lx,,·I.

If 0: = 3 and 13 = 3.5, then Aex) = 10.Sx2

50

Chapter B

8-48. Let X be a Weibull distribution with a = 2 and ~ ~ 3.5.

(a)

E(X) = i(l + 11 a) ~ ['(3/2) ~ (l/2)Jrln = 4737

lila Fs J3.5 .

(b)

VeX) = -21-[[,(1 + 21 a) - I'(l + II a)2] fJ ra

= (113.5)[i(2)-[,(3/2il = (1/3.5)(1- rc14) = .0613

(e)

P(.25 s X $ .75) = (I - e-J·5(.75l') - (1 ~ e-J5(25)') = (1-.13963) - (I - .80352)

= .66389

8-49. The Weibull density function is f(x) ~ ajJxu-le-fJxa .

E(X) = L~ xf(x)dx ~ afJ f~x· e -fJx" dx

Let u = XU X = uli" dx = (lIa)uI/a-ldu

xae-fJxa dx =ue-""'(1la)uI1a-ldu =(lla)ul'ue-fJl<du

E(X)=ajJr~(lla)uIlUe-P"du=,Br(l+lIa) = [(l+lIa)

Jo pL'u+1 s":

8-50. Let X be a beta distribution with a = 4 and ~ = 1.5.

f()= r(a+jJ) xu-I(I-X)P-I ~ ['(5.5) x3(l_X)5

x rca)rCjJ) 1(4)[(1.5)

= (9/2)(7/2)(5/2)(3/2)[(312) x3 (1- x) 5 (31)[(3/2)

= 315x'(l _x)L'2/32

Chapter B

rl 4 2

8-51. We need to find k so that Jokx (1- x) dx ~ I.

(I 4 5 6 c 6 - II J/x -2x +-X )dx=(x"/5-2x /6+x/7) 0

= 1/5 -]/3 + 117 = 1/]05

Hence k = ]05.

8·52. For the beta distribution with a = 3 and f3 = 2:

E(X) = a/(a + 13) = 3/6 = 0.6

VeX) = aj3 = _6 _ = 0.04

(a + f3)2 (a + f3 + 1) 25(6)

[(5) IS J'

PCX$O.5)= [(3)[(2) ox2(I-x)dx=12 o"cX2 -·x3)dx

= 12(x)/3 - x4/4)I: = 12[(1/3)(.5l- (l/4)(.5)4J =.3125

8-53.

8-54.

For a beta distribution with a = 2 and ~ = 4:

P(xsOJ)= ['(~i~4) t3x(l-X)'dx

= 20 L3(x-3X' +3x3 -x4)dx = 20(X2/2 - x ' + 3x4/4 _X5/5)1:

= 30[(1I2)(.3i -(.3)' + (3/4)(.3)4 - (1/5)(.3l1 = .47178

51

'j:

I I

J

52

Chapter 8

8-55.

Let X have a beta distribution with parameters a and 13.

E(X) = rixf(x)dx = qa + fJ) fix" (1 _ X)p-l dx

Jo 1(a )r(fJ) J 0

(and using Equation 8.3 3)

= na+jJ) .i(a+l)np) a

na)r(,B) r(a+ft+l) a+p

CHAPTER 9

9-1.

With a deductible of $300, the claim amount is g(x) = x - 300. for x 2: 300, and 0 elsewhere.

E[g(X)] = (g(x)(.OOle-oolx )dx I:o(x - 300)(.001 e -DOh )dx

fro xC.OOle-OOlx )dx - 300 Jro (.OOle-OOh )dx

soc .100

(the first integral uses integration by parts)

= [(_xe-OOlx _ 1000e-DOh) + 300e-OOI'] I~

.,00

= (-x - 700)e .. 00Ix I~ = 1000e-3 300

= 740.82

9-2.

If the policy in Exercise 9-1 also has a payment cap of $1500, then

\0,0< x <300

hex) = x - 300,300"-: x s 1800 .

ISOO.x>1800

E[h(X)] = (h(x)(.00Ie-001x )dx

Jl800 J~

= (x_300)(.00Ie-ooh)dx+ ISOO(.OOle-001X)dx

100 1800

= [(-x _ 700)eOO1X ]11800 -iSOOe -.Olll, lro

300 1800

= -2500e-ls + 1000e-3 + 1500e·1 S = 575.52

:1 \[

':i

54

Chapter 9

9-3.

Using the utility function u(w) = In(w), the expected utilities are as follows:

Method l : WI is uniformly distributed over [9, II], so f(x) = 112 for 9 S; w S; 11.

fll 11

E[u(WI)] = 9 (1/2)ln(w)dw ~ ~1/2)[w[n(w) ~ W]19

= (1/2)(1 lln11 - 91n9 - 2) = 2.3009 Method 2: W2 is uniformly distributed over [5,15], so

fex) = 1110 for 5 S; w S; 15.

JI5

E[u(WJJ = }IIlO)ln(w)dw

= (l/IO)[W[n(W)-wJC

=(1I10)(l5 [nI5-5In5 -10)=2.2574

Hence Method I has higher expected utility and should be chosen.

9-4.

If X is uniformly distributed over [a, b], then f(x) = l/(b - a).

I Ib tx b bt "

M (t)=E[e'x]= __ eIXdx=_e __ 1 = ~

x b - a a t(b - a) a t(b - a)

9-5.

For the moment generating function in Exercise 9-4,

r t(bebl - ae" ) - (e " - ea,)

MX(t)= 2

(b - a)t

As t approaches 0, the expression assumes an indeterminate form, so we must use L'Hospital's rule.

M' 0 _. (be br - ae at) + t(b 2 e bt ~ a 2 e at ) - (be bt + ae " )

xC )-!~ 2t(b-a)

_ r blebl ~a2eal = b1 _a2 = a + b =E[X]

- t~ 2(b - a) 2(b - a) 2

Note that in the case of the uniform distribution, it is inefficient to find moments using Mx(t).

Chapter 9

55

9-6.

If f(x) = 2(] - x) for 0 S; x ::; I and 0 elsewhere, then

JI JI I

= 2 oe"'dx-2 /e'dx

)1

tx tx I;\.

= 2(L_'£"'_+~

t t t· 0

= (2el - 2t - 2)/t2 , for t '" 0, and I if t = o.

9-7.

For the moment generating function in Exercise 9-6,

t t2(2e' -2)-2t(2e' -2t-2)

Mx(t)= t4

21e I _ 4e' + 21 + 4 t3

As t approaches 0, the expression assumes an indeterminate form, so we must use L'Hospitals's rule.

9-8.

The moment generating function for the gamma distribution with parameters a and 13 is

, fJ )"

Mx(t) =lfJ- t

Thus (_2~)S is the moment generating function for the gamma 2 -t

distribution with a = 5 and 13 = 2.

56

Chapter 9

9-9. The moment generating function for the exponential distribution with 'A = 3 is

Mx(t) = (~3 ) 3-t '

so M2X+I(t) = e5'Mx(2t) = el' (_3 _). 3 - 2t

9-10. Let X be the random variable with moment generating function M x(t) = e'+"

M~ (t) = (I + 2t)e 1+1' M~ (0) = 1 = E(X)

M~ (t) = (I + 2t)2 et+l' + 2e'+t1 M~ (0) = I + 2 = 3 = E(x2)

VeX) = E(X2) - E(X? = 2

9-11.

",112

For the normal distribution, M x (t) = e %+-2-

M~ (t) = (11 + O'1t)Mx (t)

M~(t) =(11 + 0'21)2 Mx (t) + 0'2 Mx (I)

M~ (0) =;/ + a" = E(X2). (Recall Mx(O) = 1)

VeX) = E(X2) - E(xi = J.l" + 0-2 _,112 = 0-"

9-12. [fX is uniformly distributed over [0,1], then F(x) = x, 0 S x s: 1.

If Y = eX = g(X), then X = InY = heY). Since 0 s x s 1, then 1 sex s e.

(a) Fy(y) = Fx[h(y)] = In y, for I s y s e

(b) fy(y) = F~(y)= d~vlny=I/Y, for I syse

Chapter 9

57

9.13. If fx(x) = 3x-4, then Fx(x) = 1 - x-J for x > 1. If Y = g(x) = lnx, then x = hey) = e' for y > O.

Fy(y) = fx[h(y)] = I - (ey)·3 = 1 - e-3y, for y:'" O.

9-14. Iffx(x) = h-4 then Fx(x) = I - x-3 and Sx = x-1 for x > I. If Y = g(x) = I/x, then x = hey) = I/y for 0 < y s I. Note that g(x) is strictly decreas in g for x :,.. 1.

(a) fyCy) = Sx[h(y)] = (l/y)-J = y', for 0 < Y <; I (b) fy(y) = F' y(y) = 31, for 0 < Y s 1

9-15. Let X be exponentially distributed with unknown parameter 1-., and let P(X > 100) = e-IOO}. = .64. Let Y = 2X.

P(Y> 100) = P(2X > 100) = P(X > 50) = e'SO). = .,[64 = .80.

9-16. P[F(x) s x] = P[X S FI(X)] = F[F'I(x)] = x

58

Chapter 9

i'" .I

The t_able below gives random numbers used to simulate F(x) in Exercises 9-17 and 9-18. The second and third columns give the transformed values x = F \ u) for the distributions in 9- 17 and 9-18 respectively. '

**Trial Random No. (u) I F (u)~4u F-1(u) = 31\11 - u
**

1 .90643 \ 3.6257 6.6081

2 .17842 0.7137 3.2031

3 .55660 2.2264 3.9342

4 .55071 2.2028 3.9169

5 .96216 3.8486 8.9359

6 .81008 3.2403 5.2191

7 .49660 1.9864 3.7712

8 .92602 3.7041 7.1463

9 .71729 2.8692 4.5710

10 .39443 1.5777 3.5460

II .15533 0.6213 3.1736

12 .29701 1.1880 3.3739

13 .82751 3.3100 5.3893

14 .67490 2.6996 4.3630

15 .68556 2.7422 4.4117

16 .31329 1.2532 3.4004

17 .68995 2.7598 4.4325

18 .77787 3.1115 4.9536

19 .66928 2.6771 4.3381

20 .53100 2.1240 3.8613 9-17. If X is uniformly distributed over [0,4J, then u = F(x) = x!4, for o :> x s 4. So x = F-1(u) = 4u, for 0 :s; u s l. Using the above table you can count the number of values of x in each subinterval.

II

9-18. If X has a Pareto distribution with a = 3 and 13 = 3, then u = F(x) = I - (3/X)3, for x ;;: 3. So x = F-1(u) = _3_, for \l1-u'

o :s; u < I. Using the above table you count the number of values in eaeh subinterval,

Chapter 9

59

9-19. Let S be the event that that a minor claim is made, and L be the event that a major claim is made. Then peS) = .09 and P(L) =.01.

Case I: x = O. This represents those who filed no claims, so F(O) = .90.

Case 2: 0 < x :> 1000. Since 9 percent filed minor claims and these claims are uniformly distributed over this range, P(X s x) = P(O) + pco < X s 1000)P(S)

= .90 + .09(xll 000).

Case 3: 1,000 < x:s; 10,000. Since 1 percent filed major claims and these claims are uniformly distributed over this range,

P{X s x) = P(X s 1.000) + P(lOOO < X::; 10,000)P(L) = .99 + .01 (x - 1000)19000.

9-20. For the policy in Exercise 9-19, f(x) = .09/1000, for 0 < x s; 1000, and f{x) = .0119000, for 1000 < x ::; 10,000,

J I 000 J 10000

E(X) = x(.091l000)dx + x(.01l9000)dx

o 1000

= .09 (£_ I. 1.000 + _,Ql_(£_)10.000 = 45 + 55 = 100

1000 2) 0 9000 2 1.00D

9-21. Let S be the event that a claim is filed. Pt S) = .10. Let Y be the random variable for the amount of the claim filed. Y has a Pareto distribution with a = 3 and 13 = 200.

P(Y s y) = I ~ (200/yi for y > 200

Let X be the random variable of the amount paid, X = Y - 200,

P(X :s; x I S) = 1 ~ (_1QQ__)3 , for x > 0

200+x

Case I: x = 0. No claim is filed, so F(O) = .90.

Case 2: x > O. P(X s x) = P(X = 0) + P(O < X s x I S)P(S)

= .90 + .10[1- ( ___1(lQ_IJ 1

\200+x)

60

Chapter 9

9-22.

Th . I functi -fA(I)dl

e surviva unction Sex) = eO. IfA(t) = 2/(1 + x], then

Ix 2

~I t = 21nO + x) = In(l + X)2

o + t .

e -2In(ln) = ~I- = Sex), for x ::-: a

(l + X)2

9-23. If the hazard rate is "-(x) = 1/(100 - x), for a <; x < 100, then

- L'lOd _ t dt = In( lOa - t)l: = In(100 - x) -In(1 00) = In(IOO-x)

100 .

Sex) = (100 - x)ll 00, for a <; x < [00

9-24. IfS(x) = lI(1 + xi, for x z 0, then

E(X) = fS(x)dx = fro ~x = -=l._ I~ = 1.

o D(I+x/ [+X 0

9-25. IfS(x) = (loa -x)/IOO, for 0:; x< 100, then

E(X) = flOor 00 - xdx = (100 - X)21100 =

Jo 100 200 0 50.

r

CHAPTER 10

10-1. The values of ptx.y) can be found by direct substitution. e.g. p(I,I)=P(X= I,Y= 1)=[(1)(1)+ 1]/27=2/27.

The marginal probabilities can be found by adding the entries in the columns to get pex) and the row entries to get p(y).

P(X = I) = p(l,I) + p(I,2) = 2/27 + 4/27 = 2/9

P(Y = I) = pC I, I) + p(2,) + p(3, I) = 2/27 +3/27 + 4/27 = 1/3

10-2. Since the team consists of 2 professionals the possible values of X and Yare 0:; X + Y :; 2, (2 CPA's and I actuary would be impossible). The event (X = x, Y = y), denotes a team of x CPA's, y actuaries, and 2 - (x + y) accountants.

r(x = O,Y = 0) = p(O,O) = C(5,0)C(3,0)C(2,2)/C(10.2) = 1/45 P(X = O.Y = I) = p(O,I) = C(5,O)C(3,1 )C(2, I )/45 = 6/45

P(X = I,Y = 0)= p(1,O) = C(5, l)C(3,O)C(2, 1)/45 = 10/45

PrX = I,Y = 1)= p(!,l) = C(5,1)C(3,I)C(2,0)/45 = 15/45, etc.

The marginal probabilities can be found by obtaining row sums and column sums as in 10-1.

P(X = 0) = p(O,O) + p(O,I) + p(0,2) = )/45 + 6/45 + 3/45 = 10/45

10-3. Using the table obtained in Exercise 10-1:

E(X) = Lxpxex) =2/9+2(3/9)+3(419)=20/9

E(Y) = LYPv (y) = (1/3) + 2(2/3) = 5/3

62

Chapter 10

10-4. Using the table obtained in Exercise 10-2:

E(X) = L xPx (x) = 0 + 25/45 + 2(10/45) = 1

ECY) = LYPy (y) = 0 + 21/45 + 2(3/45) = 27145 = 315

10-5. Using the table from Exercise 10-2:

;r.i

''l' EeX2) = Z> 2 p X (X) = 0 + 12(25/45) + 22(10/45)

= 65/45 = 13/9

i '1"

VeX) = EeX2) - E(xi = 13/9 - I = 419

Ecy2) = Ly2py(y)=O+ 1\21/45)+22(3/45) =33/45= 11115

V(Y) = 11115 - (3/5/ = (55 - 27)/75 = 28175

10-6. Clearly the function f(x,v) = 1 + ~ + ~ + xy ,0 < x < I and

\- 422 --

o :;; y <:; 1, is greater than or equal to 0 on this range, so we only need to show that the volume under the surface is I.

II I

rlfl (l+~+~+xY)dXdY= II (~+2C_+ xy +~)I dy

Jo 0 4 2 2 0 4 4 2 2

X"'O

= DI/2 + y)dy

= (y/2 + YI2)1~ = 1

.1

Chapter 10

63

Therefore f(x,y) is a joint probability density function.

I Ii' (1 Y }

P(O s X.:: 112, 1/2 s Y s 1) = I I· ~ + ;.;_ + ~ + xy xdy

li2 0 4 2 x

I (' 2 ,'IV)

X x : xy x"

= I ~+-+-+--) dv

li2 4 4 2 2 -

x=o

~ II (2.. + 3yt"

li2 [6 8 r-

( ·1

3y 3y2 J .

= 1- + - .' = 15/64

\ 16 16 1.2

10-7.

OS;xSI

=(II2+y), O<:;y<:; 1

10-8.

The density function f(x,y) = 2X2 + 3y is defined on the region bounded by the x-axis and the lines y = x and x = 1.

(a) tx(x)= f:(2X2+3Y)dY=(2X2y+3~2Lo

= 2x' + 3x1(2, o s x s I

(b) fy(y) = fC2x2 +3Y)dx=(2~] +3XY)~~}

= 2/3 + 3y - 3/ - (2/3)/ . 0 s y s I

64 Chapter 10

10-9, Using the marginal distributions from Exercise 10-8:

(a)

P(X> 1/2) = (, PX 3 + 3x 212)dx

(4 J)I

= I _!_ + l_ = 29/32 \. 2 2 112

P(Y> 1/2)= fl (2/3+3y-3y2 -2yJ!3)dy

112

( ') 3 2 4 Jl

=l~; ++-Y'-Y6 U2 =41/96

(b)

10-10, From Exercise 10-6, fx(x) = Y, + x for 0 :0; x :0; 1.

I I (2 3 JI

E(X) = Ix(l/2+ x)dx = fCxl2+x2)dx= 1L_+L =71l2

o 0 4 3 D

10-11. Iff(x,y) = 1/4 + x12 + y/2 + xy, for 0:0; x s 1 and 0 S Y S 1, then

P(X> Y) = f~L'(li4+ x12 + y/2+ xy)dydx

Ii

'I

Chapter 10

10-12, From Exercise 10-8, fx(x) =2Xl + 3x2/2, forO S x s L

II fl 4 ,

E(X)= /fx(x)dx= Jo(2x +3x'/2)dx

( l 4)1

= lL+k.- =31/40

5 8 0

Also fy(y) = 2/3 + 3y - 31- 2/13. for 0 :0; y $ 1.

10-13. f(x,y) = (30-x-y)/1875. forO$ x:O; 5 and 0:0; y s' 25.

P(X 2: 4. Y 2: 20) = (111875) 125 [(JO - x - y)dxdy

. w1

= (1/1875) f2~(30X - x 212 - xy)C dy

I25

= (1/3750) (51- 2y)dy

20

= (1I3750)(51y -l)I~: = 3013750 = 11125

f2r30-X-Y'~ (30Y-Xy-y2/2J25

10-14. (a)fx(x)= 0\ 1875 rY= 1875 >=0

( 2 J"l

= 60)' - 2x) -)

3750

J=D

= (35 - 2x)/150, for 0 :0; x S 50

875 - SOx 3750

65

66

Chapter 10

:-~ !

10-14. (b) fy(y) = f: CO;8;5-Y}x~(30X-1~;:-XYLo

=(60_X2_2XYI5 ~275-IOy

3750 I 3750

} .r--O

= (55 - 2y)/750, for 0.:: y s 25

10-15. For the random variables in Exercise 10-13:

E(X) = f: xf x (x)dx = (1/150) f:(35X - 2x 2 )dx

= C~0)(35{ _2~)J: 291026=1~

f~ f~

E(Y)= oyfy(y)dy=(1!750) 0 (55y-2y2)dy

_ ( 1 )(55y2 2y'i2l 40,625 __ 325 - 750 -2--3)0 6·750 36

10-16. Using the data from the table obtained in Exercise 10-1: peX = 11Y = 1) = p(l, J ) ~ 2/27 = 2/9

py(l) 113

P(X = 21Y = I) = p(2,1) = 119 = 1/3 py(l) lIJ

P(X = JIY = I) = p(3,I) = 4/27 = 4/9

py(l) 113

10-17. Using the data from the table obtained in Exercise 10-1:

P(Y=IIX=J)= p(I,J)=2!27 =1/3 Px(l) 2/9

P{Y=2IX= 1)= p(I,2)=4/27 =2/3 Px(l) 2/9

Chapter 10

10-18. Using the data obtained in Exercise 10-16:

E(XIY = I) = L xP(X = x I Y = 1)

= 1(2/9) + 2(113) + 3(4/9) = 2019

10-19. For the joint density function in Exercise 10-6:

f(x, v) 1/4+xl2+y/2+xy

f(xly) = f y (;.) = 112 + Y

(l/2 + x)(l/2 + y) (1/2 + y) =1/2+x,0'::x'::l

J 0-20. For the joint density function in Exercise 10-8:

a f 112 = 2(1/2)2 +3y

10-21. () (yl ) 2(112)3 + (312)(l/2i

4+24y 0<v51/2

5 ' .

(b) E(YIX = 1/2) = f;/~tlY 11/2)dy

= (1/5) f~'2(4Y+24y2)dY

= (1I5){2Y + S/)C = 3110

67

68

Chapter 10

10-22. Let f\x,y) = 6x, for 0 < x < Y < 1, and 0 elsewhere. (a) fy(y)= f:f(x,y)dx= J:6xdx=3y2,foro<y<1

(b) f(x[y) = f(x,y) =~= 2x ,for a < x <y< I

fyCY) 3y2 y2

(c) E(XIY = y) = f Yxf(xl y)dx = ClIyl) J}ix 2dx =~. 2y3 = 2y/3

o 0 y 3

(d) E(XIY = 112) = 2(112)13 = 1/3

10-23. Let X and Y be the random variables in Exercise 10-1. From the table obtained in that exercise it can be seen that

p(x,y) = Px(x)py{y) for all possible choices ofx and y, e.g.,

;1

\ '\

Ii

il: :

[i

p(l,I) = 2127 = (2/9)(1/3) = px(1 )pv( I) p(2,I) = 119 = (If3)(1/3) = Px(2)py{I), etc.

Hence X and Y are independent.

10-24. Using the tables obtained for the random variables in Exercise 10-2,

p(O,O) = )/45"* (10/45)(21145) = px(O)Pv(O).

Hence X and Yare dependent.

10-25. In Exercise 10-6 the joint density function is

f(x,y) = 1/4 + xl2 + y/2 +xy = (1/2 + x)(112 + y) = fx(x)fy(y).

Hence X and Y are independent.

10-26. For the joint density function f(x,y) in Exercise 10-8, f(x,y) = 2X2 + 3y"* (2x3 + 3xl/2)(2/3 + 3y ~ 31 ~ 2y3/3) = rxex)fy{y).

Hence X and Yare dependent.

CHAPTER 11

II-I. If X and Yare the random variables in Exercise [0-1, the possible values of X are l , 2 and 3, and of Yare I and 2. The possible values of S = X + Yare 2, 3, 4 and 5. Since X and Y

are independent,

ps(s) = LPx(x)PY(s~x). ps(2) = px(l)py(l) = (219)(1/3) = 2/27

ps(3) = Px( I )py(2) + px(2 )py( 1 ) = (2/9 )(2/3 ) + (l13)( l/3} = 7/27 ps(4) = px(2)py(2) + px(3)py(l} = (113)(2/3) + (4/9)(1/3) = 10/27 ps(5) = px(3)py(2) = (4/9)(2/3) = 8127

11-2. lf Irx.y) = (4/3)([ ~ xy) , 0 S; x S; 1 and 0 s; y 5. I:

J1Jl-'

P(X + Y.-:: 1)= 0 0 f(x,y)dxdy

= J ~ C\ 4/3)(1 ~ xy)dxdy

51 l~v

= (4/3) o(x ~ xly/2)L~o dy

= (213) J~(2 ~ 3y + 2i - yJ )dy

= (2/3)(2y ~ 3il2 + 2y'/3 ~ lI4)1~ = [[/18

70 Chapter 11

11-3. Iffx(x)=2e-1x for x > 0 and fy(y) = 3e-Jy fory> 0, and s= X +Y:

fs(s) = f" fx(x)fy(s-x)dx= 6 f'e-2Xe-3('-X)dx

-" 0

= 6e-3s J:e'dx=6e-)'(e' -1) = 6(e-2, - e-3'), s > 0

11-4. From Example 11.3, f(x,y) = 2 - 1.2x - .8y, 0..; x s; 1,0"; y s I.

P(X + Y s 1.5) = I - P(X + Y > 1.5) P(X+Y> 1.5)= f.lfl (2-1.2x-.8y)dxdy

.!i 1_5-y

JI 2 II

px - .6x - .8xy) ,=15-,. dy

fl 1

J -.25 + .6y - .2y )d)'

2 )1 -

=(-.25)'+.3)' -.2/13) 5 = .04166

P(X + Y ..; 1.5) = 1 - .04166 = .95833

11-5. From Example 1 1.4, f(x,y) = e-{x+y), X;;.0,y20. IfS=X+Y:

Fs(s) = P(S"; s) = P( X + Y:s:: s)

f :1:-Ye -(x +y) dxdy

r'e -v (-e -x )1'-" dy

Jo >.:",,0

S:(e-Y -e-')dy

= (-e -Y - ye -s )!: = 1 - e -'(I + s) To compare fs(s) with Example 11.4,

Chapter 11

71

11-6. For the independent random variables in Exercise 10-6, we found that fx(x) = 112 + x, for 0 :s:: x s I, and fy(y) = 112 + y, for O'S)'::;!.

P(min(X,Y) > t) = P(X > t and Y > t) = peX > t)P(Y > t)

= (If 112+ X)dX)( HlI2+ y)dY)

(Note that both integrals give the same value)

= (xl2+X212)I:Y

= (I - tl2 - t212/

11-7. Using the table obtained in Exercise 10-1:

E(X + Y)= LL(x+ y)p(x,y)

y x

= (1 + 1)(2/27) + (2 + 1 )(3127) + (3 + 1 )(4127)

+ (1 + 2)(4/27) + (2 + 2)(6/27) + (2+3)(8/27) = 105/27 = 35/9

Using the probability distribution fs(s) from Exercise II-I:

E(S) = E(X + Y) = L sfs (s)

s

= 2(2/27) + 3(7/27) + 4(10/27) + 5(8/27) = 35/9

From Exercise 10-3, E(X) = 20/9 and E{Y) = 5/3 = 15/9, SO

E(X) + E(Y) = 20/9 + 15/9 = 35/9 = E(X + Y).

72 Chapter 11

11-8. For the joint density function fix.y) = (4/3)(1- xy), 0 sx s 1, and Osys 1:

E(X + Y) = JJ~x + y)f(x,y)dxdy

J1JI 2

=(4/3) 0 o(x+y-x y-xy2)dxdy

Ii , 3 It

= (4/3) o(x "/2 + xy - x y/3 - x 2y2/2) X~O dy

= (4/3) 10<1/2 + 2y/3 - y2/2)dy

= (4/3)(y/2 + il3 - y)/6)1~ = 16/18 = 819

fx (x) = f~f(x,Y)dY=(4/3) f~(1-xY)dY

= (4/3)(y - x/12)Lo = (4/3)(1 - x.l2)

E(X) = I~xfx (x)dx = (4/3) f~(X - x2/2)dx = (4/3)(x212 _xl/6)1: = 4/9

A similar calculation will show that E(Y) = 4/9, so E(X + Y) = E(X) + E(Y).

11-9. Let X and Y be random variables with joint density function f(x,y).

E(X + Y) = C[ex + y)f(x,y)dxdy

r- [f{x,y)dydX+ [Y L)<x,y)dXdy [xfx(X)dx+ [lfy(y)dy

= E(X) + E(Y)

Chapter 11

73

11,10. For the random variables (independent) in Example 11.9:

'j x

y 't

The only non-zero terms occur when both x '" 0 and y '" 0, i.e. when (x.y) = (1.1), (1,2), (2,1) or (2,2).

E(XY) = 1(1/4)(114) + 2 (114)(1/4) + 2 (1/4)( 1/4) + 4( 1/4)([;4) = 9/]6

II, II.

For the random variables in Exercise 11-8, f(x,y) = (4/3)( 1- xy), for 0 S x s 1 and 0 s y s; I.

J1il fiJI, 2

E(XY) = xyf(x,y)dxdy = (4/3) (xy - x <v )dxdy

o 0 0 0 •

(a)

II, "11

= (4/3) (x ' v/2 - x v 13 ) dv

o "" "" :0,;-0"

II 2 2 1 II

= (4/3) o(y/2 - y 13)dy = (4/3)(y 14 - y. (9) 0

= (413 )(5/36) = 5/27

(b) E(X) = E(Y) = 4/9 from Exercise 11-8, so EeX)E(Y) = 16/S1

(c) Cov(X, Y) = E(XY) - E(X)E(Y) = 5/27 - 16/81 = -1/8 I

11-12. For the random variables in Exercise 11-8:

(a) fx(x) = (4/3)(1 - x.l2), for a s x s 1

E(X2) = J~x 2f x(x)dx = (4/3) f~(x 2 - x'/2)dx

= (4/3)(x'/3 - x 4/S)I~ = (4/3)(5/24) = 511 8 VeX) = E(X2) - E(X/ = 5/18 - 161S1 = 131162

(b) A similar calculation will show V(Y) = 13/162.

(c) vex + Y) ~ V(X) + V(Y)""- 2Cov(X,Y)

= 13/162+ 131162-2/81 = II/SI

74

Chapter 11

11-13. For the random variables in Exercise 10-1 we have already shown (see Exercise 10-3) that E(X) = 20/9 and E(Y) = 5/3.

x

VeX) = E(XL) - E(X)2 = 50/9 - (20/9/ = 50181

Ecy2) = L:)2py(y) = [(1/3)+4(2/3)=3

y

V(Y) = E(Y") - E(y)2 = 3 - (5/3/ = 219

Since X and Yare independent (see Exercise 10-23),

VeX + Y) = VeX) + V(Y) = 50/81 + 2/9 = 68/81.

11-14. (a) E(X) = 1(.5) + 2(.5) = 1.5 (b) E(Y) = 1(.4) + 2(.6) = 1.6 (c) E(X2) = 1(.5) + 4(.5) = 2.5 V(X)= 2.5 - (1.5/ = .25

(d) E(y2) = 1(.4) + 4(.6) = 2.8 V(Y) = 2.8 - (1.6i = .24

,Ii! 1,.1'1.·.1.

"

(e) E(XY) = 1 (.15) + 2(.25) + 2(.35) + 4(.25) = 2.35 Cov(X,Y) = E(XY) - E(X)E(Y) = 2.35 - (1.5)(1.6) = -.05

(1) VeX + Y) = VeX) + V(Y) + 2Cov(X,Y) = .25 + .24 - .10 = .39

Chapter 11

11-15. For the random variables in Exercise 10-22, we have already shown fy(y) = 3y", for 0 s; y s; 1.

f x (x) = f:f(X)dy = 6 J:Xdy = 6x(l- x) , for 0 s; x s; 1

f I fl 2 J

E(X) = ox(6x)(1 - x)dx = 6 o(x - x )dx

= 6(x3/3 - x4/4)I: = 112 E(X2) = f~X2 (6x)(I- x)dx = 6 S:(X3 - X4 )dx

4 I II

= 6(x 14 - x (5) 0 = 3/10

E(Y)= r3yJdy=(3v4/4)11 =3/4

10 . 0

E(y2) = fl3v4dv = (3v' (5)1' = 3/5

o· ~ - II

(a) VeX) = 3110 - (l/2)2 = 1/20 (b) V(Y) = 3/5 - (3/4i = 3180

(c) E(XY) = fJ:xyf(X,Y)dxdY= J:f:6x2ydxdY

= rl(lx Jy)IY dy = Sl2y4dy

Jo x=O 0

I

= (2v5/5)1 = 2/5

• 0

(d) Cov(X,Y) = E(XY) - E(X}£(Y) = 2/5 - (1/2)(3/4) = 1/40 VeX + Y) = VeX) + V(Y) + 2Cov(X,Y)

= 1/20 + 3/80 + 2/40 = 11/80

75

76

Chapter 11

11-16. For the random variables ill Exercise 11-14, VeX) = .25, V(Y) = .24 and Cov(X,Y) = -.05. Hence

p _ Cov(X,Y) -.05 =-.2041

Xy - JV(X)V(Y) Je.25)(.24)

11-17. For the random variables in Exercise 11-15, VeX) = 1120, V(Y) = 3/80 and Cov(X,Y) = 1140. Hence

Cov(X, Y) 1140

PXY = = = .5774.

JV(X)V(Y) J( 1/20)(3/80)

11-18. Let X and Y be random variables whose joint density function is f(x,y) = x + y, for 0 <;: x <;: I and 0:::; y s I.

fx(x) = (f(X,y)dy= fo{x+y)dy=(xy+y2/2)1:;o

= x + 112, for 0 <;: x <;: I

II fl 2 l' /1

E(X) = /fx (x)dx = Jo(x + x(2)dx = (x 13 + x - 14) 0 = 7112

E(X2)= (x2fx(x)dX= nx3+X2/2)dx=(x414+x3/6)1~=5/12 VeX) = 5112 - (7/12i = 111144

A similar calculation will show that E(Y) = 7112 and V(Y) = 11/144.

E(XY) = JIJlxy(X + y)dxdy= fIJI(X2y + Xyl jdxdy

o Q Jo 0 - -

= flex}y/3 + x 2y2/2)11 dy = JICy/3 + y2/2)dy

o x=O 0

= (116)(/ + i)l~ = 1/3

Cov(X,Y) = E(XY) - E(X)E(Y) = 1/3 - (7f12)(7!12) = -11144

Cov(X, Y) P -----r'~~~ XY - JV(X)V(Y)

- 11144 = -1111 J(l1l144)2

Chapter 11

77

11-19. Let ftx.y) = (3/8)(x2 + i), for -1 :::; x .:: 1 and -I :>: Y s 1.

(a) f,(x)=(3/8)LrX2 + y")dy =(3/8)(x2y+ yJ/3l(=_1 = (3/8)(2x2 + 213) = 3x2/4 + 114, for -I .:: x':: 1

fl" ,

fy(y)=(3/8) JX· +y")dx= 114 + 3y-/4, for-l <;:y:>: I

fx(x)fy(Y) = (1/4 + 3x2/4)(114 + 3/(4) "" (Xl + /) = f(x,y)

Hence X and Yare not independent.

(b) E(X) = [xfxex)dX = (xl4 + 3x'/4)dx

I

= (x2/8 + 3x4/16)1 = 0 -I

Bya similar calculation, EeY) = o.

fIJI JIJI , 2

E(XY) = xyf(x,y)dxdy = (3/8) xy(x- + y )dxdy

-I -I -I -I

JI 4 ,- II

= (3/8) )x y/4 + x .y' 12) x~-I dy = 0

COy(X, Y) = E(XY) - E(X)E(Y) = 0

78

Chapter 11

11-20. The joint probability function yields the following table.

x

Y I 2 py(y)

I 2115 4115 2/5

2 3115 6115 3/5

Px(x) 113 2/3 X and Yare clearly independent

Mx(t) = E(e'x) = L e'Xp(x) = (l/3)e' + (2/3)e11 My(t) = E(e'Y) = L ely p(y) = (2/S)e' + (3/5)e2t

By independence: Mx+y(t) = Mx(t)My(t)

= (lflS)(2e21 + 7e31 + 6e41)

11-21. If X and yare each uniformly distributed over [0, 2], then

Mx(t) = My(O = (e21 - 1 )/2t. (Exercise 9-4)

Since X and Y are independent,

11-22. By Example 5.25 and Exercise 5-33, for each i, E(Xj) = 7/2 and V(Xj) = 35/12.

E(S) = E(X d + E(X1) + ... + E(Xn) = n(7/2)

Since the X;'s are independent,

Chapter 11

79

4

11-23. V(Xj + x, + x, + X4) = L V(X;) + 2L Cov(X;X j)

i=! I<J

= 4(13/162) +2(6)( -1/81) = 14181

11-24. IfS = X, + Xl + ... + XlO, V(S) = 500/9, VeX,) = 25/3 for each i, and all covariances Cov(Xj,Xj) for i l' j are the same, then

10

V(S)= LV(X,)+2LCov(X"XJ)'

i=1

(The number of terms in the covariance sum is I + 2 + ... + 9 = 45.)

500/9 = 10(25/3) + 2(45)Cov(Xj,Xj) Cov(X"Xj) = -25/81

11-25. S is the sum of 500 independent and identically distributed random variables each with mean .50 and variance .25. Then S is approximately normal with fl = 500(.5) and oj = 500(.25).

P(235 s; S s; 265) = P ( 235 - 250 s; Z s; 265 - 250 J\

l .Jill .Jill

= P(-L34 s; Z s; 1.34) = .8198

For Exercises 11-26 to 11-30, we make the following table of conditional probabilities from the table in Example 11.28, e.g.,

p(0190) = P(Y = 0IX = 90) = p(90,0)/Px(90) = .05/.20 = .25.

o 10

.25 .75

.45 .55

.90 .10

II

80

Chapter 11

11-26. (a) E(YIX = 90) = Op(OI90) + [Op(l 0190) = 0 + 7.5 = 7.5 (b) E(YIX= 100)= Op(OIIOO) + 10p(101100) = 5.5

(e) E(YIX= 110)=Op(01110)+ 10p(1011IO)= I

11-27. E[E(YIX)] = E(YIX = 90)px(90) + E(YIX = IOO)Px(lOO)

+ E(YIX = II O)Px( 1 00)

= 7.5(.20) + 5.5(.60) + 1(.20) = 5 = E(Y)

11-28. (a) E(yliX = 90) = 0(.25) + 100(.75) = 75 V(YIX = 90) = 75 - (7.5i = 18.75

(b) E(y2IX = 100) = 0(.45) + 10(.55) = 55 V(YIX = 100) = 55 - (5.5i = 24.75

(c) E(Y/IX= 110)=0(.90)+100(.10)= 10 V(YIX= 110) = 10-1 =9

11-29. E[V(YIX)] = V(YIX = 90)px(90) + V(YIX = 100)py{100) + V(YIX = 110)Px(110) = 18.75(.20) + 24.75(.60) + 9(.20) = 2004

11-30. V[E(YIX)] = l)E(YI X = x) - E(y»2 Px (x)

x

= (7.5 - 5)2(.2) + (5.5 - 5)\6) + (I - 5)\2) = 4.6

V[E(YIX)] + E[V(YIX)] = 4.6 + 2004 = 25 = V(Y)

Chapter 11

81

11-31. Let S be the random variable indicating whether a claim has been filed (S = I) or not (S = 0). Then Pt S = 0) = .93 and peS = I) = .07. Let X be the random variable of the claim amount paid.

rex = 500lS =1) = .60 P(X = 1000lS = 1) = .30 pex = 2000iS = I) = .10

E(XIS = 0) = 0 and V(XIS = 0) = 0

E(XIS = I) = 500(.60) + 1000(.30) + 2000(.10) = 800 E[E(XIS)] = E(X) = 0(.93) + 800(.07) = 56

E(X21S = 1) = 500\60) + 10002(.30) + 2000\ 10) = 850,000 VeXIS = I) = 850,000 - 8001 = 210.000

E[V(XIS)] = 0(.93) + 210,000(.07) = 14,700

V[E(XIS)] = (E(XIS = 0) - E(X»2p(S = 0)

+(E(XIS = 1) - E(Xl/P(S = I) = (0 - 56)"(.93) + (800 - 56l(.07) = 41,664

V(X) = V[E(XIS)] + E[V(XIS)] = 41,664 + 14,700 = 56,364

11-32. for f(x,y) = 6x, 0 < x < y < I and 0 elsewhere:

(a) fx (x) = {6xdy = 6x( 1- xjdx.for 0 < x < 1 (b) E(X) = D6x2 -6xJ)dx=lf2

(e) E(X2) = D6x1 - 6x 4 )dx = 3/1 0

VeX) = 3110 -(lI2i = li20

(Note: All these were calculated in Exercise II-IS.)

82

Chapter 11

11-33. In Exercise 10-22 we showed that C{y) = 3i ,0 < y < 1.

11-34.

fi(x I ~.) = f(x, y) = ~ = 2x 0 < < y

s f'() 1 2' X

y Y 3y y

E(XIY=y)= fYxf(XIY)dx=f'2x,2 dx= 2x~ I.' =2yf3

o 0 y- 3y~ ,~O

E[E(XIY)] = fIE(X I Y = y)fy(v)dv = fl 2y (3'/ ldy

o • - 0 3 - 0

, 11

= (2y 14) 0 = 1/2 (This equals E(X).)

fl 1 \/4

11-35. E[V(XIY)]= oV(XIY=y)fy(y)dy= So 6dy

51

= LI =1/30 30 0

11-36.

E[EeXly)2] = fl (2YJ2 (3y2 ldy = 4y.511 = 4/J 5

o 3 - 1) 0

V[E(XIY)] = 4/15 - (112)2 = 1160

E[V(XIY)] + V[E(XIY)] = 1130 + 1160 = 1/20 = VeX)

Chapter 11

83

11-37. The number of claims N is a Poisson random variable, 50 E(N) = YeN) = A = 20. The amounts of claim random variable X is uniform over [0, 500], so E(X) = 250 and VeX) = 5002/12.

(a) E(S) = E(N)E(X) = 20(250) = 5000

(b) YeS) = E(N)V(X) + V(N)E(X)'

= 20(5002112 + 250") = 1.666,666.67

11-38. If the amount of claim random variable X is lognormal whose underlying normal distribution has 11 = 5 and cr = .40:

E(X) = el'+"t = e50~

VeX) = e1l'Ml (c" -J) = elO 16 (e16 -1)

(a) E(S) = 20eS08 = 3215.48

(b) V(S)=20[elOI6(e16-1)+ e10I6)= 606.665.15

11-39. The claim amount S is approximately normal, with

11 = E(N)E(X) = 500(250) = 125,000 and cr2 = 500(5002112 + 2502), so cr = 6454.97.

pes.e; 140,000) = p(z 5, 140,000-125,000) 6454 .. 97

= P(Z 5, 2.32) = .9898

84

Chapter 11

11-40. If E(X) = 600 and VeX) = 12,000, then S is approximately normal with ).l = 500(600) = 300,000 and

02 = 500(12,000 + 6002), SO c = 13,638.18.

We need to find So so that peS .,:: So) = .95. From the z-tables we know that Fz(1.645) '" .95. Hence

P(Z <; 1.645) = P (S - 300,000.,:: 1.645J '" .95 . [3,638.18

Therefore So = 300,000 + 1.645( [3,638.18) = 322,434.81.

,

t :

ii

i!

I.'

CHAPTER 12

The data in the table below is for the simulations in Exercises 12-1 and 12-2. The third column shows the number of coins player A has in [2-1 and the fourth shows which fund the employee in 12-2 is in at the end of each month.

**Trial Random A has Fund
**

Number

Start 3 0

1 .57230 4 0

2 .85472 5 [

3 .37282 4 I 1

4 .77133 5 [

5 .20525 4 0

6 .82496 5 1

7 .52184 6 I

8 .49837 5 1

9 .76729 6 [

10 .50986 7 I

11 .02480 6 0

12 .99954 7 1

13 .81708 8 1

14 .90535 I

15 .76227 I

16 .78322 I

17 .00067 0

18 .24844 0

19 .14118 0

20 .47417 0 I.

86

Chapter 12

12-1.

Player A wins a toss if the random number is in [.5,1) and loses otherwise. The game ends when A has all 8 coins (he wins the game) or when he has 0 coins (he loses the game). From the table, A wins in 13 tosses.

12-2.

If the employee is in Fund 0, he remains there if the random number is in [0,.65) and switches otherwise. If he is in Fund I , he switches if the random number is in [0,.25) and stays there otherwise. From the table he is in Fund I thirteen times.

12-3.

The wartmg time X between accidents is exponentially distributed with ~. = 3, so Ftx) = I - e-3,. To simulate waiting times, x = F-l(u) = ~ln(1 - u)/3. The following table simulates the waiting time between accidents and the cumulative time to the nth accident. From the table you count the number of accidents in the first few months.

Random F (u)

Trial Number WaitiQg_ Time Total Time

1 0.57230 0.28311 0.28311

2 0.85472 0.64303 0.92614

3 0.37282 0.15551 1.08165

4 0.77133 0.49183 1.57347

5 0.20525 0.07658 1.65005

6 0.82496 0.58091 2.23096

7 0.52184 0.24594 2.47690

8 0.49837 0.22996 2.70686

9 0.76729 0.48599 3.19285

10 0.50986 0.23769 3.43054

II 0.02480 0.00837 3.43891

12 0.99954 2.56143 6.00034

[3 0.81708 0.56624 6.56657

14 0.90535 0.78586 7.35243

15 0.76227 0.47887 7.83130

[6 0.78322 0.50962 8.34093

17 0.00067 0.00022 8.34115

18 0.24844 0.09520 8.43635

19 0.14118 0.05073 8.48708

20 0.47417 0.2[426 8.70134 Chapter 12

12-4.

(a) p(\}=p(O)p= [so .. 50][·65 J5L{.45, .55] .25 .75 J

(b) p(2) = p(l)p = [.45, .55] [.65 .35L [,43. .57] .25 .75J

12-5.

(a) p(ll=pCOlp= (40, .60]l·72 .28J=[.504 . .496] .36 .64

2· II [ ] r.n .28J L ~ 5856J

(b) p(l=pl p= .504, .496l1 = .)4144, .4

. .:>6 .64

12-6.

p(ll = pcoip = {,30, .30,

[4

.40] .2 .1

'.~ :~l = [.22, .33, .45]

.3 .6

12-7. F or this Markov process the transition matrix is

[.60 .20 .201

P = .25 .50 .25 and p(OI = [1,

.30 JO .40

0, 0]

p'" = p(Olp = [.60, .20, .20]

[.60 .20

p(21 = p(llp = [.60, .20. .20] .25 .50 .30 .30

.201

.25 =[.47.

.40

.28. .25]

87

88

Chapter 12

12-8. To find the limiting distribution for the transition matrix in Exercise 12-4, we have to solve the following system of equations.

[x, y][.65 .35l~[x, yJ .25 .75J

x+y=1

;'\

; .

or

.65x + .25} = x .35x + .75y = Y x+ y=l

; ~

'-oJ .

which can be rewritten as

i

:\' I

-.35x + .25y = 0 .35x - .25y ~ 0 x + y= I

The solution is [5/12, 7/12].

12-9.

The limiting distribution for Exercise [2-5 is the solution of the following system of equations.

[ ] l·72 .28] = [x yJ

x, Y .36 .64 '

x + Y = 1

This can be rewritten as

-.28x + .36y = 0 .28x - .36y = 0 x + Y = I

The solution is [9116, 7/16J.

Chapter 12

89

12-10. The limiting distribution for Exercise 12-6 is the solution of the following system of equations.

[x, y,

[.4 .2 .41

z] .2 .5 .3J = [x, y, zl .1 .3 .6

x + y + z = 1

This can be rewritten as

-.6x+.2y+.lz=0 .2x - .51' + .3z = 0 Ax +.3y - Az = 0

x+y+z=1

The solution is [11/57, 20157, 26/57].

12-11. The limiting distribution for Exercise 12 -7 is the solution of the following system of equations.

[.60 .20 .20]

[x, y, -l .25 .50 .25 ~ [x,

.30 .30 040

x + y + Z = 1

y, z]

This can be rewritten as

-Ax + .25)' + .3z = 0 .2x - .5)' + .3z = 0 .2x + .25)' - .6z = 0

x+ y+ z=l

The solution is [ISI37, 12137, 10/37].

90

Chapter 12

12-12. Let P be the transition matrix of a regular Markov process and let £ be its limiting distribution.

Taking the limit of both sides,

lim pn = (£ limpn-I)p

n....:;.-x; G ..... -':O

n= (fL)P f = [P'

(Recall that for any vector v, vL = t.)

12-13. The transition matrix for the states for player A, after rewriting the matrix with the absorbing states first, is

rIo

o I p = 2/3 0

o 1/3

o o o 2/3

(a)

1- =[ I -113]

Q - 213 I

_I [917 317]

(I - Q) = 617 917 = N

(b)

A = NR = [917 3/71[2/3 0] ~[617 II7J

617 9/7 J 0 1/3 417 317

(c) The probability that if player A starts in state 2 (2 coins) he ends in state 0 (he loses) is a20 = 417. Recall that the rows in A refer to states 1 and 2 and the columns to states 0 and 3.

Chapter 12

91

**12-14. The transition matrix for the states for player A, after rewriting
**

the matrix with the absorbing state s first, is

r ~ r 2~3 0 0 0 0

[ 0 0 0

0 0 113 0

I 0 0 2/3 0

l 0 [/3 0 2/3 0

- [/3 -H

(a) I-Q= 2/3 I

0 - 2/3

[715 3/S 1151

(I-Q)-I = 6/5 91S 3/5 =N

4/5 6/5 7/5 (b)

[7/S 3/S l/SI2/3 A = NR = 6/5 9/S 3/5 0

4/5 6/5 7/5 0

r14l15 1115]

= i 4/S 115

lS/15 7/15

( c)

The probability that if player A starts in state 2 (2 coins) he will end in state 0 (he loses) is a20 = 4/5.