To summarize, the document discusses the Borel-Cantelli lemma and its two parts. The first part states that if the sum of the probabilities of an infinite set of independent events converges, then the probability that infinitely many of those events occur is 0. The second part, which assumes the events are independent, states that if the sum of the probabilities diverges, then with probability 1 infinitely many of the events will occur. An example is provided using a sequence of Bernoulli trials to illustrate the second part of the lemma.
To summarize, the document discusses the Borel-Cantelli lemma and its two parts. The first part states that if the sum of the probabilities of an infinite set of independent events converges, then the probability that infinitely many of those events occur is 0. The second part, which assumes the events are independent, states that if the sum of the probabilities diverges, then with probability 1 infinitely many of the events will occur. An example is provided using a sequence of Bernoulli trials to illustrate the second part of the lemma.
To summarize, the document discusses the Borel-Cantelli lemma and its two parts. The first part states that if the sum of the probabilities of an infinite set of independent events converges, then the probability that infinitely many of those events occur is 0. The second part, which assumes the events are independent, states that if the sum of the probabilities diverges, then with probability 1 infinitely many of the events will occur. An example is provided using a sequence of Bernoulli trials to illustrate the second part of the lemma.
foreverym =0, 1; 2, ....Therefore,bytheindependenceoftheevents AI, A
2 , . , weget n+m ) P(Bn):5 P Dn Ak ( n+m ) =(l - Pn) ...(1 - Pn+m) :5 exp - L Pk (2-77) ( k=n where we have made use ofthe inequality 1- x :5e- X , x O. Notice that ifAI, A 2 , is a sequence ofindependentevents, then so is the sequence ofcomplementary events AI, A 2 , ... Butfrom (2-70) n+m LPk --+ 00 as m --+ 00 (2-78) k=n Therefore,passingtothelimitm --+ 00in(2-77),wefindthatP(B n ) = 0foreveryn = 1,2,.... Thususing(2-76) PCB) :5 L P(Bn ) =0 n andhence PCB) =1 - p(lf) =1 (2-79) thatis,the probabilityofinfinitelymanyoftheevents AI, A 2 , .occurringis 1. Noticethatthe secondpartoftheBorel-Cantellilemma,whichrepresentsaconversetothefirstpart,requiresthe additionalassumptionthattheeventsinvolvedbeindependentofeachother. ..... Asanexample,considertheevent"HH ...H" occurring ina sequenceofBernoulli trials. To determinetheprobabilitythatsuchan "all success" sequenceoflengthn ap- pears infinitely let A k stand for the event "Head appears on the kth toss," imd y define B i = Ai n A i + 1 n A i + n - 1 , i 1. Wehave P(B i ) = Pi = pn. Theevents B i arenotindependent,however, theevents B1, B n +1, B 2n +1, ...areindependent,andthe series2::f:,oPkn+l diverges.Hence,fromthesecondpartoftheBorel-Cantellilemma,it followsthatwithprobabilityonethepattern"H H ...H" (aswellasanyotherarbitrary pattern)willoccurinfinitelyoften. To summarize, ifthe sumoftheprobabilitiesofan infinite set ofindependent events diverge, then W\th probability 1, infinitely many of thoseeventswilloccurinthelongrun. PROBLEMS 2-1 Showthat(a) AUB UAU B =A; (b) (A UB)(AB) =AB UBA. 2-2 IfA ={2 :5 x :5 5} andB ={3 :5 x :5 6}, find A UB, AB, and (A UB)(AB). 2-3 ShowthatifAB= {0}, then peA) :5 PCB). peA) =PCB) =P(AB), thenP(ABUBA) =0;(b) ifpeA) =PCB) =i, then P(AB) =1. 2-5 Proveandgeneralizethefollowingidentity peA U B U C) =peA) + PCB) + P(C) - P(AB) - P(AC) - P(BC) + P(ABC) 2-6 ShowthatifSconsistsofacountablenumberofelements andeachsubset isanevent, thenallsubsetsofSareevents. 2-7 if'S={1, 2,3,4}, find thesmallestfield thatcontainsthesets{1} and{2,3}. -I 77) ls a 78) '9) he he 11i )- d I 'i e t { CHAPTER2 TIlEAXIOMSOFPROBABILITY 45 2-8 If A C B, P(A) =1/4, andP(B) =1/3, find P(A 1B) and P(B I A). 2-9 Showthat P(AB 1C) =P(A I BC)P(B 1C) andP(ABC) =P(A I BC)P(B I C)P(C). 2-10 (Chain rule) Showthat 2-11 Weselectatrandomm objectsfromasetSofn objectsandwedenotebyAm thesetofthe selectedobjects.Showthattheprobabilityp thataparticularelement of S isinAm equals min. Hint: p equalstheprobabilitythatarandomlyselectedelementofSisin Am. 2-12 Acalloccursattimet, wheret isarandompoinrintheinterval(0,10).(a) FindP{6 :5 t :5 8}. (b) FindP{6 :5 t :5 81 t > 5}. 013l ThespaceSisthesetofallpositivenumberst. ShowthatifP {to :5 t :5 to +tIl t ::: to} = '-=7 P {t :5 tJl foreveryto andt}, then P {t :5 tJl =1 - e-C'l, wherecisaconstant. 2-14 TheeventsA andB aremutuallyexclusive. Cantheybeindependent? 2-15 Show thatifthe events AI, ... ,An areindependentand Bj equals Ai or;;e orS, then the eventsB l , .., B n arealsoindependent. 2-16 A box contains n identical balls numbered I through n. Suppose k balls are drawn in succession.(a) Whatistheprobabilitythatm isthelargestnumberdrawn?(b) Whatisthe probabilitythatthelargestnumberdrawnislessthanorequaltom? 217 Supposek identical boxes containn balls numbered I through n. Oneballis drawn from eachbox.Whatistheprobabilitythatm is thelargestnumberdrawn? 2-18 Tenpassengersgetintoatrainthathasthreecars.Assumingarandomplacementofpassen- gers,whatistheprobabilitythatthefirst carwillcontainthreeofthem? 219 A boxcontainsm whiteandn blackballs. Supposek ballsaredrawn.Findtheprobability ofdrawingatleastonewhiteball. 2-20 A playertosses a penny from a distance onto the surface ofa square table ruled in I in. squares.If thepennyis 3/4 in. indiameter, whatis theprobability thatitwillfall entirely insideasquare(assumingthatthepennylandsonthetable). 2-21 In the New York State lottery, six numbers are drawn from the sequenceofnumbers 1 through 51.Whatistheprobabilitythatthesixnumbersdrawnwillhave(a) allonedigitmpnbers? (b) twoone-digitandfourtwo-digitnumbers? 2-22 Showthat2 n - (n +1) equationsareneededtoestablishtheindependenceofn events. 2-23 Box1contains 1whiteand999redballs. Box2 contains 1redand999whiteballs.A ball ispickedfromarandomlyselectedbox.Iftheballisredwhatistheprobabilitythatitcame fromboxI? 2-24 BoxI contains 1000bulbsofwhich10%aredefective.Box2contains2000bulbsofwhich 5%aredefective.Twobulbsarepickedfromarandomlyselectedbox.(a) Findtheprobability thatbothbulbsaredefective. (b) Assumingthatbotharedefective,find tileprobabilitythat theycamefrombox 1. 2-25 A trainandabusarriveatthestationatrandombetween9A.M. and10A.M. Thetrainstops for 10minutesandthebusforx minutes.Findx sothattheprobabilitythatthebusandthe trainwillmeetequals0.5. 2-26 ShowthatasetS withn elementshas n(n-1)(n-k+1) n! 12k k!(n-k)! /(f)"k-elementsubsets. V , . We have two coins; the first is fair and the second two-headed. We pickoneofthe coins " atrandom, we tossittwiceandheadsshowsbothtimes. Findtheprobability thatthecoin pickedisfair. CHAPTER3 REPEATEDTRIALS 11 PROBLEMS 3-1 Let p representtheprobability ofan event A. Whatis the probability that(a) A occurs at l) leasttwiceinn independenttrials; (b) A occursatleastthriceinn independenttrials? 3-2 A pairofdiceis rolled50times.Findthepro!;Jability ofobtainingdouble sixatleastthree Jr times. 3-3 A pairoffair dice is rolled 10 times. Findtheprobability that "seven" will show at least once. J.e ,er 3-4 Acoinwithp{h} = p = 1 - q istossedn times.Showthattheprobabilitythatthenumber ofheadsisevenequals0.5[1 +(q _ p)n]. 3-5 (Hypergeometric series) A shipmentcontains K goodand N - K 'defectivecomponents. Wepickatrandomn ::: K componentsandtestthem. Showthattheprobability p thatkof thetestedcomponentsaregoodequals(comparewith(3-39)) re, Ike ne. p= fS.) r3-6\ronsiderthefollowing threeevents: (a)Atleast1sixisobtainedwhensixdiceare rolled, ber C/(b)atleast2sixesareobtained)Vhen 12dicearerolled,and(c) least3sixesareobtained when18 dicearerolled.Whichoftheseeventsismorelikely? ign 3-7 A player wins $1 ifhe throws two heads in succession, otherwise heloses two quarters. :ted Ifthe gameis repeated50times, whatis the probability that the net gainorless exceeds and (a)$1?(b)$5? 3-8 Supposetherearer successesinn independentBernoullitrials.Findtheconditionalproba- bilityofasuccessontheithtrial.
ents 3-9 A standard packofcards has 52 cards, 13 ineachof4 suits. Suppose4players are dealt lave 13 cards eachfrom awell shuffledpack. Whatis theprobabilityofdealing aperfecthand eof (13 ofanyonesuit)? 17 is toExample3-15(Gambler'sruinproblem).LetN a denotetheaveragedUflltionofthe 1the V gameforplayerA startingwithcapitala. Showthat :cted one b a+b p=j:.q ;cess N = 2p- 1 - 2p- 11 _ a lould { 1 ab p =q ="2 (Hint: Show that N k satisfies the iteration N k = 1+pNk+1 +qNk-l under the initial conditions No = N a + b = 0.) 1 3-11 RefertoExample3-15.SupposethestakesofA and B areIX and{3, andrespectivecapitals areaandb,asbefore.Findtheprobabilitiesfor A orB toberuined. 3-12 Threedicearerolledandtheplayermaybetonanyoneofthefacevalues 1,2,3,4,5,and6. If theplayer'snumberappearsonone,two,orallthreedice,theplayerreceivesrespectively one,two,orthreetimeshisoriginalstakeplushisownmoneyback.Determinetheexpected lossperunitstakefortheplayer. 120 PROBABILITYANDRANDOMVARIABLES ;0erandomvariableis xis N(1), cr) andP{1) -kcr <x<1)+kcr}= Pk. (a) FindPk fork::::: Q2, and3. (b) Findk forPk =0.9,0.99,and0.999.(c) IfP{1) -zucr < x < 1)+zucr} =y, expressZu in termsofy. 4-5 Find Xu for u =0.1,0.2,...,0.9 (a) ifx is uniform in the interval (0, 1); (b) ifJ(x) ::::: 2e- 2x U(x). G WemeasureforresistanceRof eachresistorinaproductionlineandweacceptonlytheunits theresistance ofwhich is between96 and 104ohms. Findthe percentageofthe accepted units (a) ifR is uniformbetween95 and 105 ohms; (b) ifR is normal with1) =100and cr =2ohms. 4-7 Show that if the random variable x has an Erlang density with n = 2, then Fx(x) ::::: (1 - e- cx - cxe-CX)U(x). 4-8 Therandomvariablexis N(10; 1).FindJ(x I (x- 10)2 < 4). 4-9 Find J(x) ifF(x) = (1 - e-aX)U(x - c). 4-10 Ifx is N(O, 2) find (a) P{I x 2} and(b) P{I x 21 x::: 1}. .;.:.::.;.... .-' f'4-ii"'ThespaceSconsistsofall points tj intheinterval (0, 1) and P{O tj y} =y forevery '--../y 1. The function G(x) is increasing from G(-oo) =0 to G(oo) = 1; hence it has an inverse G(-l)(y) =H(y). Therandomvariablexis suchthatx(tj) =H(tj). Showthat Fx(x) =G(x). 4-12 If x is N(10oo; 20) find (a) P{x < 1024}, (b) P{x < 10241x> 96I}, and (c) P{31 < v'x 32}. 4-13 Afaircoinis tossedthreetimesandtherandomvariablexequalsthetotalnumberofheads. FindandsketchFx(x) and Jx(x). 4-14 Afaircoinis tossed900times and therandomvariablexequalsthetotal numberofheads. (a) FindJx(x): 1; exactly2;approximatelyusing(4-34). (b) FindP{435 x 460}. 4-15 Show that, ifa x(n b for every SE S, then F(x) =1for X > bandF(x) =0for x < a. 4-16 Showthatifx(n yenforeveryS E S, then Fx(w) ::: Fy(w) forevery w. 4-17 Showthatiff3(t) =J(t Ix > t) is theconditionalfailurerateoftherandomvariablexand f3(t) =ft, then J(x) isaRayleighdensity(seealsoSec. 6-6). 4-18 ShowthatpeA) =peA Ix x)F(x) + peA Ix>x)[l- F(x)]. 4-19 Showthat peA Ix x)Fx(x) Fx(x I A) = peA) 4-20 ShowthatifP(AIx =x) =PCB Ix =x) foreveryx Ix xo) =PCB xo). Hint: Replacein(4-80) peA) and J(x) by peA Ix xo) and J(x Ix xo). 4-21 Theprobabilityofheads ofarandomcoinisarandomvariablep uniformintheinterval(0, 1). (a) Find P{0.3 P 0.7}. (b) Thecoinis tossed 10 times and heads shows 6 times. Findtheaposterioriprobabilitythatp is between0.3 and0.7. 4-22 The probability ofheads ofarandom coin is arandom variable p uniform in the interval (004, 0.6). (a) Find the probability that at the next tossing ofthe coin heads will show. (b) Thecoin is tossed 100timesandheads shows60times. Findtheprobabilitythatatthe nexttossingheads willshow. 4-23 Afaircoinistossed900times.Findtheprobabilitythatthenumberofheadsisbetween420 and465. Answer: 0(2)+ 0(1)- 1:::: 0.819. 4-24 A fair coinis tossed n times. Findn such thatthe probability thatthenumberofheadsis betweenOA9n and0.52n is atleast0.9. Answer: 0(0.04"fii) +G(0.02"fii) ::: 1.9; hencen > 4556. -._-- -----