SOLUTIONS TO ~ PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Leonard A. Asimow, Ph.D., ASA Mark M. Maxwell, Ph.D., ASA ACTEX PUBLICATIONS, INC. WINSTED, CONNECTICUT SOLUTIONS TO PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Leonard A. Asimow, Ph.D., ASA Mark M. Maxwell, Ph.D., ASA ACTEX PUBLICATIONS, INC. WINSTED, CONNECTICUT Copyright © 2010, by ACTEX Publications, Inc. All rights reserved. No portion of this book May be reproduced in any form or by and means Without the prior written permission of the Copyright owner. Requests for permission should be addressed to ACTEX Publications PO Box 974 Winsted, CT 06098 Manufactured in the United States of America 109539) 65437221 ISBN: 978-1-56698-722-6 TABLE OF CONTENTS Chapter 1: Combinatorial Probability 1 amet ie EL OD aD ymVIOGCle neha cck ystems Ser ook cceccs ac scrretnsch mites estes tcsscase 1 1.2 Finite Discrete Models with Equally Likely Outcomes ..........c.sccscccssssssssessssessesesessssesesees 2 eM ATU ara Ec UO Ut ON card stn sMeree cme era. oes. doses Sage ac acsTPe Letetseoes coset cus es eee 8 TES IN RGTOS TANbY)FCGE218(6)1S 1s a SR OH yah rR oe 10 meee DaDich IPS alliplcieXAIMINAtlOU, ten wet cemtese 5, tenth tctas ts ete roa sitec es eass aeece ee ee 13 Chapter 2: General Rules of Probability 15 PM CUBE THCOLV ASUL VIVAL Koll srtce sogeen toner ei ta coc ao Ateabevioscanavh cues (Eee vues: we Resa cea eee 15 Bem COUCH LON AL ELODADIILY 6-5 es ie ctseseen irra oo is Cie chetce ogi cree AC Eon 23 MR OCC CEC Ctra tn cet Nese ocuscaode Guat Cap yc taet a SNe cea oa CMR aN oe 29 SMES ee 35 Ay CS eelNCOTSIYI Reet eia 0 ees, Ws each Ty Re nD ee AS NCO ako e a SN Re Ue eed sea ee ee eee 38 in occy0. cen Lia CoECOMDUINCY tee Pewee Ree PME Pee IU cles aliple EX aNMDatlOl sc sc1 tae tg a ers AR Aca, rye hioacts anes eserth ioe ene: 40 Chapter 3: Discrete Random Variables 51 EMIS LeosANCOMI.Y ALA DICS omen eet es aa aiced Sithcee artiste ors ete ats 51 Pee ULALIVe EroUaDIlLy LiSitiDULIONe er a etc eee a ee ca eres ne a2 Boe Mc cotirC ayn (ci ralel CLICCIICY aerrets eee nicer: ars ahans ys OC eae mer ciaeei od tos Eee a Re HC ASUITOSM aLISS CE SILOM erat eotner te tommeted cseeraten ty tnReocia aaaraat eek ante een cre ee 56 Pen ONKOL ex DeCLAUON ANG. V ALANCO. sara sees eoceaceed.coasece cassdndsaphpeueetepbes nian eset ao SomemrOiny listed Random. Variables (ROUMnG 1) i ..jcc-+..tss avn cavesoecasncssouevasttocweansndeones 62 Soe EXO UA IMI yA Oe LAtIN ee:PUNCLION: sxc.nsatectscsseet sides secs(oasis siieutcnseiariensnvacnnsuaancas@nesnsomes 65 Fea 11) teeny MUNA KATIE Oar cee ocean stent ovicaog ict oie wn sieancano dnees dun ee vacua soutemen neue haeak oe 66 Chapter 4: Some Discrete Distribution 73 SP Uist emTOUR LUE UNIOU wetge sscles rncger eeniveneeie ate ar sha ce dshy vara Snnca>yeruacieseee 73 aoe Deraoill triale and the DYimormial DISPUTOD cose tosscxencsetievanceraesesseusumagesutens-cosacanas--nadees 76 SPM rere OIAC ITTGR!ISELIN eee ecg regi can 5 Coxe ascent e datspsoesdupiaa seep asuacgdth da aalia soto ene danon 78 em ee ASV ESOT JISCUIU LOU oponcsscacarnc gapeaanissmny smcpig-te ths wwaantaasypGrtaicduoasneduda scien sve’ 82 Gee CAELY et eCICOMICITIC. IST DULIOM so,ucct:csocoecrs sndeuse sui oorensaftrsantpach coatnens saataacaruruspomcanan' 83 Sm LICH ASSET SSL PUL LULU ety orecar rie aetraks avacoedvecout creck y ck cata tanta ri dhaain sadlsea is nstuas haa hinn 84 vinesinestnsuserensny areananennrey nan are89 GIALS CoOUDArisoll OL LISCLETS LISI DULONS divyorenccenss. BES c Napict 4 matinic EX AIMAtION eerste css enshcdtns.svoegy re costal atactraecanaban< argon: 93 imme Chapter 5: Calculus, Probability and Continuous Distributions 97 LLOM ae tiasciiaa rant cater cb es? IV Assaraly as cof wats dunes soon nt vusch,casiornas iN adbssacnunanaesa ise. oF me ML eS IEE UMC MLC erPCC ON Goat ot crepe cimn ate cicwat se by sillnish oo dieovsands inges nyunsoaatenges capsinnia svsis>oxslenes quantum cannes: ao SOCEM COLL VESEY ASUONLstpee Cee te usta RY ol cae sav capaha unsaaautin ics vbrannntoactyts genk oc sSaencns sens104 Applications to Insurance: Deductibles and Caps.....cccicesecseseetnessssessnessessseeseesenens 105 5.5 fetapencsanacsien euapnqects og nen 107 esisatsreats enncaacegiqayunts Oe tCC ETAL PUG CULO Mes ecaeesagtevac UV sesenntsasnsoneens 111 sssentosestesseeaeesseneoeno scisscisesccscsstsesacossces Chapter S Sample Examination... 5-7 il ® TABLE OF CONTENTS Chapter 6: Some Continuous Distributions 115 C7 ee UDLLORTR ANG OMIeV ALLA IES acccas seater, ete nee eae ees Py eee tae pee giant tec 1 Ore) Hee XPOUCA LDIS(T DULION tc... toes xt nee aes em aioe oe yc ah lk och eneaete abe 116 6:3 [he Normal Distribution... at-0 ccm hese epee wood ocd ceo encstou ocpinsonnjaaenetee= ce 120 64 Pe ihe law or Averages and the Central Lrmit VR6 Gren ies. scstecesss-cee-ccdocesesoacvaraneudeevatene122 Orme Ne GraTniina-L)ISCITOUILL OIL cp tere crete tne ee eerede- fag ete gine te dnc ven ats atee saa nie altuna eevee | Wee G Geel Ne-Beta tainty Of L1Sttt DUNOMS se eecna cess oavevees teeter er tor et ws eaunee aa ore reese wader 127 Ow Ole © ONUTUOUS LIIStPIOULIOIS sree eastern tate neti gtentees nt tesa ea asian tacoe cies 130 GrSaar CMaptcy O-Salaple EX ALMITALIOL sa corse cr sncee ieee raed aster eer eee meine adieree tgs ecteoacernee ened 132 Chapter 7: Multivariate Distributions 135 7A, J Olt Disiibutionstor Discrete RaiCOM VAMADIES jcessesee pevercses-cercccet eecnecn-auneekeeteneeene 135 7.2. Conditional Distributions — The Discreté Case... iesc-c-secenscnen--ee ian aaitaces dome ee 137 Jedme. Independence. DISCTELe, CASE ares teccasiynnt neta. ee URC ee aati alse e ee epee a 139 eee COVarIance atid COTTCLAUOU:.«.a5 thet eoeaccts sascha cee erence Se nace tat clas eerie a ease 139 7 yee Joint DIstioutions for Continuous Random Wallabies .rec..sstece-e ee asset eee 143 HeOme SOUCIHONAL WIStUIDUM ONS =. He G OMINOUS ese peer ens see css nase aaa 147 Ueyeeincependence and, Covariance m thie: CONUMUCUS, CASE. 5, ..cc.cncesersseten coestnaesesendesnca tanec 149 Pp mee Gy armale NOrmal: LISI DU TOL scum see coe tccve: erreueeaeroe tor ae ee deseasodedia: eked tenn ne eenenenmetetee 154 fal Cevioment Getlerating Function fOr a J OUIt DISH DULLON ys. -c oop aceeeveonanryoe ee-cneacento 155 POMS AO Le igieS ATL EX AIIIITALION sernorce a hleOocorenetacus tien temsuna ant ase take eased aman a ee 155 Chapter 8: A Probability Potpourri 163 Sree be istlibution Obdslranstorimed. andor: V alialee gece. ates ee orem eden een 163 Onl meetHe WIGMENE Gener atlily PAIR CU Oi ser gecsuce au. oc steeee te Meron cou eee ete ees oe nee 178 S cig, (COV anIaNCe! b OLAlaser sam, meee paces ec sctreenscrcee ce an easuanuecs teeneaeeee ee ee DP sa ohn 179 SAge he Coma iionin eek Oni las eerste. soe beeen cae nec ae orare oe eee ee 179 Se CHAplEr GS Saliple BXaMie AOD. asc cceececs ce tete crvaeiecesase rene roeee races mace ein eae 183 Chapter 9: Statistical Distributions and Estimation 187 Dales bhessample: Mean asran PStimiatot rye ccc ctor te aecree ee eee eee De eeu Demerol Mating Lee OPUlAatiOn. V alLALICOs, ic StNGent 1 Listhi DULL cxccac ecco es er Glatt eee a eee rercsri tarts ca ieennes © oanrsts Uneaten teem eo certo tao 187 ae 188 ee eens bone AS 190 De Mme eetLEEDS LAD ULL Lyn teste’ cee tvs sua cel cares astieaeoutenen aueeie rene ter osseuen vceanvee ea tee eames aeneremete eRe 191 SP am SUITIATILLS PTODOLLIOLS 5 .acace-cesoncacovsaty oxseveaeu roan Uatetens teeta aataunreacapsr ae sar tee eae eee tees 193 DEO MeL stiiiaung thea iierence, DOtWweenl IVCALS seed sectre teeet ens enisstecsteeeat waenateemetee ea treme 194 Fee SLULENAGIIVG CLS eSALII()LS cOUZ eattgmn nrg ei eyes uate seeks oneniert ceemesits cactiacercere caeta tact eter ee anne 195 DS COADLEr DO Sar pLe Haima COM. cx csncngsnseit eines teas tsaneenceees auevenie aan santesesgesccsnte nena enema 196 Chapter 10: Hypothesis Testing 201 LOST Hypothesis: Vesting: Prarie work: nasa shen canchs sce seuraaite nine arth cache ttee tek eat ee en 201 10.2 Hypothesis ‘Testing for POpUlation VICeIS cencconreuseanreenastcmreeectcs¢eteeentnesee eee nena 201 10.3 Hypothesis T estings Tor: Pop UlariOne Vartan a acccecetenigue terrence unten otters eee ene 10.4 Hypothesis Testing Tor PropGmOnSwan. anexs-c<ccvoteereanatueeteeetereet ctaeedsere neta cereeam 10.5 Hypothesis festingetor Ditterences 1 Population ivican acececawe sce eereeeeeene 204 10.6 CHI-SQuUare T EStS sicnsdaresstneneescsasecneea soulcaveisin tte eee MER een nee tee etn 10.7 Chapter 10 Sample EX art Ati00)ceiver sccasscauurs chester earn eaeanne nner chee eee 206 TABLE OF CONTENTS ® iil Chapter 11: Theory of Estimation and Hypothesis Testing 209 SWC HESTSpO i Vie SttAtOL. .ostac tuts csr rg. mamaate MONE. <teee «slo ensecoevasccsaveneeavecveetoatascastlec 209 SU, TSUN PORE LEGS WUNRETy A ea eR Rr ae ge I al ol SE er ee ee ae 209 IS ee LOPGnL CStOltrSUNTIALOES tre oer Cores. het RUM ere sae cactsgarvactuaah abt uretaes malar mens teaten 21S Aaa ELV OLNGSISRPSUIT pk COL Yeates ae tect eee e eee Mack oo dit ca ee tae dea whet oresdaw si aly, } Deum ViOre CrcTievallla ke lINOOG: RATION LEStG, epee stat ts peates farteoetec tec cactoxsoder seause dasa eeder atone: Pee, LVDSGS LSS SUIE GH SIRS VW WETET 0 2c A eo pF Ne Oe a a Pal SR RP er Rn Zo) eit icecineals WepressiOn:, asi F OTIMUIAS 2a ssucdieeseteacosnesdvosdnec-seacue-censbanettusousesvmrtes226 Bie cm ile Linea coresstOn:, HStimation, Of PAaraincteys cnctnctodetasnsracsocstecotdcsernsoceseeanrae 231 IMCS Gramm ANIC ox AITO ate eae rate tac ene surederectectate couse eansoesaruaube caciesmeeereease aye the Internet Archive a eae Seat ion stin Foundat cme ne HaBlea aes ns ne 7? https://archive.org/details/solutionstoproba0000na SOLUTIONS TO PROBABILITY AND STATISTICS WITH APPLICATIONS A PROBLEM SOLVING TEXT Ee See a eS SS A hat SAE CHAPTER 1: COMBINATORIAL Section 1.1 Ee OORT OES SSN LSE IS OE LOI See AK PENTOSE AY ee PROBABILITY The Probability Model oe (a) We systematically list the sample space. chicken sandwich chicken sandwich |coleslaw (b) {hamburger & potato chips & soda, hamburger & coleslaw & soda, chicken sandwich & potato chips & soda, and chicken sandwich & coleslaw & soda}. (Cc) Pr(soda with meal) = Reducing =— l number of meals with a soda 4 = =—, total number of possible meals 12 a‘ ; has the interpretation that one of the three beverages is a The event that doubles are rolled is the set of outcomes 1!1,22,33,44,55,66}. There are 6 ways to roll doubles. (b) Pr(sum is prime) = Pr(sum equals 2, 3, 5, 7, or 11) = 36" 36 2 SOLUTIONS TO EXERCISES — CHAPTER | 1-3 (a) Legend Third plain M&M package First peanut M&M package Second peanut M&M package /B | Peanut butter M&M package /D | Dark chocolate M&M package Almond M&M package There are two ways to list the sample space. If you grab two packages at the same time, then the outcome of DA represents one package of dark M&Ms and one package of almond M&M’s. In this case DA is the same outcome as AD (order does not matter), and therefore only needs to be listed once in the sample space. Then there are 28 outcomes: VA Eo, PONG ea NZ. 2 BaD i A, ei tos io Ng, 128, 20, P,A, Ps Ni, Ps No, PB, PD, P,A, NM Nz, NiB, MD, MA, N2B, N2D, NA, BD, BA, DA}. If you envision the experiment as grabbing one package and then grabbing a second package (in other words, order matters), there will be 56 elements in the sample space. (b If order doesn’t matter the answer is =. If order matters then the answer is — =. Of course 2 = = , so both perspectives lead to the same answer so long as you are consistent in your treatment of the sample space. (c) 18 036 a 56 Section 1.2 Finite Discrete Models with Equally Likely Outcomes 1-4 There are 3-2-6 = 36 current outfits eee Shoes Pants Shirts Adding shoes will allow 4-2-6=48 different outfits, adding a shirt will allow 3-2-7=42. You increase the total number of outfits most by adding to the item that you have the least. That is, if your mommy offered to buy pants, you could form 3-3-6=54 different outfits. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 3 1-5 ** CIEE Red White Blue (oye ae There are six choices for the blue die. (c) Pr(first die is 3 and second die is 4) = : - = + (d) No, tree diagrams are not so useful if tracking more than 50 or so outcomes. 10° LILILI-LILI-LJLILIL 10-23-26-23-10-10-10 One could try to use a tree diagram to systematically list all possible way to rob four banks. The thief could rob any bank first. He has only five choices of bank to rob second (he does not rob the bank he robbed first). Similarly, the thief has five choices of bank to rob third (it just cannot be the bank he robbed second, but it could be the bank he robbed first). The total is 6-5-5-5=750. 8-10-10-9-10-10-10-10-10-10 er Tap ov eed A ns (c) 7! 1 Sere USI aa (dy) 5 220 (e) 4!=24 (f) Undefined (g) 23°22-21 eat; aie 4 @ SOLUTIONS TO EXERCISES — CHAPTER | (hy 3,1 (1) ee 0) Gat kh) ) 7! == 31.4! 7-6-5 32 2e1 3s Ciel ‘EE Wik Cay Ken! 2 ey Kendra gets her own exam, the other 16 exams can be redistributed to any of the remaining 16 students. Kee ae aL” sP3 6P; = 6-5-4 12 Py r=23. 365 Fy. 2 .5 by trial and error. Solve 1—-==— 365! r=4. Solve l—- sae yt 7 Pr[people get offon unique floors] = 11ehP; 352 3512 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 5 oe n=5.-solve Veaamee .5_ by trial and error for the minimumn. oo hes (a) 7C6=ier lous (b) 3012 = W2'q T =13 13! = (c) 3CC| =——~ Tho 5! (d) 5C, = 371 =10 OnCe ol Remembertar olet. ! (g) Cs = 525 =10 a ee (1) 5Cs aol G) 5367C1 = a0 = 5867 (k) HE = T= 50,=35 (1) 0Co = aie Both expressions equal 220. n C, = ' n! (n-r)! = n! (n-r)!r! = Kner Placing r toppings on a pizza is equivalent to keeping n—r off the pizza. 6 SOLUTIONS TO EXERCISES — CHAPTER | _ 1-23 6C4 1-24 The multiplication principle says that the total number of possible pizzas is Number of ways to select toppings x ways to select size x ways to select crust (a) 13C2-3C1-2C) erCare Cis (DO 1-25 = 468 aC py (Cl) Oy sCrscr (a) iss (ob) 4Ci-5Ca-6C2 1-26 123015Ci 1-27 14C3-3C4-7C2-5C, 1-28 (a) e252 = 210 39C2-290C 2=8550 (b) To match exactly one white ball, your lottery ticket must match one of the two “good” white balls. There are 7C ;=2 ways to do this. The second white ball could be any of the remaining eight “bad” balls. 16 oC ® 4Cjyes match | match! good bad white white 1-29 (10C1) oOo oe 8550 = ee match both blue balls 6 (a) ——_——_ = .01997 60 Co choose | piece from 4 choose one ofthe chosen choose 5 variety w/2 varieties varieties pieces easy ae (b) 6Cs > 5Cy (15Cy) 4 choose2 Pieces for variety w/2 pieces = ge = 2697 60 C6 4Cy~ §Cy + 6Cq = 410-15 = 4G = 50y —— — centers forwards Gs — | ae —~— guards centers forwards guards starters PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT ee 15 15! Eee = Sh3h7t = 15s ° start 5 of the 15 1-32. 10 C3 eee ON 7Cq 3 of remaining 7 ofthe final 10 are on the bench 7 do not dress 4 ! The number of words is ,i ; = saa =12. The sample space of possible words is thoop, hopo, hpoo, pooh, poho, phoo, ohop, ohpo, opho, opoh, ooph, oohp} Pr(chosing HOOP) = a 1-33 5 ees ee : 3 : : = 30) rT) WeLahIbS Pol ohh ! 1-34 og la 1-35 @ 7 (a) Deal5 rer a nrers ae SSS Ses gtr Tanase cards to the first poker player. There are 52Cs ways to do this. Next, deal 5 of the remaining 47 cards to the second player and continue in this fashion. The total number of ways to distribute six 5-card poker hands is 52! s2C5 + 47C5 + 42C5 + 3705+ 32C5-27Cs. = SE 5151515151991 (b) The 52 playing cards need to partitioned into 6 piles of 5 and | pile of 22. There are 2 je eA = Ee EL|ETAL a, s to do this this equals | Cs + —— 2Cs Nee eens frat party soup kitchen : Cis ° Se stay home : = 25! PAD). AY 25! = Sem ZOESI Aolor le oNor lS: In terms of a multinomial coefficient, ; L 5 llth WN Nm 8 @ SOLUTIONS TO EXERCISES — CHAPTER | (b) This is similar to part (a), but we need to take into account that the 5 person teams are essentially the same. versus team green. That is, suppose the hoop game was team blue If the teams are {Amy, Bree, Candie, Diane, Edna} versus {Zoe, Yolanda, Xena, Willaminia, Venus}, then it does NOT matter which is team green and which group is team blue. 25 C5 + 20 Cs Z 1-37 30 C7 - 23C7 +2 1-38 9 ! There are L4 {= sche = 9C)-7C4-3C; Prowomen e total ways to assign the jobs. assign 2 women assign 4 of the 7 assign the remaining both bad jobs menthe average jobs 3 men the good jobs ———— a —|——_ : cet both bad 106s) Jom) 1 Cie —_——<——————— Note: This can also be solved as 7C4 3C3 9 Cz > 704-33 = l o>)ON since the problem doesn’t involve the distinction between average and good jobs. Section 1.3 Sampling and Distribution 1-39 There are 16 different side dishes. specific order: 1-40 age Order does not matter. The customer must select different sides in a 816 seelow: 14 5 ; first side : second side a ae third side + So, for example, the following 3-side choices are equivalent: {(1) (2) (3) (4) corn & dressing & mac-N-cheese, corn & mac-N-cheese & dressing, mac-N-cheese & corn & dressing, mac-N-cheese & dressing & corn, (5) dressing & mac-N-cheese & corn, (6) 2.53560 dressing & corn & mac-N-cheese} 3 nen There are 1¢Cs = me = 560 such orders of 3 side dishes 1-4] 16° = 4096 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 1-42 # 9 This is equivalent to distributing 3 indistinguishable balls (selections) into 16 fixed urns (side-dishes) without exclusion. 16,3-)C3 = 1g C3 =816 1-43 Ordered samples without replacement. See Exercise 1-39 and solution. 1-44 Unordered samples with replacement. See Exercise 1-42 and solution. 1-45 Pr(order 3 servings of the same side) = As 816 1-46 (a) 345-1Cs =7Cs (b) The parent gives each child one dollar bill. She is left with two dollar bills which she can distribute to her 3 children: 1-47 4C. (a) The fundamental theorem of counting implies the solution equals: Number of ways to distribute the Snickers x Number of ways to distribute the gum 13Ci9-10C7 (b) After giving each trick-or-treater a Snickers and a pack of gum, she has 6 Snickers bars and 3 packs of gum to distribute in any way she likes. 6x4-1C6 - 344-103 — 9 Ce. 6 C3 1-48 Each child could receive multiple pickled beets. There are j9,s_;Cjo = 14C\o ways to give beets. 1-49 A bridge hand consists of 13 cards dealt from a standard deck of 52 different cards. 5213 52 Cs It depends upon the number of chicken sandwiches. If you do not order a chicken sandwich, then you have 5 dollars to distribute to 5 different $1 items. There are 9Cs ways to do this. If you order exactly one chicken sandwich, then you are lett with 3 dollars with which to buy $1 items. You either get zero, one, or two chicken sandwiches. 9 Cs 1 7C3 1 5 Cj 36 C21: 43.C21 10 # SOLUTIONS TO EXERCISES — CHAPTER | Section 1.4 More Applications 1-53 (a) (xt3)* = Ix4 44-23-3146. x2 -32 £4-¥-33 41.34 Niele 454% Ose (b) (2x+y) oe| 4 1-(2x)? +5-(2x)*- y +10-(2x)> - y? +10-(2x)*-y?+5-(2x): y+ +1-y? = 32x° + 80x*ty + 80x73 y? +40x7y? +10xy* + y° (Cy 1-54 3) (4x) + (-Sy))P° = 1-(4x)? +3-(4x)? -(-Sy)! +3-(4x)! -(-Sy)? +1-(-Sy)? 64x? — 240x7y + 300xy? -125y3 II n The key here is to write out | n—-| and | }and then combine using a common r- r denominator. (A oe r—-1 — r (n-1)! oe er)! (n-l)!r — (n-1)! iF Ter P (n-rybr-(r-)! (n—-1)!\(n—-r) (n—-r)-(n—r—-DeEr! (n—-l)!r Fi(n-1)!(n—-r) (n—r)tr! (a=r)brl _ (a-Ibr+@-Di@=r)_ - leand y=l. (1+1)” A OF +. a _a! _{”) ~ (n—r)br! GY. ) 1-55 betix 1-56 Selecting r items from n possible toppings to put on the pie is equivalent to selecting n—r to keep off of the pie. 1-57 = (n—r)!r! + nO? see Gas ; n n n+l ae : Since |}+| = | Joincoefficient of 4 5 5 : n+] io yoes = x yitt)-s s| = 3,876+11,628 = 15,504. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 11 Glee n Alternatively, -< n = O88 3,876 faa 5 = 3. Thus n=19 and the coefficient of 4 ayes = x°y!> in the expansion of ep = (x+y) is 20 J1s.s04, 5) (ay @=2ye5z)) = 1.x? 43-2. (-2y)43-x2+(52) +3-x-(-2y)? +3-x-(Sy)* +.6-x-(-2y)-(Sz) +1-(-2y)} +3-(-2y)? (52) 23 Capen 2) = x3 —6x*y+12xy? —8y3 +15x2z —60xyz + 60y2z 4+ 75xz* —150 yz? +12523 (b) (w+x—y+2z)* = 1-w? 4+1-x? +1-(-y)? +1-(2z)? +2:w-x+2-w-(—y)+2-w-(2z) \\ ake (=) ex ogee Fs Pe hee ci Jot oO) Net +U\Z - Uy 22) 2 (—y) (2 a pa + | WZ Zz. Detailed calculations follow the exércise in the text. 0 Each of the five dice has 6 possible outcomes. So the total number of ways to roll five dice is 6°. (a) Pr(Five dice-of-a-kind) = sts. =.00077. ways to select 4 of the remaining die 4 of what kind? the Sdieto match must be different ee, —__——_—, een 6C\ 5C4 (b) Pr(Four dice-of-a-kind) = 4&; 6> 5? = 64 = 01929, (c) There are 2 possible straights, 1-2-3-4-5 and 2-3-4-5-6. many ways these straights can arise ina roll of5 dice. HE Pr(straight) = 6> It remains to count how | as ssn 6° 12 @ SOLUTIONS TO EXERCISES — CHAPTER | (d) A full house would have three dice be of one value (1, 2, 3, 4, 5, or 6) and two dice of a second value. 1 Wh S chal,i hoe PrFull house) = Hl46>0 5 10.5. one = 03858 (ec) The only way to get nothing 1s to have dice of the form: 1-2-3-4-6, 1-2-3-5-6, 1-2-4-5-6, or 1-3-4-5-6. It remains to count the ways these can sequence can be arranged on 5 dice. 4.5! 480 Pr(nothing) = a eer = .06%73 Bek pag 3 Oa: aticlie ieee as aA ("a = So (f) Pr(Three dice-of-a-kind) = = aea aa ZA 2a. 2 ACE 8 B 4 " b ge (ey ePIC Eo pair ie ee 6° (h) Pr(one pair) = 6° a B PA 5 ea a aea" geal| 4010-3: a 63 = wee = 15432 = 46296 It is neat that you are more likely to get 3 dice-ofa kind than you are to roll nothing. Did you check that the sum of these probabilities is one? 1-61 1-62 Solutions tabulated in the text. Expected Winnings l = 49,999,999 .———____— 120,526,770 4,00 006) 2 Sy, OSes 20526 oon eet sAlis 0) ene That is, the expected winning is negative 41 cents per play. ee t). 117,184,724 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 13 Section 1.5 Chapter 1 Sample Examination L. Order does not matter. 2 799-104 3: Order matters here. 4 2. 15 P; C2 8G give slinky's away to 2 of the 15 students give bop-on-nose to | of the — remaining 13 ee Both solutions are 5Cs; = 65,780. rei =| 365 2— TY 14C2 give bop-on-nose away first give slinky's away to 2 of the remaining 14 students (a) 51Ci7 2 pair and 13 other cards without the old maid two pair, the old maid, and 12 other cards SSS SSS (b) 6. 25 Gg Ga-9 C7" 33Cig (4G)> gh. Gor Cy ya 7 93 Cp -(2C)) 12 (a) 32Cs (BD) aC ae fo which suit? —S Cin< 3500) —— a ways to select 4 the remaining cards from the selected suit if Ca Cog 51C17 card (a) Cae 6. C2 : 162663 (b) Pr(5 kids selected included 3 soccer and 2 football) = acetate A408. 1305 8. Using the multiplication rule gives 3-2-3=18. s) 49Cs + 42C; = 80,089,128. The Jackpots were not getting large enough, so there were more balls added to decrease the chances of winning and therefore increasing cumulative Grand Prizes. 10. (2x—y)> = 32x> —80x4y +80x°y? — 40xy? +10xy4 - y° 14 LI; SOLUTIONS TO EXERCISES — CHAPTER | There are 15 terms in the multinomial expansion. (x—2y+5z)" = 1-x4+4-x3 -(-2y)+4-x3 -(5z)+6-x? esto -+1-(5z)* 12 cards have faces —t Pr(all 3 cards have faces) = Az 7 = 00995. 3) ways to match ways to match white socks black socks ——— Pr(matching socks) = 8@ SS mai 42 a 2 8-4 =1-Pr(one white and one black sock)=1— 7 There are a total of 100 birds, therefore j99Cs ways to take six birds. Joe could take 2 Canada geese (and one from each of the other prey), or 2 ducks, or 2 eagles, or 2 cranes, or 2 flamingos. This can be done in 20C2 + 25Ci- 401-101 5Ci + 29 Ci1* 95 Co = 49Ci 0 Ci= 5 Ci + 29 + 25C1- 4o€2 > 10C1 +5 Cy + 20Cy > 25Cy > ani «10 C2 > 5 Cy + 99C) + 25Cy + an Cy - 19 Ch - sC2 = 47,500,000 ways. 47,500,000 bird ofof each each variety) £192,052. 400 = .033 985. Pr(atr(at least leastone bird variety) =~ ee Pr(passing on first attempt)= 1oS) Use the multiplication rule with the following steps: Step 1. Step 2. Step 3. Step 4. Choose any 3 rows (order immaterial) - # ways = 6C; =20. List the rows in ascending order for definiteness. Choose a column for the first listed row -# ways =5. Choose a column for the 2™ listed row -# ways =4. Choose a column for the 3" listed row - # ways =3. Answer = 20x5x4x3 = 1200 (C) CHAPTER 2: GENERAL RULES OF PROBABILITY Section 2.2 Set Theory Survival Kit 2-1 (a) (b) (Cc) (d) O=rl3 od.7-9} and, Pat 2 aio. The odd primes less than 10 are OM P = {3,5,7}. The set of odd numbers or prime numbers less than 10 is OMT Pee 28 5.7.9. The odd numbers less than 10 that are not prime is the set Age (a) (b) B ={az water}, “so N(B) =2. (c) AUB=(1,2,z,Jamaal,gum}, (a) The Luisi family pets can be summarized as: (b) The number of pets of each type is summarized as: so (AU B)’ = {water}. Named \ O— P = {1,9 Sf: 16 ® SOLUTIONS TO EXERCISES — CHAPTER 2 2-4 The shaded section denotes the indicated set. 2-5 From Subsection 1.4.3, we already have the probability of the winning prizes. Pr(losing) = 1 — Pr( winning something) > Rye le . 120,576,770 ae 120,576,770 1,712,304 probability of winning the grand prize probabilty of matching only the powerball _ 117,184,724 120,576,770 ° PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 17 Doubles = {(1,1),(2,2),(3,3), (4, 4),(5,5),(6,6)}. Sum is divisible by 3 = {(1,2),(1,5),(2,1),(2,4),(3,3),(3,6), (4,2),(4,5),(5,1),(5,4),(6,3),(6,6)}. Doubles - Sum is divisible by 3 = {(3,3),(6,6)} Doubles U Sum is divisible by 3 = {(1,1),(1,2),(1,5),(2,1), (2,2), (2,4), (3,3), (3,6), 1) a) 0,1), (024), (,.2), (0;3),.020)} Pr(Doubles U Sum is divisible by 3) = + = < + oa— = Let K equal the percent of king size mattresses sold, O equal the percent of queen size mattresses sold, and 7 equal the percent of twin size mattresses sold. We are given: K+Q+T = 100%, Solving for the variables, K Q.= LK +7), ane Kae 3 2 =60%, Q=20%, and T = 20%. The probability that the next mattress sold is king or queen size is K 2-8 A’ = {2,4,6,8,10} B' = {1,4,6,8,9,10} AUB ={1,2,3,5,7,9} ANB =33,5,7) A'UB' =(1,2,4,6,8,9,10} A' 0 B' ={4,6,8,10} (AUB) ={4,6,8,10} 2-9 A-B = ACB’ is represented by the shaded area. A B +Q = 80% (C). 18 D210. SOLUTIONS TO EXERCISES —- CHAPTER 2 “Cte eee (4UB) =4'0B' 2-11 AU(BOC)=(AUB)A(AUC) 2-12 N(AUB) = N(A\B)+N(AQB)+N(B\A) [Nv(A\ B) + N( AMB)|++[ N(B\ 4)+N(ANB)|-N(AOB) = N(A)+N(B)—-N(AMB) _ PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 19 2-15 (4 Bo c) is the shaded area AM(BUC)". 2-14 Method 1: Using the Axioms N(A) N(B) 6 = N(U)-N(A). 10 = N(U)—N(B’). From the inclusion-exclusion relationship, N(AO N(4U B)+N(AMB) B)=6+4+10-12=4. Method 2: Venn Diagram Method 3: Venn Box diagram (Subsection 2.2.4) = N(A)+ N(B). 20 @ SOLUTIONS TO EXERCISES — CHAPTER 2 2-16 NM(AUBUCUD) = N(A)+N(B)+N(C)+N(D) —-N(ANB)-N(ANC) —-N(AND)-N(BOC) —-N(BAD)-N(CAD) +N(ANBOC)+N(ANBOD) +N(ANCND)+N(BOCOD) —-N(AQNBOACOD) There are 4C, =4 ways to select one of the four regions. There are 4Cz =6 intersections of two of the four regions. There are 4C3 =4 intersections of three of the four regions. . 2-17 A four region Venn diagram could be drawn as: 2-18 There are seven male children without purple hair in this family. Adult Female Bin Purple hair PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 2-19 (B) 50% = 40%+10% of the population owns exactly one of auto and home. Auto_owner 2-20 (D)_ Home There are 880 young Tharried females. Married 2-21 (D) 52% of these viewers watch none of the three sports. Gymnastics aN yy Soccer Baseball @ 21 22 2-22 SOLUTIONS TO EXERCISES —-CHAPTER 2 (A) 5% of the patients need both lab work and are referred to a specialist. The percentages in bold are given in the problem. Try this just using a Venn diagram. Lab work | No lab work Referred to Specialist | 5% 0 Bier ae Not Referred to Specialist 35% 60% 2-23 (D) Pr(A) =.60. The key here is to note that Pr(AU B) = 0.7 implies that Pr(4’ B') = 0.3. Similarly, Pr(A U Be)=0.9 implies Pr(4'n B) =0.1. These are versions of De Morgan’s laws. We do not have enough information to find Pr(A A B) or Pr(AB'), but fortunately we are just asked to find Pr( A) = 1—(.1+.3) = .6 A (D) B We have two equations with two unknown variables x and y (see box diagram below). The first comes from the fact that the sum of the probability of four disjoint outcomes must equal 1. The second equation comes from the statement of the problem. 22%+x+y+12% 22%+y = 100% = 22%+x+ 14% —.,—_——“~- We —-—- probabilty visit chiropractor probabilty visit PT. exceeds by 14% x = 26% and y=40%. implies that x+y imphlesthat y = x+14%. Pr(visit P.T.) = 66%. = 22%+26% = 48%. Chiropractor 22% + y No chiropractor x+12% 22% + x 100% PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 23 2-25 Draw the Mickey Mouse diagram, with Auto, Homeowners and Renters labeled as shown below. Since H and R are mutually exclusive the subsets comprising their intersection have cardinality zero. Statement (i) gives N [AUBU cy |Sa, Statement (v) gives N(H \ A)=11. The remaining unknown subsets are labeled eye Z AWN From (ii) x+ y+z = 64. Also, 100 = 174+114+(«+y+z)+w => we 8. From (111), y+11 = 2(x+8), and, From (iv), x+y = 35. Solving the last two equations simultaneously gives Section 2.3 2-26 = 174+11+64+w x = N(AXMR)=10 (B) Conditional Probability There are 3 levels of choices in the following tree. The first branch concerns whether the professor put the batteries in an empty pocket or in the pocket with the two new batteries. The second level concerns which pocket the professor withdraws batteries from. In the third level, if the four batteries are in the same pocket, we will use counting techniques from chapter | to determine the likelihood of grabbing either two batteries that do not work, two batteries that do work, or one battery that works along with one battery that does not work. 24 SOLUTIONS TO EXERCISES — CHAPTER 2 Professor selects Professor places dead batteries in empty ; Select two Bee . 1 pocket with dead eee batteries : 3 we : Professor pocket er Select two selects pocket 1 __ good batteries with good batteries 2 ‘alt Select two dead batteries Professor places dead batteries in pocket 1 with 2 good batteries. 4 oe | gi ees l Fe 1 4 elect one professor selects beket with all :attenes dea ie and one good battery Select two good batteries (a) It Pr(get at least one good battery) = =.Bers 1 ») eee eee (b) Pr(two good batteries at least one good battery) Peay _ G Pr(two good vat least one good battery) Pr(at least one good battery) 2 Pr(two good) Pr(at least one good battery) ; mall Tiga, ae 4 24 (c) l able l play m0) 4 Pr(batteries in same pocket|at least one good battery) Pr(batteries in same pocket Mat least one good battery) Pr(at least one good battery) The audio device will play PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 25 2-27 (a) This is the scenario where we will keep our original door. You will win the car if and only if you selected the correct door originally (it just doesn’t matter what Monte does). There is a one in 3 chance of this. Pr(win auto|keep original door) = +. (b) If you originally select the door with the new car, then Monte can select from either of two doors to show you a goat. If you select a door concealing a goat, Monte only has one door that he can open. Consider the following diagram. shows you Monte : You pick a door hiding a If you change doors, ? ei the goat below l you win a new car goat ou pick a door hiding a Monte shows you aA goat the goat above If you change doors, > | you win a new car ' 3 If you change doors, Monte shows you a door hiding the first goat you win a goat vof— shows you the second goat Pr(win auto change doors) = ~ i)[o <) ee w+ a new automobile If you change doors, you win a goat vi- ‘L-1+4-1-1 = veto |= Consider the following Box Diagram with the information in (i), (11) and (111) filled in: b = 1000-14-22 = 64 From (iv), > High] = Pr[High|[rregular]- Pr[Irregular] = +-15% — hi a = Pr{Irregular Similarly, from(v), c = +: 64% = 8%. Regular Irregular High | Low | Normal ee Row Total oy SES ies 85 5 2 8 15 100 Column Total Pr[Regular ~ Low] = 20% Completing the Box Diagram gives, (E) | 26 @ SOLUTIONS TO EXERCISES 2-29 - CHAPTER 2 (C) Let x equal the probability that the male is a smoker, given that he does NOT have a circulation problem. Pr(circulation problem] smoker) = Pr(circulation problem ~ smoker) Pr(smoker) oe ak 5 5 Oe ee 4 ae Smoker 2x Smoker x 25% circulation problem 2-30 Cyne (C)_ 2-31 (B) Pe oS ees : Pr(smoker A died) .10-.05 .005 ker|died) = ————-___—. = —_——____—_____ = —— _ = 36° Pr(smoker/died) Ne Pr(died) <N= oe 10-.05+.90-.01 iS 014 ] oe = Pr(N <4) = ; Ni : + ms + a 3 + - a +35 +96 a] role ; | h233 2 6 et. : PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 2-32 @ 27 (a) Pr(Penguins win series) =.55-.35+.55-.65-.55+.45-.35-.55=.47575, (b) Pr(Devils win at least one game) =1—.55-.35 =.8075. \ 35% |Pittsburgh wins the series 2-0 Pittsburgh wins 55% | Pittsburgh wins New Jersey Pittsburgh wins wins : ew Jersey wins 45% New Jersey the series 2-1 ‘ ae Pittsburgh Pittsburgh wins wins Game 2 played : in Spc New Jersey New Jersey —. Game I played in ne Game 3 played in Pittsburgh, if necessar VW 65% eM CLSeD 2-33 New Jersey wins New Jersey wins the series 2-0 (B) This is a standard Bayes’ formula problem, easily solved by drawing the appropriate tree diagram. The data may also be displayed in a Box diagram similar to the one in Exercise 2-28: | | Critical |Serious |Stable |Row Total | Sie a ee a Lert >oa lr ec (SS el ae 10 Column Total 30 d 100 d = 1000-10-30 = 60 a = Pr[Die Critical] = Pr[Die| Critical] Pr[Critical] = (40%) -(10%) = 4%. Similarly, b = (10%)-(30%) = .6%. = 3%, and c = (1%)-(60%) the box gives, vai fon 10 30 Critical |Serious | Stable |Row Total Z 59.4 92.4 Column Total 7 Pr| serious } survive i Pr serious|survive =—= L Pr| survive | 60 J oro a7 100 wal ayNaicaaeae ha ey Ee B (B) Completing 28 @ SOLUTIONS TO EXERCISES 2-34 — CHAPTER 2 — Pr(exactly 1 hit in 3 at-bats) =.25-.6-.75+.75-.25-.6+.75-.75-.25 =.365625. 400 Hit 400 Hit No hit 600 250 250 =F Hit No hit -600 1 No hit .750 Hit No hit Hit .400 No hit maul — at Hit First at-bat 250 750 —<. No hit Second at-bat No hit Third at-bat 2-35 Pr(call “dad” ~ Packy is your father) Pr(Packy is your father) Pr(call Packy “dad” |Packy is your father) e. .01-.90 .01-.90 +.99-.05 One Car | Multiple Cars | Row Total Using (iv), a = Pr[Sports Car > Multiple Cars] = Pr[Sports Car |Multiple Cars]Pr{Multiple Cars] = (15%) -(64%) = 9.6% The complete Box Diagram is now easily calculated: 2 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Thus, Pr |One Car 7 No Sports Car|=25.6% Note: @ 29 (C) One may also display the information in a tree diagram: Insure a sports car Does NOT .64 15 insure a sports car .85 Insure a sports car exactly | ca Section 2.4 Independence 511 Pr(Male-Male) =.511* =.2611 Female gq Pr(Male-Female) = .511-.489 =.24988 First person chosen 511 Pr(Female-Male) = .489-.511 =.24988 499 Pr(Female-Female) = 4892 =.2391 489 Second person chosen™ 30 ® SOLUTIONS TO EXERCISES = CHAPTER 2 2-38 | You are encouraged to observe that this is structurally the same as problem 2-37. 2 z Heads >) Pr(Heads-Heads) = (2) ae, 3 2 . 3 Head Tails ae Pr(Heads-Tails) = eae 2 Px(Tails-Heads) = 44 : = 3 é paresip 3 Heads Tails l 3 First coin flip | Maes pyTails-Tails) = (+) = 3) Second coin flip Number of Heads | 0 | l Probability 9 cathe thoagh the srobsbibiy tree that have one head and one tail 384 tee Make Makes his 3 point attempt 384 384 Miss 616 Nree Misses his 3 point attempt Pee 1 Make : 616 Make 384 Miss 616 Make 384 eit. 616 384 John Stockton’s _ Miss 616 Make 384 Miss 616 616 Ea first attempt second attempt third attempt Pr(Stockton makes exactly two 3-pointer) 3 384 .384 .616 ways to make exactly two 3-pointer on three attepts probability of making a 3-pointer probability of making a 3-pointer probability of missing a 3-pointer = 2725 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 31 2-40 iy Pr(B| A)=—. 2-41 (a) Since Pr(B)=b, we know Pr(B’) = 1—b. To show that4 and B’ are independent, we observe Pr(4> B’) = Pr(A)-Pr(B’) = a-(1-b). Pr(B) api 51 The events A and B are dependent. 52 Q Ca [a RRS w=) 2-42 ‘Not necessarily. 2-43 (a) (b) B B' b eoah) C = {HHH, HTT,THH,TTT}. Thus, Pr(A) = Pr(B) = Pr(C) = 4. (c) Pr(AB)=Pr(HHH,TTT)=2-=1. Pr[A 2-44 Similarly, VC] =Pr[B AC] =Pr[HHH, TTT) =. (d) Yes, because for example, Pr(AM8B) = + = +. (e) Pr(4 (f) ING Weiter Arie Let x=Pr(AM 8B). Using independence, solve (.2+ x)-(.3+x) SGIUUONS, X= =ePTCA)- PTC): BOC) = Pr[HHH,TTT]=+. Cyt ee = SPs A] PLB) Pr. = x. There are two 2 0lL =. Pi eet Obar 0. Since the coin is fair (all outcomes are equally likely) and we do not need to track probabilities, listing the sample space will be sufficient. 2-45 U =|HHH, HHT, HTH,HTT,THH A = first coin 1s tails ={THA.THT B =the third flip is heads = THT ,TTH,TTT. TTH,TTT}. {WHH, HTH ,THH ,TTH }. (b) Yes, the events are independent. 32 @ SOLUTIONS TO EXERCISES — CHAPTER 2 2-46 (a) Now is when tracking probabilities through a tree diagram is critical. PrANB) =_ 4-.6-.6+.4-.4-.6 = 60%. agPr(B| A)a = ———. Era) (c) 2-47 A Yes, the events are independent. You can see this by computing Pr(B) = Pr(third flip is heads) =.60. Let O be the event of using orthodontic work, F fillings and E extractions. Using (1) and (11) we have: tele eae Ss Gerth Ol) Primal 1—21o—9 b=j12=1)2. By independence, Completing the O and F box gives: 3 d= be=>c=a/b=213- Row Total 1/3 Bes) O | O' | Row Total | E 1/4}1/4 le2 E' [1/4 ]i/4] 1/2 |Column Total | 1/2 | 1/2 We now use Pr| F |= ly | | anc Pri Efe l /2, plus (iv) to fill out the F and E box: ip F E Row Total LS [ 1/8 | al E' 7/24 Li 1/2 Column Tota! Note: F and £ are not independent. Pr| EU F )=1-Pr| EV F'|=1-7/24=17/24 (D) l PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 2-48 33 We take the statement that a “different brand of pregnancy test was purchased” to imply that these test results are independent. Pr(both tests are incorrect) =.01-.02 = 0.0002. 2-49 First we fill in the missing information. ivajor | ee Grades yeas | TAT BIC] DE |Totals | | Business| 2| 3 |3 [5 [1] 14 | PSs lem ee | Totals |9 |10}4[/5[3] 31 | (a) Pr(student assigned ‘B’) = =. students - business majors : b (0) : : number of science majors — 17 Pr(science major) = -———_-_—__>—_ = — = \ yor) number of students on (c) Pr(science major ~ assigned ‘B’) = +. (d) This is most efficiently calculated by looking at the table. We are told that the student is a business major. Therefore the student lives in the shaded row of the table. Pr(assigned *C’ b usiness 2-50 major) 31-14 at 3 et is 1 (e) Pr(business major assigned 'A') = =. (f) No, because (a) Wewill use counting techniques (combinations) that we studied in chapter 1. The number of ways to select 2 batteries from a flashlight containing 5 Pr(assigned *C’ business major) # Pr(assigned *C’). batteries is 5C> =10. Pr(both batteries lost their charge) = at eho: (b) Pr(exactly one battery lost its charge) = : = (c) Draw a tree diagram with appropriate probabilities. > Pr(exactly one tested battery lost its charge) = 2- bo in| 34 2-51 SOLUTIONS TO EXERCISES—~CHAPTER2 ~ Let Pr(purchase disability coverage) = P[D]=.x, then Pr(purchase collision coverage) = Pr[C’] = 2x. 0.15=2x? implies that x =.274 and 2x =.548. The completed Box Diagram is: ee KE oa [ea Sonn Tour]sasa2[1 Pr(select neither insurance) = Pr aoN D'] = 328. 2-52 (B) Pr(select two consecutive face cards) = =-- ae or using combinations Pr(select two consecutive face cards) 2-53 2-54 number of ways to deal 2 face cards 12C2 number of ways to deal 2 cards 52C2~ (a) Pr(next roulette spin is black) = 4. (b) 8 Pr(next eight roulette spins are black)= (48 = U02 53: (E) Pr(no severe and at most one moderate) = (0.5)? +0.5-0.4+0.4-0.5 =.65. minor moderate severe 0.5 0.4 First Accident Second Accident”! PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 2-55 (A) Let B denote the number of blue balls in urn 2. Pr(both ball same color) =.44 = (0.4)- 16 16+B B + (0.6): _ 6.4+.6B [6% Bael6e Solve for B to find there are four blue balls in urn 2. 16 16+8B 16 16+B a4 10 Red ball Blue ball B Waa Red ball 16+8B Blue ball 16+B 16 Blue ball 2 ae 10 B rm Urn 2 Section 2.5 2-56 Bayes’ Theorem Draw the standard tree diagram to track probabilities. ~ 51 Ace pais on 52 Non-Ace 48 uy 3 2 Non-Ace 51 Sade 51 First card Non-Ace A 51 Second card : number of aces - Pr(first card is an ace) = (b) Pr(second card is an ace}first card not ace) = number ofcards 4 a (a) 52 35 36 @ SOLUTIONS TO EXERCISES —CHAPTER 2 (c) Pr(first card is an ace|second card not ace) Pr(first card is an ace ~ second card not ace) Pr(second card not ace) Be ° AGS AS ip Se FAT Sil 2-57 40 11Face 50 a 40 eh See 5D Non-Face Face sy) First card Non-Face Face 11 Ae 50 oh 51 on-Face Non-F 39 50 Face 50 51 Non-Face = 39 Face rr 12 = Second card & Oo 50 12 ae 5 Non-Face Third card Number of Face Cards Probability .F¢ A 0 9,880 22,100 9,360 22,100 2 2,640 22,100 + 4 220 22,100 2,860 (b) Pr(at least 2 face cards) = F100" (c) Pr(3" card is face|first two cards are face cards) = Pr(all three face cards |first two cards are face cards) = e PROBABILITY AND STATISTICS WITH APPLICATIONS: (d) A PROBLEM SOLVING TEXT # 37 Pr(all three face cards| at least two face cards) ___Pr(allthree face cards) 220 — 22,100 _ 220 ~ Pr(at least two face cards) 2,860 2,860° 22,100 (e) Solution Method |: Tree diagram and Bayes’ Theorem Pr(at least 2 face cards| last card has face) a Pr(at least 2 face cards ~ last card has face) Pr(last card has face) this is the probability that the first and third cards have faces . iP Mim = 1 0Nt B40 gio 5ST 50 Bes 50) Oh oD Py i Pane ol Th eee 3) GS i i SS ee = This equals S Does it make sense to you? — 11880 _ 3999 30600 Solution Method 2: Using Combinations Imagine that you know that the third card is a face card. That leaves 11 face cards in a deck of 51 total cards. Now imagine dealing the first two cards from this reduced deck. The question asks us to find the probability that at least one of these cards is a face card. 40 C2 se Hey Pr(at least one face) = 1— Pr(two non-face cards) = 1— NE Ge 5102 38 @ SOLUTIONS TO EXERCISES— CHAPTER 2 Section 2.6 Credibility 2-58 Ay 0.10 Good driver Bad driver Type of driver A 0.50 Accidentor not in year| Accident or not in year2 (a) PrA. As) =.8201)* 2-5)? Cha =.058. Pr( 42 OA SoG Lyhed) SA bee pete mee (c) Pr(G|A ro yy Pr(4,) 28.5 (1) ba2ed) O58 LEE 18 90 ese ater boa Pr( A, C SS 29 >. (1a. 2 < (d) Pr( BlAt least one accident) = ae Lee aaa .2-(1-.5*)+.8-(1-.9°) 302 497. PG A Az) As A Pr(A, 7 A>) mel. )s + .2-(.5)- 008 imi O58 Ales: (e) PROTAro AAS A>) 8 .(.9)2 he) _ .8°(9)*4+.2:(5)e 2-9 a) Pr(1997|1997 or 1998 or 1999) = =e .16- 2-60 (D) Draw atree diagram. _ .6482 — 9284. ~ .698 ae (.05) +.18-(.02) + .20-(.03 ) Pr(ultra-pre ferred dies) = = 455. (0.10) -(.001) =.0141. (0.10) -(.001) + (0.40) - (.005) + (0.50) -(.01) a PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 39 2-61 (D) Drawa tree diagram. (.16)(.08) Pr( young adult collision) = (.08)(.15) +(.16)(.08) + (.45)(.04) +(.31)(.05) — 2-62 (D) Let xdenote the probability that a non-smoker dies. Dies Heavy 20% Smoker |. Lives 1—4x Dies Smoker 4x 2x 30% Lives 1-—2x NonSmoker 50% Type of smoker l-x Live or Die? fi fo ce Pr( heavy smoker}died) = (.2) +(.3) (4 (2x) x) +(.5)x lp 19 (.08)(.06) it)1 lon)ios) (B) = Pr(16—20jaccident) = (.08)(.06) + (.15)(.03) + (.49)(.02) + (.28)(.04) 158 40 SOLUTIONS TO EXERCISES Section 2.7 - CHAPTER 2 Chapter 2 Sample Examination 1. A Z. B Pr(dealt a fullh NE Cus) houses eeber of full ee number of 5 card hands me 13Ci C519 Ci aC a 3744 52Cs 3. 2598960 © You may wish to create a Venn diagram. The key is to note that that you must have some insurance to be a client. Auto Insurance No Auto Insurance Dismemberment | No Insurance Dismemberment il7/ 45 (a) N(both auto and dismemberment) =17. (b) N(exactly one kind of insurance) = 65 = 2( a aw NM (c) No, because Pr(auto ™ dismember)= 17 62 go * 95" 89 = Pr(auto) - Pr(dismember). PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 41 (a) Venn Diagram U =College class Engineers Female Venn Box Diagram (b) Pr(female > engineer) = a oo (c) Pr(femaleU engineer) = ca = pit a a2 by) (d) Pr(female| non-engineer) = =. (e) No, since Pr(male Mengineer) : Ke ea aey) # Pr(male)- Pr(engineer). number of ways to be dealt 3 Kings = : = number of ways to be dealt 3 cards ES (b) Pr( NOT dealt 3 Kings) = 1 Cy ce Oy 22100" 9) (c) Pr(3™ card King|I‘' and 2" card King) = av (d) Pr(3 Kingsat least 2 Kings) = 4C3 ae ancy =ACTON = 4C3 Z 52 C3 = 4 22100 e 42 6. SOLUTIONS TO EXERCISES — CHAPTER A) (a) There are 16 elements in the set.: The bold numbers denote a sum of 6. ‘a= {22,23,24,25,32,33,34,35,42,43,44, 45, 52,53, 54,55}. (b) Pr(sum is nine) = =. (c) Pr(at least one die is 5) = (d) Pr(sum is 6at least one 5) = ws, 9° (ec) Pr(at least one 5|sum is 6) = we 3 It (a) (b) Pr(sum is 4) = (c) Pr(at least one 2) = (d) Pr(sum is 4) ball from Urn 4 is 2)= Pr(ball from Urn B is 2) = i st (e) Pr(ball from Urn B is 2/sum is 4) = ~ = PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 43 8. Se] Coke i |Ripple | Water |Red Bull|__| eager [alte dt oaMeeSESEY eee [aes |e facia lela Ma cl PE arian SO Oe ese es Se 0! | b) (b) 5 Pr(Red Pr(Red Bull) Bull) =—30 (c) Pr(female| milk) = N(female 4 milk) 8 Namilk) SG d) Pr(male U milk) = 22. 307 (a) Pr) =07) (b) Pr(A’)=0.6. (Cy PH Arm.) =.255 (een eyPr(AA B) Pr(B) (eyo Pr AGB ORCI )\=HOs. Pr(B es eerra A) _ (g) No, they are dependent. ‘40. Pr(4 7 B) # Pr(A)- Pr(B). 10. Assume for definiteness that card with 5 credits is in the left (L) pocket and the 4 credit card is in the right (R) pocket. She could either spend all 5 credits in Z, or spend one from L and 4 from R, with the last one coming from R. These actions can be displayed as the following 5 letter words using LZ and R: LLLLL, LRRRR, RLRRR, RRLRR, RRRLR. 5 l 5 Each word has probability Gy , and so the answer 1s 5- +) ‘= 32 5 44 RL. SOLUTIONS TO EXERCISES —CHAPTER 2 Assume first that the left pocket is depleted leaving 3 credits in the right pocket. This means a total of 6 credits are spent, | from the right pocket, 5 from left, and the last one used is from the left pocket. This can be portrayed using 6 letter words with 1 Rand 5 L’s, ending in L. That means the first 5 positions in the word consist of 1 R and 4 L’s. There are 5;C; =5 such words. Next, assume that the right pocket is depleted leaving 3 credits in the left pocket. This means we have 6 letter words ending in R with 2 L’s and 3 R’s in the first 5 positions. There are sCz =10 such words. 122 Thus, the total probability is Consider what can happen to the team that is leading 3 games to |. 50% Series won 4-1 Win 50% Series won 4-2 Win Lose 50% 50% Series won 4-3 50% Series lost 3-4 Lose Game 5 50% Game 6, if necessary Lose Game 7, if necessary Pr(win) =1—.5° =.875. 13; Pr(win) =1—.43 =.936. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 45 14. Column Total pes b=1-1/3=2/3 and a=Pr| AN B|=Pr| 4] B|Pr| B]= Tee Lk Complete the Box Diagram to give: Row Total B' eS Beal 13 Column Total | 2/3 y 1 Then’ Pra B)=1—2715=13 715. tS. Begin by writing out the sample space of possible dance couples. Betty’s Danielle’s Dance Partner Dance Partner Danielle Dan Danielle | Boris Dan | Danielle | Andy Andy | Danielle | Boris Pr(3 wives dance with non-spouse) = 2. U = fabed,abdc,achd,acdb, adbc,adch, bacd, badc, bead, beda, bdac, bdca cabd,cadb, chad ,chda, cdab, cdba, dabc, dach, dbac, dbca,dcab,dcba}. Pr(all wives dances with non-spouse) = Nae tO 46 @ SOLUTIONS TO EXERCISES We -CHAPTER 2 There is one good young teacher. U = math faculty at School University Old 18. Bad Pr(first person wins) Wei Sea Nae ‘On 5 Wy AY 5 Red bal i Se 4 Red bal White 4 ball The First ball selected =|~ Red bal = White ball IZ Second ball selected, if necessary. sak, l White ball Third ball selected, if necessary. 19. (a) i 2} Pr(1* person wins) = Bama +| 6 6 7 -—+,..=40% ¢5 Use geometric series. (Oy > + ers person wins) =. G26 L978: eos CT COeTera PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT ® 47 20. To solve for x, we know: a =Pr(A44 BAC|ANB)= PHA BAC) = ex 0G 3 Prodan yee ao ee Pr(no risk factor| A)= =~ =.46. Zzby, (C) (C) We have four equations with four unknown variables: 1 5 it : a , and w+x+y+z=l. ee 4 Vg et e 12 e Adding the first three equations together, we see 2D 22 = => Xxytzy = |nN — Pr(no coverage) = w=1-(x+ y+z)=1 = = 5 48 @ SOLUTIONS TO EXERCISES —-CHAPTER 2 De. Let A represent the event of heart disease. Let B represent the event of at least one parent with heart disease. Then: Pr( 4]B’) = —_ = 1712. ZS, (B) P| One Car |Multiple Cars | Row Total | | SPOnICACIeNs |ae St en ey NSIS ports Cates |Seeea | Stae ee eee Column Total a 70 Pr |Sports Car M Multiple Cars | Pe |Sports Car |Multiple Cars |Pr |Multiple Cars | (15%)(70%) = 10.5%. The completed Box Diagram 1s: pS One Car |Multiple Cars |Row Total | Sports Car 9.5 No Sports Car | 20.5 Pr(one car ~ no sports car) = 20.5% 24. (B) The Auto/Homeowner Box Diagram ts: Thus, 35% have auto insurance only, 50% both. Then Pr(renew)= (.50) + auto only (.40) renew given insure only auto have homeowners only and 15% ~=+(.35)-(.60)+(.15)-(.80) = 53%. (D) have PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 49 25% Begin with the Emergency/Operating room Box Diagram. Pr| E’\O') =1-Pr[ EVO] =1-85% =15%. Column Total iso, by independence, 13% =" 25%7a) => a = 00%. The complete Box Diagram is: Row Total 40 ee Pr| O|=40% 26. (D) 60 (D) A tree diagram would be nice, but would have infinitely many branches. The key is to figure out how many ways there can be a total of seven accidents in two weeks. There could be 0 accidents the first week and 7 accidents the second week. Or there could be | accident the first week and 6 accidents the second week, and so forth. Let (a,b) be the pair representing a accidents the first week and 4 accidents the second week. Pr(exactly 7 accidents) - Pr[(0,7) ]+ Pr[(1,6) ]+---+ Pr[(7,0) | = ie gi ° al 28 a I : I cay) yi prannerd ee ie prob of ; 7 accidents in Or i 2 o irst week second week Sealy feeet 28 . ie 2! = eel 99 = Saran 64 °F: “ah a oe = ee "y —— ’ et : ge < ee = : : ee Pe (EE ee ae 7 : = ’ i. tr » - ~ 2 = -, = af CHAPTER 3: DISCRETE RANDOM VARIABLES Section 3.1 Discrete Random Variables 3-1 Y denotes the number of starters among the three players that get off the bus. Par) LO phe se eer 12C3 220 Z denotes the integer selected. “At random” means each integer has an equal chance (probability) of being selected. Equivalently, is>)' ies) Pr(Z =i) = 5 fora 6 1 0} 18 38 (b) The first two throws must miss, and the third throw must hit. (a) PriS$ =1)= ae Ps =.) -(#) (43) y (c) p(s 5—] =s)=( 22] (&) [Oty = l,2, 50°". > (a) This is exactly the same probability distribution as Exercise 3-3. 18 (b) For example, shoot a 3-pointer until you make one, where p= —— 1s the 38 probability of success (making a 3-point shot). Di 52 @ SOLUTIONS TO EXERCISES - CHAPTER THREE Section 3.2 Cumulative Probability Distribution 3-5 Value Of The Random Variable | Probability |Cumulative Probability 10/36 6/36 0 2 4 6 8 10 Sum of Dice Cumulative Distribution Function Sum of Dice 12 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Bat a) Pr = hyenG, (0.3) *-(0.7)* for #=0, 1.2.3, 4: cpr |o0ost |0.0837 [03483 |0.7599 |1.0000| 3-8 Let H be the number of heads. = Pri) Pr[H >1] = 1—Pr[H =0] = 1-— = - * Then, Pr sr ST IRR C3 | t AY SAS 7 it) ole is F(t)=Pr(T <n) 3-9 0.75 —0.40 =35% of the homes cost between 4 and % million dollars. Probability Cumulative 0 200 400 600 800 1000 House Prices (in 1,000s) Section 3.3 Measures of Central Tendency 3-10 E[W]=0- 5C2 +]- 30 +2: LS 11 C2 55 55 = 60 . 55 This is a little more than 1, which is reasonable because the female party-goers outnumber the males by 6 to 5. ii 28 Bee ED Pere te as a 55 # 53 54 SOLUTIONS TO EXERCISES - CHAPTER THREE Satie I OMe 117,181a4,724 Toe — ; ve Seal e ee Iv 18 SOc aah aST aeaes VIR Verhsins ba“ IRROLIRYION 50,361,822 Oo = in en ; Bea 3-12 Since the probabilities must sum to one, © =.04. E[X]] = 32-(.16)+39-(.04) + 45-(.42) +57-(.10) + 62-(.28) = 48.64. 3-13 We need to determine the discrete probabilities from the cumulative distribution. hice Xie [eae] <0 bol aaa Seale Ge Perey Pe ptrfasPas 18|05 E[X] = —2-(.12)+0-(.11)+1-(.25)+3-(.28)+5-(.18)+6-(.06) = 2.11. = =<© <43 o) =} eae II iS= Za = EM | = 5 1(.2)+ 2(x-—.2)+3(.8-—x) + 4(.2) sey PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT ay F kas T 5 4 3 2 OQist O0000 ree 15 7300 5 lela 15 738 2rye l pit 5 = 220, 3-17 (C) (a) A starting salary of 45,500 corresponds to the third data point. 3 = 60” percentile. 4+] (b) Median = 43,412 + 45,500 = 44° 456. 2 (c) Find the 66.67" percentile. =f = in35. Then wt | X3.33 = 45,500 + (53, 750 — 45,500) = 48,250 . Top third starts at 48,250. (d) The 60" percentile is 45,500. The so" percentile is 53,760. 49,000 — 45,500 ~ 47 4%. 53,750 — 45,500 A starting salary of 49,000 corresponds to the 60 + (.42)-(80—60) percentile. 3-18 (a) 25'"to 75" percentile. (b) Linear interpolation using data points x, and x7. Namely, 15 — Xx). 25 —~ xy +—( this will be negative = 68.48" # 55 56 @ SOLUTIONS TO EXERCISES - CHAPTER THREE 3-19 We begin by deconstructing the cumulative distribution function F(x) to find POS Pegabe |PRA 1)€9°6 By OG See Fee pees oh FE | Re (a) ELX] = 4-(.2)+6-(.2)+6.5:(.4)+7-(.1D+7.5-(.1) = 6.05. (b) The median is Q) =6.5. (c) The mode is 6.5. (d) The midrange is “min =*mx. aS S575: (e) The third quartile is QO; =6.5. 3-20 (a) See solution to Exercise 3-5. l 2 l = 2-——-4+3-—+---4+12:— = 7. b)-E eet a Bee eae (c) Median, mode, midrange are all seven. random variable. (d) 3-21 O; =5 and This is an example of a symmetric O; =), Let N be the number of distinct pairs of the form (a,b) , where a and 5 were dance partners at some point. Then the mean number of distinct dance partners for a boy is ine ree ne ie and the mean number of dance partners for a girl is m implies mplies thatthat ee Pe 3.5=——. n This ie ———=— —— =—. ey In other words, the ratio m:n 2°| ofgirls to boys is a2 ee eT BR po Section 3.4 3-22 Measures of Dispersion _ Lacrosse player heights (in cm) 1n ascending order: (aj (¢) 1S 3erl64 51627-1624, 168, eed 181, Ess —— — —— ene) minimum QO, median=Q) QO; maximum aS mode Mean height is (153+ 154+162+162+168+181+183)/7 Midrange is 183 +153 9 = 166.14cm. = 168 “a (d) The first 162 corresponds to the 37.5" percentile. The second 162 corresponds to the 50™ percentile. Linear interpolation verifies that 162 will correspond to all percentiles in between the 37.5" and 50" percentile. (ec) Range = maximum - minimum IOR = 0O3— QO) = "270m. = x7 —x, = 183cem—153cm = 30cm. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 57 5-23 1 3 3 1 12 (a) E[G] E[G] = 0-—+1-—+2-=—4+3-—= SaRehG Ge = age— | =: (7 -— EIG? [G*] 3 3 {\ Sie oy ate at iReese gt? Ye gt 3 (b) VariG)=3=1,5" =15, 3 = 0.75. The standard deviation for the number of girls in a 3 child family is og = y.75. 3-24 We already found (Exercise 3-20) that the mean for the sum is E[S]= 1s =7. Why compute it again? l 2 oe} ES Ps ee Seibaaeh ee nl36° gre [ee 36 254 VartS] = E[S?|-(E[S])" = <2 -49 ss2 E583 3-25 OF —— min alae) — —O7j » Be me median een ong Oe a O03 nn“ max ae mode 1 2 3 1 l as 1 E[Classes] == 0 0-—+]-—+2-— The Tt ee +3. or a: “it 8-— Fr +24. 7 = 2 Midrange = at 56 1 = Ge Inter-quartile range = 8-1 = 7 E|Classes“]earl = 0°-——- +1 CoinaeGyeda” ue1] eal 11 2,2 1] es ie Var| Classes |= 42.45. 3-26 §=9E[X] = 32-(.16)+39-(.04) +45 -(.42)+57-(.10)+62-(.28) = 48.64. E[X2] = 322 -(.16)+392 -(.04) +45? -(.42)+57? -(.10) + 627 -(.28) = 2476.4. Var[ X]=110.55. 3-27 +23-(.01) +22:(.05)9) 9-(.2 +20-(.38) +21-(.09) (a) E[X] = 18-(.18)+1 = 19.57. + 222 -(.05) +23" -(.0D. +21? -(.09) E[X?] = 187 -(.18) +19 (29) +202 - (.38) (b) Var[X]=1.2051. 58 @ SOLUTIONS TO EXERCISES - CHAPTER THREE 3-28 .(a)(-b).- 48 ,59, 69, 69, 78,78, SS 84, 87, 93, Se min — 93,;..96,°-102 —> ad median 10 _ 784+84 eae trae 5) as =169" percentile E({Exam Score] = 795. There is no mode (3 values occur twice). Midrange = ie — a (c) 80 percentile corresponds to the i = .80(12+1) = 10.4” data point. X10.4 = 93+.4(96-93) = 94.2. (d) 3-29 Oxam Score = 15.429. E{Claim Size] = 20-(.15)+30-(.10) +40-(.05) +50-(.20) + 60-(.10) +70-(.10)+80-(.30) = 55. E[{Claim Size2] = 20? -(.15) +302 -(.10) +402 - (.05) +50? - (.20) + 60? -(.10) +702 -(.10) +802 -(.30) = 3500. Thus, oy =21.8. Pr[55-21.8< Claim Size <55+21.8| = Pri 33.2 < Claim Size < 76.8 | Pr| 40 < Claim Size <70|=.05+.20+.10+.10 = A5%, 3-30 (A) The median is 50, so Pr(28.2 <claim size < 71.8) =55%. The mode is 80, the percentage of claims between 58.2 and 101.8 is 50%. BS a) zg= AAI 1.5625, (b) 10° =100—lh75(16) 3-32 = 72. Ryan’s apple is z ed = eee 2.848 een eee ar} Jen’s orange is z= Fag ge Both have impressive fruit, Jen’s orange being a little (relatively) larger. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 59 Number of standard deviations from the mean gee ear deviations of the mean 3-34 rede elo oo 7s eee) (oo984 | 9999 For the apple, the coefficient of variation is 100° 2.865 9,= 250%. 11.19 For the orange, the coefficient of variation is Seren =33.)%. 3-35 {sum of two fair dice] =7 and Ogum =2.415. The coefficient of variation for the sum is Wee, =34 5%, Section 3.5 Conditional Expectation and Variance 3-36 Probability Cubs win in 3 games (.59)3 Cubs win in 4 games 3-(.59)3(.41) Cubs win in 5 games Cardinals win in 5 aACye(59)2( 41)" 6 -(.59)2(.41)3 Cardinals win in 4 3-(.59)(.41)? Cardinals win in 3 (a) ELX]=31(.59) +((41)5} +44ee 3(.59)3((41) | te +3(.59)(41)} aa eee probability that 3 games are played probability that 4 games are played " 5|6(.59)5(.41) + 6(.59)2(.41)3 mo ability that 5 games are played 3(.27430) + 4(.35109) + 5(.37461)= 4.1 EL X?] = 3° (.27430) + 47(.35109) + 5? (.37461) =17.45139. Oxy =.799. 60 SOLUTIONS TO EXERCISES - CHAPTER THREE (b) If the St. Louis Cardinals win game 1, then we have the following (conditional) probability distribution: Probability | Cubs win in 4 games | (,59)3 US379 Cubs win in 5 games | 3.(.59)3(.41) | .252616 Cardinals win in 5 3-(.59)2(.41)? |.175547 Cardinals win in 4 2:(.59)\(.41)2_ Cardinals win in 3 (.41)? | .198358 1681 E\X |Cardinals win game 1] =3 {1681} : +4 {403737} HO E[ X? |Cardinals win +5 {428163} SSS probability that 3 games are played probability that 4 games are played =4.26 See probability that 5 games are played game 1]= 37 -{.1681}if + 42 - {.403737} ! J + 52 - {.428163} s =. LOO ae oxG ardinals win game 1 — * EAs 3-37 Let Hbe the event the husband survives and let W be the event the wife survives. From the given information we have: obs es W | .960 | .025 |andthe completed|W w' | .O10 ae oe ee Box Diagram is: The conditional distribution given H is in the first column. Let X be the excess of premiums over claims. The conditional distribution of Xgiven His: Event xX Probability How 1000 OF H AW’ | 1000-10,000 | .01/.97 1.00 E|X|H] [X|A] 0.01 = ((1000-10,000) = area 7 SS excess if wife dies conditional probability wife dies 0.96 LOOO Wear pais excess = "0.97 if wife lives 690.91 ; eocn (5 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 3-38 @ 61 The 11 relevant outcomes from the 6 by 6 dice table are: (4,1) (4,2) (1,4) (2,4) ©G,4) (4,3) (44) (5,4) (64) (4,5) (4,6) 2 2 E{Sumlat least one 4] = 5 austere: 3-39 Let A be the amount of the loss. 06 03 E[A|A>0] Tat 00 70 A|\A == 500500-—+1000-—— a2 l eae 2 2 82 AE ee 008 001 001 siaide are+ 100,000 -——— 70 +10, 000 - —— + 50,000 00- -—— = 2900. 3-40 (a) Let x =Pr(D=5000). 600 = 0-(.82) + 1000(.10) + 5,000(x) + 10, 000(.08 — x) ganado 2 Eee | 225 ESTO 0G (b) E[D?] = 02-(.82) +1,0002 -(.10) + 5,0002 - (.06) + 10,0002 - (.02) = 3,600,000. op =/3,600,000 — 6007 =1800. 10 06 02 (c) E[D|D > 0] = "1000s (Wes, 000-[16) +10, 000: [ 7 BO 3833,33. are 10 06 a2. (2 ( dy) EID* D> 0] = 1,000 (42 )+5,0007 (}- 10,0002 Mei = 20,000,000. OFp\p>=2981.42. 62 SOLUTIONS TO EXERCISES - CHAPTER THREE Section 3.6 Jointly Distributed Random Variables (Round 1) 3-4] Probability Probabilit Ets ade | fe | | (b) b) E[Y] EY) = Oseot a ee Var[y |= el ; oael ey eel = 02 ee Beer | : Cola ee tants ly 2 (ec) FIV+Y] = 0-41 +2 Ae 43 eA er hs, CNotethis equals say 1 E1Y) haar ECMEY)? | 0 en ; 4 eee pea St VarlX +Y | = 3.5—15* =:1.25, oR eee (d) 1.25#.25+.50. 3-42 (a) We need to verify that PrL¥ =xOY=y) x =1,2,:--,6 and y=1,2,:--,6. = PriX =x)-PrY=y) Yes, they are independent. (b) By independence, Var|X + Y] = Var|X]+Var[Y] calculate Var_X] = 1? eee 6 for eee 6 = 5—. = 2.916. Of course we need to bg SOby bg ht Se bl X+Y Probabilit 4/36 12/36 13/36 6/36 E[X +Y] RRSIES > ><S — Se Seby — be (c) ~ | l l -—+2-—=]. ~ 4 (a) ELX] ={) -_—+ 4 San E = — ~ Probabilit i T — . =FPTrarTryTr — ileell vr Stes a SPTarT —— ee — Il SN 3-43 — Tare ST PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT i fot° 2 3 ra {= # 63 64 SOLUTIONS TO EXERCISES - CHAPTER THREE (d) Yes, the joint distribution table below shows thatX and Y are independent. The joint distribution for X and Y 1s given by: Pie) 7) at 0 1 2 1 pate) 3-44 (a) 1/8 e+ i) e+ e rt bo WI} l 3/ 3/ 1/ (b) EX] = 0-5(oat 41-5 (c) EX, 4+Xo Xs) | AU ALIS es = |bo to Var Xt Xo tks). “Su 5 Varlaeelt rartxe lePareye) = is ee + 5 4 by independence Or we could use the probability distribution for XY, + X> + X3. Var[X;+X2+X3] = Bl(Xy+X2+X3)?]-(E(X1+X2+X3)) = 3-1,5? = 0.75. + PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 65 3-45 Shoe Size 25 a ee ieee i = = 2 (a) E[Height] = 68-(.25+.05)+70-(.32) + 73-(.38) = 70.54. E{Shoe Size] = 8.5-(.60)+12-(.40) = 9.9. (b) E[Shoe Size Height = 73]= 8.5. =a) (3) STs (c) E[Shoe Size|Height =73]= 8.5- 3+£12. (: = = 9.083. = Shp ssOe (3s (¥) = 84.2083. 30 30 E[Shoe Size? Var[Shoe Size|Height = 73] = 1.701. Conditional coefficient of variation is 100- NS 083 eur Section 3.7 The Probability Generating Function 346 (a) ns) = (5+ = RtostSst ais) 3 ieee g innar =a Vee B-47 ~ es? g (1+)? is A's) = 2a} 8 243-(3| =k (a) hytsy=5-C7) (3475) EB lahe (1) =3 5: (b) A&(s) = 20-(.7)? -(.3+.7s)3 hp(1)=9.8 Var[B] = hg(l)+hp(l)—(hg()) = 9.8+3.5-3.5° (c) hp(s) = (3+.78)> = (.3)° = 1.05. +5-(.3)4 +675) +10-¢.3)? (75)? +10-(3)2 «(.7s)? +5-(.3)' -.7s)4 +€7sy° 60 40 1.00 66 ® SOLUTIONS TO EXERCISES - CHAPTER THREE So the probability distribution for the random variable B is: LB _| Probability | a a ee Note: This is the probability generating function for a binomial random variable (wait for Section 4.2) with probability of success p=.7 and number of trials Peale), 3-48 _ (-.21)-(8)-G8))-(2) (a) hg (s) = aaa] 77 aaa _- 8 Geant E[G] oe. 3 eee ee ORIG Ye errs 559 CVA aoe ars a Note: This is the probability generating function for a geometric random variable (wait for Section 4.3) with probability of success p=.8. Section 3.8 Im (a) Chapter 3 Sample Examination Pr(S <s) s, sum of two fair die 2? 2 3 4 5 6 7 8 9 10 1/2 2/25 3/25 4/25 5/15 4/25 3/25 2/25 1/25 (n (b) E[S]=6. (c) E[S?2]=40, o2 = 40-62 = 4, os = ,/40-6? = 2. 2 3/25 625 10/25 15/25 19/25 22/25 24/25 25 —- 4] _| PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 67 1 2 eee, 3 4 5 Am 98 ee eae ee re! (Ou alc meomme TOR 08"tySa 19 ~ jon CM eee eon Sie1 a 2 a A 3 ee lhl (Oe elon oe 546 eee 74 19 a4 Be SrIRSh ao Fe5 ig 4 le ot 10 Var[S |S <8] = 28.74-5.16? = 2.13, ofS|S<8]=2.13 =1.46 1 with probability + X =0 a heads = 1 De X =1 with probability heads tails @2 -; heads tails rf) 2 tails aii D (ate nO eo OPO S078). MIN 17, fe 8, 5 O50 0,9509, 95.9; Dy 9, 10 SS median FT mode ees max =(7+7)/2 pyre as Midrange = a =e (c) 80" percentile corresponds to the i = (80%)(24+1) = 20” data point. (d) oy =25938:. (c) ”) CV cheeseburgers = Z= Z 11-6.5 Ahi 2.598 / = 39.97%. x29 =9. 68 ® SOLUTIONS TO EXERCISES - CHAPTER THREE 4. (a) ELX] = 0-(.74)+50-(.12)+100-(.09) +200-(.05) = 25. (6) oF =3200—= 257 = 2575, oy =V5200—5 650.74. (c) Since the expected value after a year is 25 the prudent investor should not buy the stock at 30. (d) E[10-X¥ —300]=10- ELX]-300 =-$50. (ec) Var{10-X-300] = 102-Var[X] = 100-2,575 = 257,500 25 49.3%, (f) CVy = ae — ys = ae 0 Geometric series. Expected Claim f=0 1000 1000(.95) 1000(.95)? ae f=2 f= vee 1000(.95)78 f=29 1000(.95)?? t=30 The expected claim for month ¢ is 1000(.95)‘" and, Expected total Payments = 1000 + 1000(.95) + 1000(.95)? +--- +1000(.95)?? as e\30 awlegi 15,707. 7 000) 1—(.95) S|S is prime fe (a) E[S|S is prime] = 6.13 (b) 0 =2.5027. cS and Var{S|S is prime] = 6.26. (D) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 69 30 hit 30 miss hit ; hit 30 70 miss — 4 > hit 70 Y~ miss HS) : DS miss hit 5 30 A miss TY) .70 miss 5 hit Probability Chee = 39375 C3) )Cs>) = (25)07)7 (INCITS \25) = 41125 (3200) = (3)CC25) + (.7)(.25)(.3) = 168 (.3)? =.027 (a) Efhits] = 0-(.39375)+1-(.41125) +2-(.168)+3-(.027) = 0.82825. E{hits?] = 02 -(.39375) +12-(.41125) +2? -(.168) +32 -(.027) = 1.32625. (Dy Ope (a) = v1.32625—.82825" = 0.8. 22-(.05) + 23-(.01) = 19.47. 20-(.41) + 21-(.05) + ELX] = 18-(.23)+19-(.25)+ Median is 20. Modal class is 20. Midrange is “ee 70 ® SOLUTIONS TO EXERCISES - CHAPTER THREE - (b) Range is 23-18=5 years. E[X2] = 182 -(.23) +192 -(.25) +202 -((41) + 21? °(.05) + 22? -(.05) +237 -(.01) = 380.31. Var| X|=1:2291, oy =1.1086. Al .05 .05 01 (c) E[X |X >20] = 20-[) +2198) +22 [-B)+23-(-2) = 20.346. (a) ELX] = .10(0) +.18(1) +.12(2) +.37(5) +.23(8) = 4.11. (b) ELX2] = .10(02) +.18(12) +.12(22) +.37(52) +.23(82) = 24.63. oy = ¥24.63-4.1122 = 2.78. (c) Q3=5. Let p denote the probability that the damage is $2,500. Solve 700 = E[ damage] =0-(.73) + 2500- p + 5000 - (.27 — p). oN (a) E[damage |damage > 0] = 2500-52 + 5000-2 = 292.59. 2 .26 (b) E[damage? damage > 0] =2 500° an © damageldamage>0 a (a) fy te 5000- - oT Ade ko: = ELX] = 12.09) + 2(.32)+3(.42)+4(.13)+5(.04) = 2.71. The 50" percentile, median, is Y = 3. E[X?] = 17(.09) + 27(.32) + 37(.42) + 47(.13)+57(.04) Gy =8.23-2.717 (b) £[Payment = 8.23. =.941, at least 3 TVs].: = $75:2-2= alee e025 oe GTeS ines iy 59 eS) Nn PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 71 LZ (aye elses tel 4el42 15.91 =< eS) a median ae mode =(14+14)/2 O71 = X3,25 =13. 5153.15 9868 18 Sa Hntebs O; = X075 — 4ies Midrange is 15. jis RY Wee (b) JOR =15—13 =2. 1) __ 10-14.416 — | 2.845. eea0ee Not an outlier. E{merit money] = 0-(.09) +500-(.61) + 1000-(.15) +2500 -(.09) +3750-(.05) +5000-(.01) = $917.50. E|(merit money)? |= 02 (.09) + 5002 -(.61) +1000? - (.15) + 2500? - (.09) + 37502 -(.05) +5000? -(.01) = 1818125. O merit money = $988.09. Let p denote the probability of winning $10. Let W be the random value of your winnings. (CS Ge ae E(W]-= 0-(.90—p)+10-(p)+100-(.10) Var[W] = 100p +1000—(10p+10)? = 10p +10. = 73(10p +10). Solving for p we have p=.2. (a) E[W]=12 and Var[W]=876. 9) (b) EW |W > 0] - 10-4100: = $40. (Cc) CVy = 12 876 = 40.544%. Ty 72 155 SOLUTIONS TO EXERCISES - CHAPTER THREE SEIR SSE A) = SCS 202 2) ane Gi eo, (48) You might earn a raise if you informed your boss that hp(s) = (.8+.2s)3 83 +3-.87 -.25+3-.8-(.25)? +(.2s) = .512+.384s +.09657 + .008s7 So the discrete random variable R has probability distribution eS Eee CC 16. eee eo ee ee Let x be the common difference between consecutive probabilities, po, p}.---, Ds From (1) po is the largest, with the remainder of the probabilities arithmetically decreasing in size. From (i) and (11) together we have, Mi PossPe CPE Pe: Then, Py =(ParX; and imeenetalsy First, ase 7, = Dia ps = Pp = Pix Dy iis Po s,s + (Po = (pox)— x — po OL ee —X) +++: + (po P| Second, from (iii), 2 = potp 2x, —5x) = 6 po —15x. Ds = pot (po-x) = 2po-X. Solve simultaneously the two equations, 6p9 —15x = 1 and 2pp —x 5 Po aT : = 3 to get, l and x svat Then, Pr| #Claims > 3 | = pat ps =(Po —4x) +(po —4y) = nee fet (C) CHAPTER 4: SOME DISCRETE DISTRIBUTIONS Section 4.1 The Discrete Uniform Distribution 4-] (b) To prove i = 14+24+3+--4n n(n+l) 5 = i=] by mathematical induction, we need to 1 I. Verify that the statement is true for n =1. Me = [= =vk i=l II. Assume that the statement is true for n =k. ew tiie: eyseal III. Show that the statement is true for [+2+34ek + (ke = 9) n=k +1. (en = oe. (c) The statement is clearly true for the case n=1. Assume that the statement is true for Pepe ay n=k. k(k+1)(2k+1) 6 iy aay Then, + (k+l)? (k+1)| k(2k-+1) + 6(k+1) | 6 (d) showing the statement is true for 6 n=k +1. The statement is clearly true for N =0. ; ataxt+ax* +ace+---+ax Rote ay cee 6 Assume true for k+l ad—x*™") l-x + l-x a(l—x***) l-x showing the statement is true for ee 1104 i = Il + s|-Sei — + s|-— Saw — We) s|- a + Se to a ee) + SS) ars + N=k +1. oo “fe = S se! Tal — ——— \| n-(nti) wn 9 N =k. K+ ax*™(1-x) Shue) 9 Then, 74 4-3 SOLUTIONS TO EXERCISES - CHAPTER 4 (b) The total accumulation of rice is 1+2+27+---+2° = 2° —1 (geometric series), which is a bunch more than one billion grains of rice. . 4-4 2 n?(n+1)* We make use of the algebraic power sum formula | es =—_———.. | k=l es aes ~ 1 EV Xe | =o) Uke Ea 2 n n?(n+l)* ee = n(n+1)? See n 4 Then, ; One may expand E|(X-p)3| = E|X3-3uxX? +3n2X -13| (APG RVI BC APREIILIA eye II = E[X?)-3uE[X7]+3u3 -pwe = ELX?]-3MELX2]4+2° This can be evaluated using the formulas, errr vite a FLY?) a ee and 3 (AO Ga i n(n+l); After a little algebra the result is seen to be zero. Or, one may simply observe that the values of X — w (and hence (Y—y)*) that are negative exactly cancel out the values of X — wu (and hence (Y¥—y)*) that are positive. Thus, the expected value, by this symmetry, must equal zero. Discrete uniform distribution on {1,2,3,4,5,6'. (b) EX] = = 3.5, (c) There are 6 data points, the median is (d) oy = 67 -1 12 3+4 =3 2 Lo) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 4-7 (a) Pr[(X =69) U(X =70)U---U(X =85)] = a (b) E[X]= 100 +1 2 = ehh, (c) There is no mode. All 100 outcomes are equally likely. ip eae 100? -1 4-8 (a) Pr[Y 25] =<. Let X =¥+1. Then X is discrete uniform on (Oop (b) E[Y]=ELX -1)=ELX]- £11] =73h 124 (c) Chy 4-9 2 (7) = — = 64.55%. (a) Pr(Z>5] = 1-+ = z. (yFIZ pe BI2 eX] = 2° RIX Ie 224) P10) (c) Var{Z] = Var[2-X] = 2? -Var[X] = + 5] = 26.6. oz =5.164. 4-10 +45] = 5E[X]+45 = 5(3)+45 = 60. - xk b>) = Pr|A>= leer | = Pian Il. EY] = Ela-X +6] = a-E[X]+b iy Vary |= Vvarfa-X +b) = a* -Var[| X]. —h | 9) # 75 76 SOLUTIONS TO EXERCISES - CHAPTER 4 Al? a) PrLy is prime] ==, (b) PrLX >11]=30. Section 4.2 Bernoulli Trials and the Binomial Distribution 4-13 E[M] = 0-(.1184)+1-(.3681) +2: (.3816)+3-(.1319) = 1.527. E[M2] = 02-(.1184)+1? -(.3681) +22 -(.3816) +32 -(.1319) = 3.0816. o3, = Var[M] = E[M?]-(E[M]) = .749757. Ay WewePrMe=kh) ="35 Ce 8) 02) fork = 0,2 roi ELM |:=A2"(8) =.9.6 Var[M] = 12-(.8)(.2) = 1.92. Pr[M <10 = 1—Pr[M =12]—Pr[M =11]—Pr[M =10] = 1-.0687 —.2061—.2835 =.4417. toa 10)" 19, = a0 Cie G2) sy ee Pal Ye> 2 li = op Cond) $8)" ona 2) G8) E\Y| = 20-(2)i=4 Var\Yo) 92022). C8) = O88: 4-16 Pr[ Yankees win 2 or 3 games] = 3C> -(.6)? -(.4)'+(.6)? = 64.8%. 4-17 Sy ueaa aera Pr[2 < X]=1-Pr[X =0]-Pr[X =1] =1-3C,-(.9)!-(.1)? — 3p -(.9)° -(.1)? = 97.2%. E[X]=3-(.9)=2.7 and Var[X]=3:-(.9):(.1) =.27. PROBABILITY AND STATISTICS WITH APPLICATIONS: 4-19 A PROBLEM SOLVING TEXT ® 77 A twenty question multiple-choice particle physics examination written in the language of Lower Lausatian, each question with four answer choices. PrX=2] = 20C2-(.25) -(.75) TEP ET Ny 00-025) = 5 Pr[X <p] = Pr[X <5] = Pr[X =0]+Pr[X =1]+---+ Pr[X =4] = .4148 OF 20825015 Oo; =3.75 4-20 Draw a tree diagram, where the first stage denotes from whence the shipment came (X or X’). The second stage denotes the number N of defective vials (probabilities from the binomial distribution), where we are given that N =1. Pr[X]=.2 and Pr[X']=.8. Pr|N=1|X | = 30C-.1!-.929 and Pr[N=1|X"] = 30C;-.02)-.98. Using Bayes’ Formula, 20+ (30 C, -.1! -.9??) PrLX vee s 20:+Go Gyeul! -.92?) +80: 4-21 = Pri[no shows =0 4-22 = 0957 (A) or 1] = 32Co-(.1)°-(.9)?7 + 2) -C.)'-C9)?! = .1564 Let N be the number of hurricanes. Pry 02 4-23 - 66 G-.02'=.982?) |= (.95)7° + a9C\ -(.05)! -(.95)!? + a9 C2 G05)" (.95)'® 5)-np=n- p-q=Var[X], (a) (.25)-E[X]=(.2 7 so q=.25 and p=.75. ae (b) PrLX =7]=10C7-(.75) -(.25) =.2503. A24 -Pr(X >3.25) = 5Cy*(.65)* -(.35)' + 5Cs £(65) 7(bo) ="9245. cette 8.703 (B) 78 @ SOLUTIONS TO EXERCISES - CHAPTER 4 4-25 hy(s) = (qt+p-s)". hy(s) = n-p-(qt+p-s)"". E[Y] = hy(l) = np. hy(s) = n-(n-l)- p?-(qt+p-s)”*. E[Y(Y-1)] = hy(l) = n-(n-1)-p? Var[Y] lI= h'(y+h'ay—(h')) n:(n—1)pp? +n-p—(n-p)? = n?+p?—n-p? +n: p-—n?- p* n-p—np> =n-p-(l-p) = n- p-g. 4-26 Let N be the random variable that denotes the number of employees that achieve high performance. NV has binomial distribution with parameters n= 20 and p=2%. The first few rows of the probability distribution are E24 Biles ad 7 Pe as Prv=n) |6076 |2725 |0528 |. | Fo) [0675 [9401 |9929 |. 66.76% of the time, no employee would earn high performance, and C could be set at anything. Looking at the cumulative probabilities, 2 or fewer employees will earn high performance 99.29% of the time, so the probability of exceeding two awards for high performance is less than 1%. Thus, we can set C=120/2=60. (D) 4-27 The only time a penalty occurs is if all 21 ticket holders show up. Efrevenue] = 21-50 —-100- (.98)*! —— = 1050-—65.43 = 984.57. (EB) — revenue from sale of 21 tickets probability all 21 people show Section 4.3 The Geometric Distribution 4-28 Pr[D <2] = (.8)+(.2\.8) = 96%. E\D]=-2=(25. 4-29 E{D*]= Var(D]+(E[D]) See +{- = 375. The probability of success (not being rejected) is p=.10. Pri R< 2) = GLC 9)GL) = 19%: E[R] == | 9. OR = Aas, 15 = We are assuming that the geometric distribution applies, so in particular (1) your advances in a pub are independent and (2) your probability of success remains constant at 10%. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 4-30 Fi Macon failures] = 4 ais a-31 Let px = Pr| X =n] = q"p. Then, n Pr| X >n| = Pat Pal t= q’ptq™ pt -= aay l-q # 79 n = gor P = Gs Pr} (X¥ >a+b)n(X = Pr(X 2a+}|X >a) = Prix 2a+b)O(X2a)] Pr(X >a) P[X>a+b] = gt? Sg = Pr[X Da] cage iat cate emesis eeepe D 4 Pr{(X <1.5)] = Pr[X =0]+PrLX =1] = (.4)+(.4)(.6) = 64%. 4-33 No. But you better have a reason. (As in the previous exercise, the variance uniquely determines the value of p.) peg 7—50 p implies g=(2)=.5848. poet Tey \ gp eee ee LX] =1.41. stor y—1,2,3,---. ee el p p p Note: Var[Y]=VarLX +1] =Var[X]=—4 Pp 4-37 >° This strategy will cause you to lose your bankroll if you have 4 consecutive failures. 4 Pr[lose bankroll] = (2 INCH 80 SOLUTIONS TO EXERCISES - CHAPTER 4 4-38 (a) For n=0,]1,2,---,48, (b) Distribution of X =e S 2 i) = i 10 20 30 40 Cards Drawn Before (c) E[X]=9.6 and Var[X]=160-9.22 =75.36. (d) If Gis the geometric distribution with p= | Nn , then £[G]= Aus (Nnbho Var[G]=156. 4-39 This is the alternate definition of the geometric distribution. E[X]= = iS P implies p=.08. Pr[ X =6]=(.92)5(.08) =.053. (B) 4000 3000 2000 A | 1000 E[ Benefit] = 4000-.4 + 3000-.6-.4+ 2000-.67 -.4+1000-.6% -.4 = 2694.4 (E) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT ® 81 4-41 Here Xis the trial of the first success, where success means rolling a 5. This is again the alternative definition of the geometric distribution with p=1/6. e[x]= Set (8-1 as co Thus, n-l Pp Now, let Z=X|Y=2. Given Y =2, we know that the first roll was not a 6 and the second roll was a 6 with certainty. Thus, either Pee Z=1 or Z >3. Weeetsit2 3.4015) = 1/5, Pr| Z = 2| =(); 4 Pitz 3 Tei 3 > a I - .-( 4 \f 1 ie (+4). Similarly, TO fistrol oer rob 31droll is not 5 rle~4l (56) rz-s}-(s)es): is 5 acs =gs203(3)e}-+(s) aa) -S(SPO+Oe- Od and so forth. Now, | of we n|— i ~S—Seee This sum is by definition the expected value of X +2 Ty 42) 5 5 1 (4 [eLx]+2| = =+(= = —+ 2 (6 +2) 82 @ SOLUTIONS TO EXERCISES - CHAPTER 4 Section 4.4 4-42 The Negative Binomial Distribution | Repeated differentiation of Aj (s)= p”(l-qgs)~” shows that AO(s) = (rt+k-l(rtk-2):-r- p" -q* (-qs)-"*) for any integer k 20. Then, A“) (0) = (rtk-l)(r+k-2)---r- p” -q*, and the coefficient of s* in the power series expansion of hy(s)= p’(1-qs)" is WOO) (reRaIKP+k—2 e r, ge es 4-43 hy (s)= p’(-r\(-g)-qs)"", PRI = Pr] M=k|. so ELM] =hy (I) = p’ (r\(q)-9)"* ao hiy(s) = p’ (r\(q7\(r+)(1-gs)"7, so Ahy (1) = ae and Var(M] = hhy (1) +h) (Av) _— OOH?) eg Pe? p és rq? +rq* ' 444 P De +rgp—r*q" ts p = =. rq oe (a) 4, (35)2905)4(35) = 1087 (6) Cy 635) 065)2 Fa Cp C35)" 065)! 2 E35)? = 2352 (\ EI MV = yt iets ST and Pert] Pp (d AAS) rq? +rqp — Sie p- 22) E[Total Attempts] = £[M +3]=8.57. (ay Pr y= 8) (Ce C8040)? = 00ed: (bo) Ely] = 7-2 = 28, (c) .oy-=/140 =)1,83 45 (d) Pr(¥ <45)= he a k=0 4-46 E[X]=6=— 2) and Var[X]=12=—+ Pr(X <3) = (.5)o+4g p- implies p=.5, q=.5, and r=6. Ci 35)" 05) 399 G3 G5) eat a PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT. 4-47 Instead of px, =Pr(M =k) = +44-1C,4~p”-(l—p)* [Ye 4-48 = PrOh4-— Kk) = ey er aie Wes for Kr # 83 for k =0,1,2;---, we have Opec, Let success mean a month with no accidents. Then p= Be 4. Let E be the event there are 4 successes before 4 failures. We need the probability of E. Method I: Binomial Distribution. in the first 7 trials. Thus, Pr{E] For E to occur there must be at least 4 successes : 2, 7CEC4)* (6) k=4 nA) 6) 94 Cs (4)> C6)? 4, Ce C4)° (0) +(4) eno Method II: Negative Binomial Distribution. Let be the number of failures before the 4" success. Then £ occurs if and only if M <4. Pr[M <4] |= xyCO (44)| 1+ (4)(.6) + (10)(.6)? + (20)(.6) |29 (D) k=0 4-49 Let X be the number of hurricanes and let success mean damage occurs, with p=.4. Let M be the number of failures before the second success. Then X = M+2 problem calls for the mode of X. Mis negative binomial with, Pr{[M =k] = 244-1C, -.6*" -.42. A partial probability table is: k |Pr[ M =k] iy eeakan team: | 1920 jalan) 798 Ea wey: . The mode of M is | and so the mode of X is 3. (B) Note: It can be shown algebraically that Pr [M = k| is decreasing for k>2. Section 4.5 H-300 The Hyper-geometric Distribution “(aye Pr Dies) = 27Co + 14C5 F. 27Cy -14C4 Fi 27C2 +14C3 41Cs 41 Cs 4iCs ae 2092 and the 84 4-51 SOLUTIONS TO EXERCISES - CHAPTER4 W denotes the number of women on the seven member committee. Pr, Se ( 4-52. _ & C6 a -34C 100 C7 100 C7 an ) Cz ga. Cyee, ee C Oe Oe 100 C7 Oa 100 C7 A denotes the number of administrators on the eight person committee. the number of faculty. Note: F =8-— A. F denotes E[A]=8(.75)=6. So, in a random selection of committee members the expected number of administrators is 6. On the other hand, again assuming a random assignment to this onerous committee work, we have, Pri Bs 4) et es eae = .1047 Comment to President Uptight: What do you mean by fair? 4-54 E[X]=E[Y]=£[Z]=2. Var[X] = 5-(.4)-(.6) = 1.2, vary) gat= 5(22) 200)(53071) 49 1= 1.102,199 and (3) [$F] Var Ze 5-(1000 awe }{ on )(33] = 1.195. ) \ 1000 } \ 999 4-55 There are 10 defective modems, 6=30-20% from company A and 4=50-8% company B. The probability that there are 2 defective from 5 selected is MESES he WO 80 C5 Section 4.6 The Poisson Distribution 4-56 4-57 Pr(X =0)=.001007785 =e, so A =—In(.001007785) =6.9. E[X]=2=6.9. Worm 100Co (.03)° | ,o0C} - (.03)! Pr(Worm =k) rcae = .0476 -(.97)% |.2252 |.2275 |.1706 |.0074 = 1471 | | L Poisson ePes)" e*:(@)E Approximation 0! 1! A = 100(.03) = 3 = 0498 = 1494 .2240 |.2240 |.1680 |.0081 from PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 4-58 @ 85 Let C denote the number of cars that arrive at the Wendy’s. e 24. (24\3 0! =.091 4-59 E[Z?] = 12 = Var[Z]+ (E[Z])" = A+A/?, Solving this quadratic equation, we have 2=3 or 2 =~—4, but the parameter must be positive. 3! e Pr(Z <1) = Pr(Z=0)+Pr(Z=1) = e3 + = —~ 749. 0 AS9A 4-60 Consecutive Days of Rain| E| Insurance Payment] Insurance Payment Probabili 2000 1219 = 0-(.5488)+1000(.3293) + 2000-(.1219) = 573.1. E(Unsurance Payment)” | = 07 -(.5488) + 10002 - (.3293) + 20002 - (.1219) =9$16,893: © InsurancePayment — NV 816893—573* 4-61 =699. (B) Number of Supreme Court Vacancies, k Actual number of years from (1790-1932) with & vacancies Poisson Prediction of number of years with k vacancies 14.3. p=4755 = 88.883 The number ofperiods is 1932-1789 = 143. (87)(0) + (47)(1) +(6)(2)+ (3)(3) = 68. The total number of vacancies is The overall mean rate iS A = 68/143 =.4755 vacancies per year . / Pel (a) E[ Vacancies] Pe= 0-775 87 a Var[ Vacancies] = .4592. 47 ae 126 ae ahve tO bes lea ATER 4755 and 86 4-62 SOLUTIONS TO EXERCISES - CHAPTER 4 (amet bom Hitetper sector=0 576 ee576 a Var{bomb hits per sector] = 1.797 —(.92882) =.9342. Bomb hits on London, k ee see Number of sectors with k bomb hits Poisson Prediction of} 576 eo© 93 35 1 97.6.) 30.2 ea ota number of sectors 0! with k bomb hits = 220.3 210.2 The number of periods is 576. The total number of hits is (229)(0) + (21 11) + (93)(2) + (35)(3) + (7)(4) = 535. The overall mean rate is A =535/576 =.9288 hits per sector 4-63 ; e22:| (a) SR ha =i} oar for f= Us2,7 (b) The mode occurs at T =2. (CG) Var T= Aw, 4-64 (a) wc = E[C]=30-2.5=75. Bale Se ies Ov (b) Pr(C =70)=—| 4-66 e-"(n)? A=n, SOoPrl A=] n! — en = (n-l)! l =9PriZz=(n—1)). PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT ® 87 4-67 Let =P i Z=n|=e" a for n=0,1,2,... The ratio of consecutive probabilities is —A4Nn ! fn = ioe ec Pr CA Eto ew(n—1)1 Uasincet enor ar integer, this ratio is ln greater than one for all n < 7, and is less than one for all n >A. Thus, the values for Py are increasing for n< A and decreasing for n >A. Therefore, the largest integer n less than 2 is the mode. 4-68 Approximate using the Poisson distribution with 2 = ae e7 02026 (.02026)° Pr(3 souvenirs) ~ =.000008. That lucky son-of-a-gun. Of course 3! Davis had a nice seat directly behind the dug-out and was a cute kid, so the requirements for a Poisson process likely do NOT apply. 4-69 Solve for 2, which is also the variance. ma Vad A> 4-70 eA? i BigSha implies A~ =4 andsoA=2 ead 2 accidents 4 weeks _ accidents week 4weeks § 4-week period © : : =o Pr[ Accidents in 4-week period <5]= e° + se + (D) Q2 -8 a 83 4.0996; E{ Accidents in 4-week period] =8. 4-7] 4-72 a . (a) E[mistakes in 300 pages] (b) peter4-151 WN 4|«\0 " eet -15/16)l \S) 0)! Pr(Cusses = 2) = 1! = an | The variance is 8 as well. mistakes eee =15 SOU pAnes Gae5-15(15)4 (Loy 4! ih = .000857 —8 7 + Q4 —8 A 88 SOLUTIONS TO EXERCISES - CHAPTER 4 4-73 (b) No, for the discrete uniform distribution all probabilities would equal 1/5. (c) Pr[(S <4]=1/9+2/9. snowstorms ea (eo)! 0! els en 1.5 TeoP a 5)" 10,000 | 10,000 ae 10,000 ee (1.5)9 ee ees 10,000 ae Note that the last column, Payment + ¥, is equal to 10,000-S. 20,000 i 30,000 40,000 Also, E(¥] = (0)(0) + (10,000) Pr[S >1] = 10,000(1—Pr[S = 0]) Thus, E{ Payment] = E[Payment + ¥]- E[¥] = 10,000. E[S]—10,000- Pr(S > 1) = 15,000-10,00011—e'?) = C. 4-75 Binomial Model, n=80 and p=.05 Poisson Model, A=A4, 95° = .0165| .0695 1978 .2004 | .1603 0733 .1954 .1954 4 e+ = 0183 1563 1 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Section 4.7 Summary Comparison of Discrete Distributions : AY ) Uniform, n=5 (Oye | Ql 4 | 0 = Prob) 4-76 (b) an 1 > 0.2 . = 2 0.2 a 3 0.2 P Binomial, i ~ n= 10, 4 0.2 5 0.2 [= = eee ead p=.3 3000 2000 - 4). 1000 0 0000 | El i) a 1 BA 3 a 4 5 = 6 — $3 — =| 9 ;— Prob 0.028 0.121 0233 0.266 0.200 0.102 0.036 0.009 0.001 6.000 0.000 ee yr te Hypergeometric 4-76 (c) 0.3000 0.2000 0 L000 OOOO Prob Ul i 1 Z 3 rn 0.2166 04165 OT 0.0793 0.0086 a 5 0.0004 @ 89 90 # SOLUTIONS TO EXERCISES -CHAPTER 4° 4-76 (d) | | ares Poisson, Mean = 3 |os t _ | |0.2000 | ; | 0.1000 - ll 1Ih » §.0090 1 i a a 5 | 4-76 (c) 9 10 Prob0.049 0.149 0224 0m 0.168i 100 00.050 9.021 0.008 9.002 8.000 - - ; Geometric, p=.25 7 0.3000 0.2000 0.19000 0.0000 - | | — ii Hla a | ra) i 4-76 (f) TET TE URGRS SIT 10 0.033 0.025 0.018 0014 or Nesaive Binaitia r= 3 and p=.5 0 3000) 0) 2000) 0 1000 0 0000 I 9 10 Prob 0.125 0.187 0.187/0156 0.117 0.082 0.054 0.035 0.022 0.013 0.008 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Random Variable [a RS irromia] Sng | pe yee 4-77 aa Pp Negative Binomial 4-78 91 (a) E{coins}]=5<16=Var{coins]. distribution. (b) 2-5 and ie joe P yy P 5Q ; This information suggests the negative binomial ple ep 16 andin= iG= wiCalculating pronapninies with a non-integer r is not straightforward, although it can be accomplished using the Gamma function, which will be introduced in Chapter 6. 4-79 Let S be the # of banana bunches purchased. (ajeels|= 72. and, Varis|=1.26 {wba =.7416. (b) Since the mean and variance are close, the Poisson model with 4 =.72 is an obvious choice. ~(c) We need to calculate the probability of at least one success in 1000 trials where p=1—Pr| S<4|=1-.99911=.00089. Then the probability of at least one success Is given by, | (.00089)°(,99911)!00 |= 5894. 92 SOLUTIONS TO EXERCISES - CHAPTER 4 4-80 The following table shows the specifics for each of the 6 discrete distributions. Fy es Seat a dd je eed Mean _|__ 5.5000, _6.0000,_ 1.2500,__4.0000|__3.0000|__ 3.0000 2.44 : pa area ae tee gs |Tenaya ere ee! N 7 0.187 2 0.187 3 0.0793, 0.1954, 0.1055] __0.156 0.1000{ 0.1115] 0.0096 0.1954, 0.0791] 0.117 5 0.0593 0.082 | 0.1000, 0.2508) | 0.1042 0.0445 =) 7 | 0.1000 0.2150 | 0.0595] 0.0334,__0.0352 0.1209 ss |__0.0298| 0.0250, 0.022 S | 9 [0.1000 0.0403, ~~ |_—0.0132| __0.0188| 0.013 10 [| 0.1000 0.0060 ~~ |_0.0053|__ 0.0141] 0.0081 nu | |[| 0.0019 0.0106 0.0048 re el es ea ea re a 13 | 0.0002{ 0.0059, 0.0016 14 | L_|0.0001| 0.0045] 0.0009 fis | eee eS eR hi code cost aces sei — ~) Nn[LW SIN ln|n (oe) i) lo - =)¥) - | +o w-2o +20 u-30 u+3o 332 8.3723] 7.549 7.5492] 2.16721 | 8.3723| 2.167 | 0.2446 2.9016 _-0.5845| | 11.2446 9.0984] 3.0845 | 3.1168] 1.3524 -1,5017| | 14.1168 10.6476 4.0017] 6.0000, 6.4641| 5.4495 0.0000, -3.9282|_-1. one 9.9282, 7.8990 _-2.0000|_-7.3923| 4.348 10.0000|_13.3923| 10.3 [in (rr | 0.8665, 0.730 (a) 0.6000} 0.6665) 0.6941 0.7977] (b) 1.0000; 0.9817} 0.9900) 0.9786, 0.9437) 0.945 (c) 1.0000} 0.9983} 0.9996} 0.9972) 0.9822} 0.988 |W I|N co > PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT . 93 Section 4.8 Chapter 4 Sample Examination eras) (7 ) rae (4) i pe (as (3)-5 (b)— Geometric distribution. 6 (=) (Z| = 3.58%, 4 a (c) Negative Binomial distribution. (2 )(Z| | = 3.35%. — i>) © — —_~ 10gia 77-100 (b) on mea 88.5. (c) Res =5 (2 ae12 = 1714, (b) ELX] o arty=rounrant? «(2)2)2) —6X2+3] = 5-(3.933)-—6-(1.714)+3 = 12.378. E[5X 94 Di SOLUTIONS TO EXERCISES - CHAPTER 4 (a) A=2 > = 4 sneezes per 2 hour period (b) o=V4 =2 Ce (d) 1-e*[1+4+8] = .762. Fl=OX 213+281 =~ 6E[ X72] 413E[X] +28 - — 6{50-(.96)-(.04) + 487} +13-48 +28 = —13,183.52 6. E[(3X +4\(7-2X)] 7. (a) Pr(2 injuries next month) = (b) Pr(5 injuries next year) = eo (3). 7 = .0747. ee 51Go) ==a |N, Number of Claims [0 [7a] 23 , pm , 2 [om [SJ , 3 (4) | 2 \| -10 Nine WE / mf 3 l | ] Pot (+) Pot (+) Po + Gg Porte a - poyis(2)+(4) vf = rae This implies that po =.8. 0. Pr(N >1) = 1—.80-—.16 = .04 (A) Let A be the number of people completing the study from one group of 10. Then 4 is binomial with n equal to 10 and p equal to .2. Then, Pr| A >9] = 8194 19C, -.89 -.2! =.376. Let B be the number of groups in which at least 9 people complete the study. Then B is binomial with 1 equal to 2 and p equal to .376. Then, Pr{B =1] = (2)6376)\1=.376) = 469 (B) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 10. # 95 The scenarios where there are at least two more high risk than low risk include: \hhhh, hhhm, hhhl, hhmm\. Nae 2) ale n-p=8 Prex Wer: Cy (2) 03) £1, (25) and n-p-q 10) = = 4.8 implies 20 Cio -(.4)!° =O)ue =e g=.6, aCe (2)"23)2 = 0488 (D) p=.4, and n=20. Ly. Let X be the number of low-risk drivers with no accidents in a given month and let - Y be the number of high-risk drivers with no accidents in a given month. Then X and Y are binomial random variables with, E{_X ]=400-0.9 = 360 and E[Y]=600-0.8 = 480 The total number of bonuses paid over 12 months 1s, S = (X,4+:--+X12)+(% +--+ H2), 5S. E[5S] = 5(12-360+12-480) and the total bonus payment for the year is = 50,400 (B) a oe en aca ee ‘ “oi hp ) 7 tetas Saray shikai et Sard Fe hw ' 2%, oe 4 | neae Si seen Gieg eeca ‘ CHAPTER 5: CALCULUS, PROBABILITY, AND CONTINUOUS DISTRIBUTIONS Section 5.2 5-1 Density Functions (a) F(x) has the properties of a cumulative distribution function. G@) 0 itpex <0 F(x) = f@) = , eo ih x 0! so f(x)20 and therefore F(x) is non-decreasing. (ii) lim Fy(x)= lim 0=0. (iii) lim Fy(x) = lim(—e~*) = (1-0) = 1. x— 00 (iv) Pr(a<X <b) = Fy(b)—Fx (a). (b) See (a) part (1.). (c) Using the CDF, Pola 2) = FQ)-Fll) =U —e°) (le 4) a lew ee’), Using the probability density function, Pree [4e ditt ent ind = (e+-e78) ene (d) ae S) Prpx 2] (b) fx(x)= = Pr| (X >2)(X >))| ls Srv Be eee 2) » Tee Sh ea 0 former sxe for 0<x<l oe 2 & o-4 Ouror a = | (c) Using the CDF, Pr(.l<x<.5)=F(5)—F(1)=(.5) -(.1) =0.124. Using the probability density function, Pride = 50) = Ve fe CG MEAN Ne oe 3 pePrleny (ik 5) bP rl a Wi<5) ear ae) Pr(X <.5) wae Eri 5) aimee 4124 = 992% 125 97 98 5-3 SOLUTIONS TO EXERCISES - CHAPTER 5 (a) {- "a dy = ke” he = k =0-——= =. Since the total probability is 1, & =3. aa Ves |” BSok ke 3 te F(0)=5=1 implies k =3. (b) F(y) = [ke*at = ke Silo (c) Using the CDF, Pr(2<x <3) = F(3)-F(2) = (l-e38)-(1-e3) = (e% -e). Using the probability density function, 3 Pr? 2455) D8) i= I,36 “dv=—e—* X=zZ (d) Pr(Y >-3) [fay Ik3 (y)dy + IPSt(y)dy [.0dy+ [-3ePdy = 041 = 1. (a) Let f(t)=k for 82<¢<90. Then 1 = [kat = 8k, sok=—., E 80 | IT] = leg = 86. (b) Zero. The probability of any individual value is zero for a// continuous distributions. 87.5 | (c) Pr(86.5<T <87.5) = [86.5ght = 5-5 5-6 PrlX <2|X >1.5] ‘ : Prix alse as) RS PriX¥S16) PrL¥ >16|.X >8]= ee ; 20 ek .1-—.005x) ax ee osaaa e : TG I, (1-.005x)de = = PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 99 5-7 4 F(x) = kf (W041)? at = k or (VANS aa 10 oa ees 10+40 1 eP : |Pls 10 104+x]- icy uabas Pr(X <5) = F(5)=12. 5|5: es TER |p 10 (C) 12 Section 5.3. Great Expectations 5-8 Py ; (a) 4 E 33 4 x4 ELX] = [x aac = > < 5 — —— 5 le — — 4x Seo ss = 4x9 4) = x=4 wa x° 45 — x=4 (b) LX] = [x tatpele CAvk Ea CeGd ee 0 (Cy ip c= 0 756 1 0 = 4 = 4, l lie 4 co Ey (7 = oe ——“— = Sy ee Ae cy aaa E SS SS (dy Pron ane <3) = [2 2 kea=ye ada =ee 256 : 5-9 > k-(l+x)7. kek) s ibe j,—*_(ax? C dy = Pe ———_} (3140) is Lk3 = =. Therefore the proportionality constant k =3. ELAS fa = 5-10 ot 0 (14x) ax © 3(u-l) , [ free = E| VX | = |v ls Letu=l—x. AN eo ba eae 2e4\) [—. a {tiued’ Sa Aa 4-x-(1—x- ee )dx Pe 4.2Faoe x Aes 5 en5a x OD ¥ C) ( ees ill 100 SOLUTIONS TO EXERCISES - CHAPTER 5 2 5-11 Var[X] = ELX?]-(ELX))" = B18) : §-12 2 c=2. = 426. pr, dude =ee 4: Ee 2xde PS Bags ELX]=Ban iF: 5. ELX PAB 1= [x ==. peti’ =.0555. ratx1=3-(3| oy =.2357. The coefficient of variation for the random variable X is Pt A ipmet 5-14 5-(8) = 2211. ier’| xdu=—. Pao ELX Ai eee (a) ELX]== le ai (b) E[VX+1] = f° (eet a 2 2302, ] 5-15 1 = 35.36%, 0 (a) fry)= | tor Vez 0r 26 pa Boe ee: Graph of the Density Function Ie:aS cei B-(3) bo toe = 2.027. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 101 5-16 Gaowie eo Cd eS0 FOF Crew |= 1.27 Var|Coig |= 1.44-260 = 374.4 (E) f'(x) = 12x-6, so the only critical value is x =.5. Since f(0)=1, f(.5)=1.5, and /(1)=0, the mode of this random variable is x =.5. 5-18 f'@ = 7.5-37.5t+45t? -1503 = (7.5)-(1-t)-(2¢2-4t+1). P= land tea 1+0.5./2 . The critical points are We need to use the 2"! derivative test and/or check the values of /(f) at the endpoints and critical points. This random variable is bi-modal with t = 1-0.5/2 = .2929, and ¢= 1-0.5J2 = 1.7071. f(t) (By el2) Cie a7 Oe jiew ne PO 1 Ne I e702 <0 = el ove = =. Alternatively, l EZ] = 0+ |0 [I-27]a = ne (c) The mode forZ occurs at the maximum value of f(z). () (75-25 = .366. The mode equals |. 102 ® 5-20 SOLUTIONS TO EXERCISES - CHAPTER 5 (ayre=43 0 (Dei 4) t= for eo ee for, 1 for x<0 OS 1st x>0(~~’ (c) py =aut[jx-4xe-(-x 2 )dx =coTe. (d) The mode forX is jt: (e) x4 = Jase Both parts (e) and (f) can be solved using the quadratic formula. (f) 5-21 KO = 33% The answers are easy if you graph the density function. wy =xs5 =z and the random variable is bimodal, x = z/2 and 37/2. Probability Density Function f(x)=.25]sin(x)| 5-22 (a) No modes: | : /(x) 3) on [5,8]. wiles a. : Since the density function is constant there is no single (or dual) maximum point. All values 5< x <8 are equally likely. (b) One mode: g(x) = 2x-—6 on [3,4]. The one mode occurs at x =4. (c) Two modes: sin z f(z)= ea on the interval 0< z<2z. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 103 5-23 (a) f(w)= a for 0< w< 20, zero otherwise. 0 (b) F(w) = Pr(W<w) = for aa for 0<w<20. l (choppy for 20<w 20 |= \, we dw = 10 _ using the density function f(w). = E[W - w<0O 20 = 0+ if i Ww ass P ‘| =) using the CDF F(w). Also, since the density function is symmetric about 10, the mean is 10. (d) No mode since the density function is constant. That is, f(w) =.05 for all O< w< 20. (e) The median is 10. (f) we9 =13.8. 1 (a) Veer 31 te ee7A |,ise ere F(x)= x5 =J/13 =3.6056. 5-25 tor lay The mode is x =5. The mode occurs at the maximum value ofthe density function /(.x). . , f(x) = F(x) = for x<0 ‘ 4x —— x tote) 5 ee 3, 0 for. ax 4-2x The maximum value of f(x) occurs when f'(x) =— T= vis the moce.. (1) A hee ee 0. 5; 104 SOLUTIONS TO EXERCISES - CHAPTER 5 $15 Jac pane 200)" Pe = 93.06 (B) 8-27 BLX] = [x-flade = fn (Ele «(3 Jee BASRA yy 1/2.5 1/2.5 oe 200 2) -200{12 Section 5.4 Mixed Distributions 5-28 f(x) = x—-1 for 1<x<2 and there is a point mass of p => at x =1. We see this from the graph of the CDF F(x), which has a jump from 0 to 4 at x=1. barns ai ys ia +2 EX) PF slyeared t[e @Vae= es 7 el =. ee: E[X2]apa =1 sth? (Gl de el = S47 5-29 E[Loss] = E[Loss| no accident]- Pr| no accident |+ E{Loss |accident] - Pr |accident | ——— (90) - ele 0 probability of no accident expected damage, given noaccident I 5-30 (a) fas ree ——C10)——Y |00 x-2e*5 de = (.10)-—2l = .05.* probability ofaccident $7 expected damage, given accident 50<x<110. 110 x E{Payment]= 1000 |. Sp & = 80,000. E{Payment*] = 1000? ae x ak= 6,700,000, 000. Var[Payment] = 6,700,000,000—80,0002 = 300,000,000. Grane slr a20a) 90 (b) E[Modified Payment] = [/1000.x. wt 000 a lI 46,666.67+ 33,333.33 E|(Modified Payment)? ]= [,, 1000) pe O Modified Payment p = $16,966. = 90 ‘ = 80,000. Gp dk +1000? ya! - ae” = 6,688,888,889 Ss 19 20000 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 105 Section 5.5 Applications to Insurance: Deductibles and Caps SS 5-31 a E[Benefit] a 10 NE y, \ OR = 1.8+10(.01) = 1.9 5-32 (D) Note: The maximum payment is | million dollars, so solution choices (D) and (E) are ruled out. i x(4—-%) 3 x(4—x) eee a ed E{ Payment] 1204+ .8148 = .9352 5-33 iy er Xe) Payment = - = 0 (C) tor Vax <1) ieee ise fi OSC .64 = Pr| Payment <.5]-= Dipeaai tee = Pr Xx 225 tC) 5eCy. (.5+C)=.8. Therefore C=.3 (B) 5-34 The loss NOT covered is equivalent to the loss subject to a cap of 2. Thus, E[Loss not Covered] = eho] CUS 0 Big a ae 2.5(0.6 ee Alternatively, using the CDF: 2oy 2509" ax = Oa wee) z F(x) = i-(4) Oe k= Ol, Let Y =losses capped at 2. Then, aC [-£) 6 dx = 93. E[Y] = 6+ |GNX 5-35 E{[X]=500. We want to set a deductible, D, such that E[{Payment ]= 25%(500) =125. : 1000 | 4 (rn))4 125 = 0/Pr(X<D)+| (e-D)- Baa Ute Therefore, D=500. (C). r=1000 a L000): pean - 106 ® ye SOLUTIONS TO EXERCISES - CHAPTER 5 Tapas [oy Eat 4 PLY > 4] =1.644-(2)=24. FLY?) = fy dy +42. Pr[Y > 4] =4.26 +16-(.2)=7.46. Var[Y]=1.706 (C) 5-37 E[Not Paid] = [x5 dr+10-Pr[x>10] =.066+10-.99 = 9.96. 5-38 f)=35 for 0<1<75. me _ 10 E(Benchit)|= {- iedt +40: Pr[40 <T<75] = = ip ‘eit Se I ares (100)° ‘ae, eG oe eee: 160-507 100 ): t=0 O<ses100 E[Payment | X >10] = mre| sap 100(100—x [ =| \" dx i ae | (100-10) 1-F(10) 3-100? 100-10 3-100 100 5-40 i l=K ili tata Net Premium Insurance Payment, w/ deductible 60 137 ted| implies K=——. oe OK hen ee = .95-(0)+.05- aed [. 2. 15 4. 12iih =~ 0313 \ tes ogee Ue ls Ae a 5 iw" PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 5-41 The health care cost variable is represented by Y. The reimbursement random variable is 0 TOU R=,5X-20 pAes20 [Ore 100+ .5(X -120) for 170 X >120 A reimbursement of 115 implies that the loss Y is more than 120. Thus, 115 = 100+.5(X-120) gives X =150. The reimbursement is positive if the health care cost is greater than 20. G(115) = Pr(R<115|R>0) Pr(X¥ <150| X > 20) _ Pr(20<X<150) _ F(50)-FQ0) _ a9, Feit 30) ee oer ae Section 5.6 The Moment Generating Function 5-42 In evaluating the following integral, we use integration by parts. ne Ge = EX] IsNed d= ELX2] = lex ede 5-43. S44 545 (-xeek H x=00 l x=0 2 ; j = 0.5. oy =y.5-.52 =.5. My(t)=.4e* +.6e™. E[X]=My' (0) =3.8. M(t) =.8e7! +3e*. E[X*]= My" (0) =16.6. M(t) =1.6e7" + 15e*. Var|X ]=16.6 —3.87 = 2.16. My(t)=.3e +.7e' =.3+.7e'. My (t)=.7e'. Ee leew jae ee M}(t) =.7e’. Var(Y]=.7-(.7) =(.7)(.3). (a) My(t)=.5e7 +.2e# 436°", (b) M'(t)=—.5e~ + 4e” +1.5e%, so ELX]= My (0)=—.5+.441.5=1.4. (c) You could find the probability distribution for Y and the MGF from the definition. Mii = May_4(t) = a". My (3t) l| = ew. (Sao +.2e° +,.3e!> ) 4 5e7/t +.2e% + (3et", ® 107 108 @ 5-46 SOLUTIONS TO EXERCISES - CHAPTER 5 (a) M(t) =.3e! +.2e7'. M(t) =.3e! +.4e*". EX M7 (0) = 5.00 Van Oo. na ey 25 = 45: Re Een) EIEN (c) The mode is 5-47 x= (Me (0) 03) x=0, the median is x =0, and the midrange for_X is = =1. 4 (a) M’(t)=5(.8)e" -(.2+.8e') 3 M3 (t) = 4e! -3.2e! (2+.8e') +4e' (2 +.8e") (aVara E[Z2]=16.8. 4 Var[Z]=.8. =e" (2487). (b) My (t)=M3-7z(t) =e" -Mz(-Tt) You may use My (ft) to verify Uy = E(W]= E33 -7Z]=3-7E[Z]=3-7-4 5-48 and op =|7|-J8. The key to this problem 1s to find the probability distribution for Y. the time (whenever any of the Y¥;=0). and X; =1. Y=1 Y =0 most of ifand onlyif X; =1 and X>=1 By independence, this happens with probability = Thus, Yisa Bernoulli trial random variable with p= + . Therefore the answer is (A). Note: All MGF’s have the property that M/y(0)=1. So (B) and (D) are NOT valid moment generating functions. 5-49 My (t) = My,-2x,43x,-4() = My, ():My, (-2t): Mx, 3t)-e™ PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 109 5-50 Mt Mae fe i |= |e as ees ion Ie ; 3 Mery. x(t) G-12 M‘ (t)= ah B- Dah: y2]-2 er) ea ess ee 3 Omit We use the shortcut method: — A(t)=In M(t) =—4In(1— 25002). A'(¢) = 10,000(1 —25002). h"(t) =10,000-2500(1— 25002)*. oy =,/h"(0) =100-50=5000. (B) Alternatively, M‘ (t) a AU (1- 25007) ELX]=M'[0]=10,000 and M(t) — 125,000,000 (125007) and ELX*]=125,000,000. oy =5,000'= B. 5-52 Since X and Y are IID, we have M()=MyirO=Mx(t):MyQ=Mr(t): My(O=Mx(O’. Thus, M(t) =0.09e~*! + 0.24e~ + 0.34 + 0.24e’ + 0.09e7’ must be a perfect square and so My(t) must take the form My(t) = ae‘ +bh+ce’. It follows that a and c must both be 0.3. Since a, b, and c must sum to 1, we have that 4 must be 0.4. It can now be easily checked that + 0.24e! +0.09e7" M(t) =0.09e~*" + 0.24e7 + 0.34 Therefore My(t) = (.30e7% +.40+.300e'). So we can reconstruct the distribution for X : PrLX <0] = .30+.40 =.70. (E) = (.30e%+.40+.30e')? . 110 @ 5-52 SOLUTIONS TO EXERCISES - CHAPTER 5 fie sok see ee > MOQ) P Mx() == Oe ElX]=Myo)=£3 a by (1-q-e') Dp legal raheemlee) 2 Pp (ere) 1G (oe ee P?-4-2p-q-P-(-4} ae ee ee_g(p+2¢) Re »,_ eee Pp Vari x)= P 2 ‘(p+ 42+ D2 isc +2 ie 5-54 Pp p usto=| ie 1-ge' Me =Io( M0) =o ayes P }=r ; P l-qe —qe' (0 Teter J=rincey rina-ge' E[X]="@=—t. } meo=r (oe!) 5-55 euee ke en ; Dp Pp This is the MGF for a Negative Binomial random variable with r=4 and p=.3. (b) Pr{V =2]= 5C3-(.7)° -(.3)4 =.0397. 5-56 0 Var[X]=h"(0)= Petre _ (a) My(t) een a) (b) My(t)=Mrx43(t) =e* ae = ee" +31-6 | PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 111 oy 5-58 Yea h(t)=In(M x(t)) = n(eHe) ale -1). ELX]=h'(0) =. Var[X]=h"(0) =A. If Mz(t)= eles then the random variable has Poisson distribution with A =3.7. The mode occurs at z =3. Section 5.7 Chapter 5 Sample Examination ies (a) ae fae 1=[ or = x == [fax and Se Solving for a and 5, we find the constants are (ob) ieee cee a=3.6 and b=-2.4 Pr(X <.4)= iff(x) dt= 2368: (c) -ELY2] = [G63 -2.4x4) de 229) Var|X \=".42= (6)* = 06) so oy =0.245. (ay meeey (t= .0% 2 Ay (x) 9, 6 = 2Ax = OC ye = 2/52 sincesthe graph of the density is a downward directed parabola, we conclude the mode is x =.75. 2. This is an exponential random variable that we will study in chapter 6. tae Pree) = GS) 97709: (bo) 5=F(x) = S=1l—e"" => the median is m=.2310. ix) ae '(x) =3e°>* forx>0. E|X] = ux (oS xy Thus, the mode is x =0. =k [1- F(x) ee | = I.eo ax = = 65 = Le mgs = 3499. 112 ® SOLUTIONS TO EXERCISES - CHAPTER 5 Re. ere = flee wet) dy = 84375, Fy) = (a (6b) = 13y-¥3 +2]for -l<y<l The medianis_ ys =0. The mode is y=0. The mean is fly =0. (CymlGsS=F Oy)= 43 y-y 342] = yes =.20278. (Use a graphing calculator or Excel Goal-Seek.) (ely 4. Jape A = = f,0? -y\)dy E[T]=8.5 and o? ==. The coefficient of variation is §, = =. Since wy =0, oF ==. 100-6 We = 100- 53 6 0% = 16.98% Ell Cee E[(X-2))] = ELX? -6X2 +12 -8]= E[X3]-6E[X2]+12E[X]-8 = 0. 9-6-E[X2]+12-2-8=0. 7 OS X*|=—. [X7]== E| ps | Var|ar| X|)=—-4=—. X| r r 6. TheCDFisf (x)= [cencus ars pire —e?.004~) Let Y be the benefit paid. Since F(250) = 1—e~%*29 for X occurs to the left of 250, where of X. The median m=250 ln 2elio29 Te ie = ae 004 (Ch: of X Thus, the constant c=.004. = 1-e7! = .63, the median Y =X . ike the median of Y equals the median is the solution to és = ]—¢ 004m) ‘ 250, the mode is x =0, and the standard deviation is oy = 250. Which is PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 113 The CDF is F(x) = f°c(l0+1)? at = a5 os | 0<x<40. 2,8; a 2 Sones (40)=¢ 55° 80 ¢ = jesKore l=F(4 Then Pr(X¥ <6) = F(6)=12.5- P | = 46875. (C) l has 0 F(x)(Cele =1.25 -—= fio for 0< x < 40 7140) iiaE —F (x)| dx lI ("[1 F(x) Jd 40 The mean is J Sie WAS ae 08 +25] < = (—.25)(40) +12.5In5 =10.118, the median is 5=F(m)=> m= S. and the mode occurs at x =0. This is an example of Beta distribution. (a) Tra= 12; (b) Pr X>.80)= [sgh2x(x-l)? jie aiDae (c) l l Var[X] = ih (x? )12x(x-1)? 2 dx -{[, (12x01? 2 as|; = .2-(.4)? 7 = .04. (d) The mode of the random variable X occurs at the maximum value of f(x). f(x) = 12x(2)(x-1) +12(x-1)? = 12(x-1)[2x+x-1]. The critical points are ; ; | x =1 (not in the domain of f(x) ) and x= = The mode is x =>. (Awe PA eS) eas) (b) The mode is x=0. (c)) ix =: (d) 1-—e-* (6) sox =I) = le? = .80 implies xg9 = 551, = —In(.2) = In(5) = 1.60944. 114 ® SOLUTIONS TO EXERCISES - CHAPTER 5 12. (a) EfDamage] = (.92)-0+(.08)-1750 = 140. (b) ETE E[Ins. Pmt.] = (.92)-0+ (08)}say 0 —- 1000 1 3000 i ] =10 0)- ee) as meas .08| 800 | = 64. 13. (a) My() = Ele] = [oe% Sede X= = Ctds= = ets] fess = = 5 [e : 14. er 5—f rcs. : aay (ye 15 G20) (b) Ae (ey OnBr (a) Graphing the CDF allows you to see that X is a mixed distribution. The peer point masses are PrlX¥ =2]=.3 and PrX =3] =.2. (2) dv+3-(2) = oie24543See = 5 (b) ELX] = 2-(3)+ fv ct Miners cers J (x) 1S; If My (t) a — 4e then W has geometric distribution with p=.6. (a) ow = Pe ) 16. (b) — PrfW =3]=(.4) (6). Since the man purchases the insurance at age 40, we are given that he has survived to age 40. E{Payment] = 5,000-Pr(40 <T <50|T > 40)+0- otherwise conditional proability an insured dies between ages 40 and 50, given he is alive at 40 = 5000- F(50)—F (40) 1—F (40) _ 348. (B) CHAPTER 6: SOME CONTINUOUS DISTRIBUTIONS Section 6.1 Uniform Random Variables 6-1 (a) f@== for 60<d<90. F(d)=2—™ for 60<d<90. 6 : : (Db) a = “as = 75, conveniently located midway between the endpoints. 30" o5=7 55> (c) 9D = 5/3 =8.66. d 69 oa 60 Solve ——— . th 2 J =.69 to discover the 69" percentile is d 6 = 60+.69(30) = 80.7. ON=67-5 by Dee Ea a 30 30 6-2 (a) fr(y== for -2<y<5, Fy) = b (b) +5 wy = —— a =1.5. =1.5. ie Dees Need: oyo7 =—. 12 (c) The median for a uniform random variable is always the midpoint of a domain, same as the mean. Here m=1.5. occurs for all -2< y<5. The maximum value of fy(y) is +, which Therefore there is NO mode. 2 (dierent ye< 0) ace (ey 0Ptcy = 1) = 1), 6-3 We let 7 denote the arrival time of the cable guy in hours (after noon). So a value like T =1.4 corresponds to an arrival time of 1:24 pm. (a) f(t )=— and F() =~ for 0<t<5, (b) sr =2.5 which corresponds to 2:30 pm. eS) o7 = 17 square-hours. (c} Pr[3<T <4]=— =20%, (d) Pr[3<T<4|T>2] = > = 115 116 ® SOLUTIONS TO EXERCISES - CHAPTER 6 tae Pfs oy SX say roy] = eb tater eth 7. b=a ne Pe ae tee 128 02 (b) Pr| wy -20y <X <py +20y |=Pr Gi peheey 2ae 2c ath Convince yourself that _,b-a a+b < fi ee eee + b-a ie: so that this probability is 1. as 6-5 w4=10 and oy oily 6-6 This uses the methods of Section 5.5. Suppose D denotes the damage to an automobile, D ~ Unif(0,1500). Then the insurance payment is given by 0 if if0<D<250 (with probability 4) - |D-250 if D>250 (with density 7) | Et] = 0-44 [-”(x-250)-—41500 ae = 520.83. 6 | 4250 EUI?] = 0? + e) a} [i (x-250)? zor de = 434027.7. ] 1500 2) ] = = 403.44. 5 (B) Section 6.2 The Exponential Distribution 6-7 | [ae Sketch the graph of f(x) = lo if x20 _. _. Itis clear that the mode occurs at otherwise x=0 regardless of the value 2. Using calculus, there are no critical points. Checking the value of f(x) at the endpoint(s) gives 6-8 x =0 as the maximum. Fy(y) is the CDF of an exponentially distributed random variable with mean +. The > variance 1s (4) and so oy = 4, You could also calculate Var[Y]= E[Y?] —(E[Y]) ra using eer ers 2 the density er t ion function ron AS ye {0 for y <0 |2e “y for yp20 : PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 6-9 (a) Oz = 23 (b) (c) The median of Z is 5-In(2). Themodeof Z is’z=0: (d) Prl<Z<3) = i —2/5 dz = e > —e3/5 = 2699, Oryusing F(z).= 1—e-¥*, Pril<Z2 <3) "= F(3)—F CG) = ee" (ce) Oal0 Sa) = 2609; Careful, f(z)=0 for z<0. Pr(-1<Z <5) = Pr(0<Z<5) = 1-e7, 7 (f)i= t 1 4.0750 6350 ite 0 7 =0 otherwise t Also, F(Z) = le 07> 0. —750 (b) Pr(T <750)=1—e 70 =.6321. (c) m=750-In(2). Conditional density for X, given X>p 6-11 E[X|X>] = [x oo (22 e749 dx Substitute V=X—-Q CO ss = | TeOe 2 ged “FF Vie fo@) i,(y+): Ae? dy 00 oe) \,yp he A) dy+o | AeW*) ay = 6-12 [47 dx+o-1 = E[X]+Q. : ——. ) Q) (b) The mode for an exponentially distributed random variable is always zero. In(2) (Cc) A=—. 7 —5-In(2) (dyer <S)panl—er 7 == 0.39, 6-13 B=a=10,s0 O;-Q, = —10In(.75)+101n(.25) = 10.986. # 117 118 a 6-14 SOLUTIONS TO EXERCISES - CHAPTER 6 3 < ‘CO E[X°] = 2 : [,35 e —x/7 7 = av =) 2058: The integral can be calculated using multiple (or tabular) integration-by-parts. Also, we will learn how to calculate this using the gamma distribution in Section 6.5. However, the easiest ee is to use the MGF: = ELX?]=M0) =-7 L20--1| 5 =C28)2° = H(29 4 = 2058. 6-15 Since 8? =9, wehave B=3. ELX*]=M™)(0) =1944. 6-16 E[7Y2-11¥°] = TELY?]-116| ¥°| = —5,773,554 (using the MGF). 6-17 The median equals 4, so the mean must be £ = aa =5,//1. PKG 5) 6-18 == Foy =e oo (Dd) First find the parameter { of the exponential distribution. 30=Pr(X <50)=1-e9/F, 6-19 = 429 B=140.18. Let L denote the lifetime ofa printer. E{refund] Pr(X <80) =1-e-80/140.18 =4349. / 2. Then Recall Pr(Z </) =F, (/)=l—e 200-Pr(0< L <1)+100- Pris L<2)+0-Pr(L 2 2) = 200-.3935+100>.2387 = 102.56. Since there are 100 independent printers, the total refund is 10,256. 6-20 (C) Let X denote the auto loss, ¥ ~ Exp(300). losses that exceed the deductible. ¢ OS Pt X= ACA : ak 00) = (D) Let 7 be the 95" percentile of actual We need to solve the following for 77: Pr(X >7 0X >100) Priv > 100) = Pr(X > 77) PrX>100) oat. = e 300 3 7 =100—300-In(.05)=998.7. (BE) =~ a 3067), 100 ~ PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 6-21 # 119 Let T denote the time from purchase to failure, T ~ Exp(10). 1000 = x-(l—e7/10) 4 2. (g-10 _ 93/10) — y..1772. SS SSS ———— SS — Pr(O<T‘<1) Solving for xgives, 6-22 x =5644. Pr(1<7'<3) (D) Method 1: Recognize the MGF as that of an exponential random variable with mean 2. E[L00(0.5)*} = ["100(0.5)' ser? ee = 50 [eno eo X!/2 dy 25 i) [ eX (1/2+In(2)) hy = 0 lee. 0 = 41.9. + +In(2) Method 2: Using the definition of MGF (with some slick work). Recall, My (1) =Ele*]= os C2) 0k 10ex Ie= 100E[(0.5)4) FALOOCO:5)2 II 6-23 : 100 100M x(In(1/2)) = iain = 41? F(x) =1- e 0: x>0Q. Let Xy be the payment per loss with a deductible of d. We use the CDF method of Subsection 5.5.3 to obtain, x BX gh = [ [1- F(x) |de = oo IkC0 dy oct = Gee, 100 Then 2000=6e ? . Also, E[Xs00] 6-24 = (200021 Ge500/0 F(x) =1 -e 8 -x>0. |e-5009 = 2000e4? Let X¥% be the payment per loss with a cap of d. We use the CDF method of Subsection 5.5.3 to obtain, | [I-F(a)|dv= im |e —x/0 dx-= 4 O(1-e74/), ei) ELX¢]aby =— I, 1 d / 8), Then 20 = @(1-e7 190 Thus, EL) § (C). (1 — e7500/8) = 60 =e") = 205(1 ——_* ae = 100/€ ) ) (8). (C) 120 @ SOLUTIONS TO EXERCISES - CHAPTER 6 Section 6.3 The Normal Distribution 6-25 (a) Pr(-2.98<Z<.99) = &(.99) -D(-2.98) lI @(.99) -[1-D(2.98)] = .8389-1+.9986 = 83.75% (b) Pr(-1.75<Z<-1.04) = [1-.8508]—[1—.9599] = 10.91%. (c) Pr(Z >~.23) =59.10% (QePHIZ 12073) ei P13 27 73) = 1-2 {lO = 1-2-0907 = so (ec) Pr(|Z|>2.05) = 2-[1-®(2.05)] = 2-.0202 = 4.04%. 626 (a) 10s Pr)2227 |31 —Py| Z sz, Z 9 That is, one needs to be 1.28 standard deviations above the mean to be (b) in the top 10%. Zo5 =1.645. (c) P49) = Pee (dy 95 = Pr 222,|=I Pr| Z<7, |= Pri 22, | 05 =e 05 \| ss NR — = II Oana 1 O(=2Z,) — 29) Se = Oe (5) Sn e705 ea OF O Thus, 27 ='Zos =—1.645. (e) Paar ® 122 (f) Probabilities are always numbers between 0 and 1, so such a zy does not exist. 6-27 (a) 1-.2119 = .7881 and z78) = .80. (b) ®(z_)-[1-D(zq)| = .9030 = @P(zy) = 9515 =e 6-28 2 = 255;35= (a) Pr(X <84") = Pr| Z (b) X >79"+(.842)-2.5" < 156 84-79 )= Pr(zZ <2) = 9772. 2.5 —————_- = 81.1 inches. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 121 6-29 IO > 100+2.055(15) = 130.8. 6-30 Let T denote the completion time of a competitive marathon runner, T.~ N(41=200min,o? = {24min}?) (a) T = 200-—.955(24) = 177.1 minutes. (b) T = 200+2.17(24) = 252.08 minutes. 6-31 (a) Pr(F>16) = Pr{Z> ent) eau (b) F > 11.7+2.33(3.9) = 20.8 inches. hee Pr Ae 60) = Pr[Z< 2-862. 764 (b) Pr(48< A<59) = Pr(-1.64< Z <.56) = .6618. 6-33 (Gc) A = 56.2—1.28-5 = 49.8. (a) Similarly, me = —.33. A test score of 79 corresponds to a z-score of z =.5. A test score of 96 corresponds to a z-score of z =1.92. (b) G = 73-1.62(12) = 53.56 and G = 73-.47(12) = 78.64. 6234 6-35 #(3) F295 = C3005) = Pre i 7 <7 1) S52. (b) C 2 30.00+1.645-.07 = 30,115 inches. (a) Bob would be assigned a *C’. (b) Course Grade | Percent of class Range |69.5,81.9) |46.5,69.5) |34.1,46.5) | (-co,34.1) 122 SOLUTIONS TO EXERCISES - CHAPTER 6 6-36 Pr(50</O0<70) = Pr(-3.125< Z <-1.875) = 3.0%. 6-37 (a) 1.22 E[Y]=e*2 =e°82 =336.97. Var[¥]=e26DH2 .(el2’ 1) =365,710. (b) Pr(l50<¥ <250) = Pr(150<e* <250) = Pr(In(150) < X <In(250)) =Pr(-.07 < Z<.35) = .17. (c) The top 10% of the normal distribution corresponds to z =1.282. In(Y)—5.1 19 =1;1.282 implies impli Y == 763 6-38 (a) leet : This implies (e?* -1) = 3 and oy =,/In(4). (b) 6300) (b) Has 2 =f" eit tells us that wy =0. Pr(l<¥ <3) = Pr(l<e* <3) = Pr(0<X <In(3)) = Pr(0<Z<.93) = 32. ee bi] som 2e=e! 1197, Vary] = ee te 1) = eo. Pr(.9<L<1.5) =Pr(.9<e* <1.5) Pr(In(.9) < X <In(1.5) . py{Be 4 (c) : E[Y]=e "2° =2 and Var[Y] = e2#**%% -(e% -1) = 12. See lise) 4 ) ere The bottom 5% of the normal distribution corresponds to z =—1.645 aoa = ~1.645 implies Y =.5724. Section 6.4 The Law of Averages and the Central Limit Theorem 640) BUY eG ance or ee eee The exact binomial probability is Pr(¥ =8) = Cg -(.3)®-(.7)* = .1144. Our approximation using the normal approximation is Prix =8) = Pr{778 7 eo —| = Prgsee e122) = .12. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 6-41 # 123 (a) Let H represent the number of heads on 100 flips of a fair coin. H ~ Binomial(n =100, p =.5). We will approximate H with a normally distributed random variable X ~ N(u=n-p = 50,0 = J100-0.5-0.5 =5). Pr(H =60) = 100Ceo «(.5)™ -(.5)* = .0108. This is the exact binomial probability. Pr(H > 60) = 100Ce1 -(-5)® «(.5)°? + 100 Con + (5) -(.5)° +++ 199Ci99 -(.5)! = .0176. Pr(H = 60) = pr(2550 By < 05-50) = 4821—.4713 =.0108, 60.5-50 (b) Pr(H > 60) = Pri ZS aoe A 6-42 Mis Pretty nice approximation. The number of no-shows is NS ~ N(u=9.3,0 = 2.95). PrOvs = 2) 5) Se ET(WA 6-43 Pr(Z > 2.1)=.0179. 4.5—9.3 = Pr(Z 2 -1.63) = .948. 795 Let S be the number of sales. Then Sis binomial with n =500 and p=.1. hus; 6-44 (a) 5 99:5 =G00)0:1) Pris = 56[°= m2 ~ 1(500)(0.1)(.9) “|el 939 =e 20/6/ae BY. E[X;]=75 and VarlX;]=l| 35. (b) E[X]=75 and Var[X] =a) =(3) . (c) -Pr72<.4;<77)\= ay =.1. That is, only 10% of the time will a randomly selected number will be between 72 and 77. On the other hand, for the sample mean, Pr[72< X <77]* Pr)Uo <Z< | 6-45 (a) Let X= then (Dyin X, ae X?2 Fama eet X 100 100 Pr(98<X wet X= = Pr[-1.8 < Z <1.2]=84.9%. = <102) Xy + Xp +--+ 400 —2 2 pref2 <7 <2 = attttow X00 , then = Pr(98 < X <102) 5 ap = pr{ 2 <z < 2 = ,9670. C (c) There is a larger probability that the sample mean is within 2 of the mean. 124 SOLUTIONS TO EXERCISES - CHAPTER 6 6-46 Theerror is U ~ Unif (—2.5,2.5). Let U = al PH 252 and Var[U] = 2.083. be the sample mean. <5) = |ee ee E[U]=0 J -0434 fe er = see ALE 5000/25 X;~N(u=3,0=1). and Var[U] =.0434. (A tes )= 77. .@) /.0434 NTN Ligh THO 6-48 os +U 4g E[U]=0 T = X)+X2+---+X,, is the total lifetime of n bulbs. E\T|=3n and Var{T |= n. ape mT > 40) = Pr(Z > Ie }Since @(2) =.9772, we have nf 2 Ee Oran ee Ore nh x=Jn. Then 3x2 —2x—40 =(3x + 10)(x—4)=0 implies x =4 and n=16. (B) 6-49 Let S=X,+-+-+ X2925. Then ps =3125-2025 and os o=n[sss]-| 254 ee | ae : = 250-/2025 Sel = 6,342,548. integrate f(w) = 12w? -12w°; 0<w<§<1 to find: eer [2m = 2 , E(w] Ps == oi [2m 2 | lawi)dw = Aaa 2 a 4 ‘ Ot -_— Psi == = = = law? dw = ee 9) i) = tps 250:,/2025- 6-50 You will learn in section 6.5 that W ~ Beta(a =3,8=2). E{W] . Thus, Elin abe EEE @) g§ = 3125-2025 +1.282-250 4/2025 = 250-2025 = = = = 2 i = =->eee = 5e. Then, Var[W] (a) E[S]=225.>= 135 and Var[S]= 225. =9. (b) Pr(S<141) = pr{z < HS 9 = 773 (C) In the meantime you can PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 125 6-51 You should recognize the density function as exponentially distributed with mean 1000. Thus, each policy is sold for $1100. The variance of this exponential random variable is 10007. If X; denotes the claim of the i" policy holder and T = X; +-::+ Xjo9 represents the total claims, then we are asked to estimate Pr(T > 100-1100). E{T] = ELX, +---+ X10] = 100-1000 = 100,000. Var[T] = VarLX, +-:-+ X00] = 100-VarLX;] = 100-1000 = 100,000,000. or =10,000. Pr OrodOy hyena) mene Zot 0 S100 DOU Ise oe Sy 10,000 SoD) Normal Approximation 6-52 N ~ Poisson (A=2), so un =2 and oy =./2 . Let S = Ni, +---+ Nj250. Then “5 =1250-2 = 2500 and os = 1250 we Si 2200FUe Pr[2450<S < 2600] = P| _ 2450-2500 50 <7< 2600 — 2500 _) 50 = .9772 —(1—.8413) = .82 (B) Note: Technically, since N is discrete this application of the CLT should use a continuity correction. But, (i) the statement of the problem makes it indeterminate which way the correction would go, and (ii) os is sufficiently large that it would make little difference in the final result. Section 6.5 The Gamma Distribution Geom ye i al(275) = 22 Dal sell) (b) 1(5)=4!= 24. (c) (2)=1. GAT) 6-55 = [, xe td “to mar md pester dees TitA\s = 30: ee (5) = = 2.071197,01 15 vr 8 126 ® SOLUTIONS TO EXERCISES - CHAPTER 6 6-56 (a) M ~ Poisson(A = 4) is the number of insults permonth. 23.34 Pt 3) = ‘ik 1680. Och. YP Gv is spell je a he ~ AGS (b) W ~ Poisson(A =1) 1s the number of insults per week. -1,40 Pry=0) ee 7 = 3679. (c) (d) PrW>1) = 1-.3679 = .6321. 3679 - 6321 — Vi = .2325, —-———" not insulted at least once first week next week (ec) Let S be the time in weeks until the 4" insult. Since insults are occurrences in a Poisson process with mean rate 4=1 per week, we have S~ /(4,1). Let Ys be the number of insults in a 5 week period. Then Ys; ~ Poisson (As =(5)(1) =5). Pr[S >5] = Pr[4” insult arrives later than 5 weeks] = Pr[at most 3 insults in a 5 week period] Then, =o Priye= 6-57 (a) 0, 2.3) Z 3 os(ies+ 4S45 |WG Dis Soe! 2ES: W ~ Poisson(A =1) is the number of failures per week. =1 42 PrW =2) = <=" = 1839. (b) The mean rate of failure is one per week, or 1/7 per day. Let T be the time in days to first failure. T is exponential with mean {=7 days. Let D be the number of days such that .5 =Pr| F< DI. Then D is the median of T and so Doin =F ine = 4s 6-58 (a) ELX]=10=a-f Ly es= an (D) efx) q BS (C)SWSerrat ee yee Srdays. and the variance: V[X¥]=20=a-f?. SE re erenoe oe ee An edt esi ene) lay oat aS =. lee d 2 Lheng Prix = 6) a (6) So B=2 (Ax) ; 7 3 as without integrating? Rte 24 = .1847, 4 (d) E[4X -5X?] = 4E[X]-S5ELX?] = 4-10-5-120 = —560. how to find EX] I with A= sn =5,5=6. = Ine|tsaedrs a om yan ] and a=5 Do you remember E{X?]=Var[X]+(E[X)) = 20+102. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 127 6-59 (a) E[X] = a-8 = 12 and the variance: VLX] = a-f? = 24. (b) Property (6) of the Gamma following integrals: Distribution is quite helpful in evaluating the a=9, B=2 x EV Ye BEA oo) ] i,)x 3: see) = 5, >. pp li2x xe z dx | esi 4 nn ioe) See 2% ihXYre dx = AY (c) E| VX | = Se = PRG = 2688 ER a=6.5, B=2 = a= he erent dy =acinsall neeee 6.5 -I"(6.5) 2 Aa 6.5): (4.5)- 3.5) (2.5): (1.5)-(5) Vr = 3.393. 6-60 (a) D~ Poisson(A = 2) is the number of business trips by your dog over a 2-day weekend. Pr(D <3) = Ce) 01 echo Seaiy he 0 a H er 5 = .857. (G4 1) 80 f= ate"; t>0 (c) Pr[T <2] = Pr[4” trip occurs before end of weekend] = 1—.857 = .143. In other words, the event 7 <2 is the complement to the event that 3 bags last through the 2-day weekend. 6-61 S~I(50,2)= 72(100). E[S]=100 and Var(S) =200. (a) Pr(S<5-J2 -o5)=Pr(S <5-J2 -J200) = Pr(S < 100) = Pr[Z <0] =0.5000. (b) Pr[S' <100]=CHIDIST(100,100) =0.5188. (c) Pr{S <100] =GAMMADIST(100,50, 2,1)= 0.5188. Section 6.6 The Beta Family of Distributions 6-62a B(4,2) = [B(l-x)lde = ~5 = {0 c'(l-x)3dx = B(2,4). . he canes : Of course we would not integrate in general. Mae LCM) oa ee stl! ee 1 rr 128 @ SOLUTIONS TO EXERCISES - CHAPTER 6 sete RAs HOO ens ay EB 1215 3960. Tyo) 25} B(3,1.5) = 6-64 TAS) 235225 15.0805) = 1524, X ~ Beta(2,6). SINCE I ee FB) eeTOTO (a) °=Fa) = 2 ee 42. oer? (b) Kase Vary x |= OEOs substitute w=I—x () PX> 5) = [42x18 de & f242(1-w)u5 (du) = 42 [G8 -u®) du = .0625. ‘(O) Oe a 42x(1-x)>dx “implies Q, =.138. 6-65 X ~ Beta(2,2). (@) fy (x)= 0 if x <0 43x°—2h ei 0s): 1 Ne ames (b) ELX2] = hee Someta (Cy Van xa 3-(4| S105 al ee ee 25 fo fa 2236. med iE 6x(1—x) dx implies med =. N (e) The mode is the maximum of f(x) = 6-x-(l-—x). 6-66 X ~ Beta(2.5,3). Aer EGS) ; (Seats) i eee (hae a Ole (d) This occurs at x =.5. ease os 2 eek: ae eS (5.5) (Gr5) ioe Pr.4<X<.9) = [, 19.6875-x° 2. (1-x)2dx = .582. iy) ie) > PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 6-67 E[TX-5X°] = [,(x-5x6)(60x2 (1-2) dx = 60.|7[8-2 27 de—s i;x8(1-x)3 as LAOS r(8) oro T(13) 81:3! ey= ii Peeskst he,PE 7! a a 6-68 faye Y= Betala = 1,B= 5). (Dye =o) (cy y= PY ay) =] 1=1= yy. (d) y=1-.25'> =.242 thousand dollars. That is, the students needs $242. (fe) Y= =. On average the students spends $166.67 on beer per month. (f) Note that Y lives on the interval |0, 1]. L ] Pr(.8<Y¥<1.4) = [,50-»)4ay = 00032. 6-69 f= 30x! -(1—x)* on the interval 0<x <1. ig F(x) = Pr(X¥ <x) = 41-6(1—x)? +5(0i—-x)® r Here socal) for O<x<l. fOr ead Use a graphing utility to solve for the requested percentiles. 5 = 1-6(1-x5)> +5(1—x5)° implies that the median is x5 =.26445. No) \| 1—6(1-x9)° +5(I—x9)° implies that the 95" percentile is x9 =.5818. # 129 130 ® SOLUTIONS TO EXERCISES - CHAPTER 6 Section 6.7 More Continuous Distributions 6-70 (a) 2 5 5, Pr(3<X <4) = F(4)-F(3) = (2) HH) |= 1004. (BELLY =2.5. (©) pA 2 2=B[X?]-2.5? 752 GB _{_@B _ 20(10) Var[X] = = CE (2)) -4 (0). This implies ELX?]==2. FY ReB | = 7-2,5-3--2 205. 6-11 (a) On EX] geope =F 14 = T2 = 56, abo = = egy Sone 3.5 Q ( yiesolving.> = 1-( (dye gives median =4-27/7 =4.88. median ekccallihatey (x) =O fora 0 =4. Pr(3<.X <9) = Pr(X <9) = i-(4] = 9415. aph ege= i-( > 08 x+5 3 (a) 9 = -(.| (b) Let Y, be the payment with a deductible of2. x+5 Givesco. = 571 ee OL E[Y2] = [ -F@)] dx = hl = =)dx 6-73 X is Pareto type II with @ =3, 8 = 2000. (a) E[X] = an = 1000 (b) nae F(x) _ 2000)". 1-200) 7x >0. ee Thus, ; 2000 wages Pr |A > 1000] =| 1000+2000} 27 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 131 (c) Se) 2000’ | [ eee im ne ae i We= 519.84. == 2000(2"°-1) givesxs5 (d) Let Y be the payment. 3000 [1-F(x)|dr = iF3000/2000)’ bs {, a dx = a G +2000) 3 2000" 2006" (+2000)- 5 {3000 i. (00 ne Lee i -2 | 50002 20002 _or 2000], 1(2000) ||= _ Ho in) 6-74 X is Pareto type I with a =3, B=3. aruba a = =. (bo EG) Pr 3 = 1-(3) ByeeeLIS E(6)=—F (4) X <6|X >4| = Sar 3 (cy = 1-(3) Sives Kh = 321 = 3/8 t 6-75. X is Pareto type I with @ =2.5, 8=200. Thus, F(x) I| | — ~ EES pos) paee 1-(22] gives, X25 = 2004 2/5 = 224,39. 2.5 13 = 1-(22) re gives x75 =200- A'S = 348.92, The difference is 123.83 (A). SN i)S fn) SSS V ~ S Se 132 @ SOLUTIONS TO EXERCISES - CHAPTER 6 6-76 X is Pareto type II with a =3, B=1 and so E[LX] = see = < (C) DS 4p) 6-77 2X 1s.Pareto type Lwith o@ = 2.5, 6 =0.62 Thus, Fix) — 1-(4| 3 he cabs Let Y be the losses not paid, 1.e., losses capped at 2. Then, . 25 E[Y] = 6+ [1-F(x)]de = anf, ieee 6-78 (a) (b) 25 LS dx — | 9343 (C) Pr(X>1) = 1-F(l)) = e2" = 0.1353. FLX] = are me er(i+2] age r(¥] = 70876. Note: The Gamma function can be evaluated in Excel using the built-in function GAMMALN (natural log of the Gamma function). Thus, r{¥) = EXP(GAMMALN(4/3)) = .89298. Section 6.8 Chapter 6 Sample Examination 1. (a) E[B]=100. The mean, median, and midrange all equal 100. There is no mode. (b) Range is 200, JOR = Q;-Q, =150-50=100, Var{B]= 200° , and OR = 57.74. 12 2 Let F denote the number offailures, Pri’ $66) = mz 3.55. 25 ede < 66.5 — 70 ef F ~ N(70,0? = 45.5). |= Pr(Z s$—.52) = 3015. ee ee Tb, In(.75) The third quartile is found by solving .75 = 1—e-*/*476!, Thus, Q3; = —347.61-In(.25) = 481.88. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Ames oO Galerie Se ClO anceiu ad aL 6? -8! 0 inact 78. —cl6 mere; Cag Ld ato @ 133 iy arene this expression equals 1, since integrating a density function 5. The mode of D occurs at the maximum value of f(x) = cx? -(l-x) on 0< x <1. This maximum occurs at x =.75. 6. L=e*, where Yis normal. = Priam Note: It isn’t necessary to find c. Let m be the median. Then, ml = py]zs S| = Pre =F Mie Ol, So, m=e>"!. ie 8. 9. Let N denote the total sales revenue. E[N] = 64-7 and Var|N] = 64-9. The 90" percentile is 448 +1.28-/576 = 478.72. PrO>120) = Pe Z>0 ie 14/./19 = Pr(Z >2.49) = .0064. Let L denote the individual loss covered by the insurance company. We will use the Central Limit Theorem to approximate total losses, S = L, +---+ Lago. 20,000 E[L] = ee (x-5000) - i] 7 555 dx == 56 5625. 20,000 ELL?) = [.., (%-$000)? -5— | dx = 56,250,000. So the standard deviation of a single loss is 4961. S —(200)(5625) Then E[S]=(200)(5625) Os = (200 }(4961) . Thus,fe Z oe es wire § <1,200,000) 00,000 Pr(1,0< <1.07) < Z 78 = Pr(—1. AN = .82. (D) and 134 @ SOLUTIONS TO EXERCISES - CHAPTER 6 10. The loss random variable X has a Pareto type II distribution with 2100 ———_ Thus, (x) = = |-| (in| EY] 4 ae) If Yis the payment then, for x>0. ik L ‘ F(x) |iis 4 _. ® 2100 oe fl ao |a ae (x+2100)3|" 00E 3 a = 468.95. 24003 = 11. @ =4 and £ =2100. 300 Let X and Ybe the two errors and let W = 5 (X4Y), Then gy =0 and of = —-[(.0056h)? +(0044)? ] = ow =.00356A. Then, Pr|W |s.005h] = Pr Z| Osh aa = 2@(1.4)=1 = .8384. (D) CHAPTER 7: MULTIVARIATE DISTRIBUTIONS Section 7.1 Joint Distributions for Discrete Random Variables ae =| a) Pa) a ee tors z = 2 PC7i) 3 00 EY Te ee = d x! y! “ = oe 00 3x71 = 3e 2) ) eam x! . ee Bes — 3” x! . l = e3 x! = 0F))235: 9) ELX |] = ais : fe fs aes a s = ee) 3 3e <i a——== x : 3 y e--e i ek Sh [eesye ap(x,y) = 0?- p(1,0)+1? -p(1,1) + 4?- p(1,4) + 0? - p3,0) +1? - p3,1) +---+ 4? - pS,4) Using marginal probabilities this is = Oe(30)1a (25)44°>C45)1= 7:45 (b) E[J¥ |= -32+ 2 22 =1.253, 136 @ SOLUTIONS TO EXERCISES - CHAPTER 7 7-4 ae vpsari 6apts Sp Sige pela +2. 1 qed te rae te pay 1/4 Pty 7-5 I teatRA By7ndi. eae ees)aay (8mest aero -—+1]?.(g+4)+ Curae 12 Pe 07 Var = “t-(3) eS E 2 gas) (a) and (b) The individual probabilities are determined from the hyper-geometric distribution. ee po G s (he 0, Ces | oe eee See apenas 220 | 220 | 220 123 36 220 48° 220 nOUR 220 ile 220 120 (c) E[C] == 0 0: a0 90 SYN Che ee TI ea) 10 oDs 590 = - et =. 20 0 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT ® 137 Section 7.2 Conditional Distributions — The Discrete Case 7-6 (a) (b) ELX |¥ =3] On LY = yes] EAE) = Ven 0. ee ee cec: Var[X |¥Y =3] = 5.5—(2.2)? = 6173. (dq) ELX] = 1-(037)+2-(.42)+3-(.21) = 1.84. BUX 3221) 1 C37) ees C42) + 3 94: Var|_X ]=.5544. 0(0) + 1(13) + 2(23) + 3(6) = (f) ior , = ge = & E[V¥ 1x=3] = SOND LD a2 C)S A Save) Let T equal total accidents and W unreimbursed accidents. Then Since we are given that_X is zero, the possibilities are: EIEAPAA CANOE S wr] nlp] Nl oO; S| Oo|o|& | Se W = max ie = 2,0]. 138 @ SOLUTIONS TO EXERCISES - CHAPTER 7 Then the distribution for W (given_X is zero) is: 7- EQ ILX =1) 05 = 0: ORES a) pl2e |- Er Si OE She aad SG ile La) EU 3 We have p(m,n2) 1 Ay = +(4| E oe 9 (E) = .741286. Vary (X= 1) = 741286741286" 7-9 ae7 = .741286. a 05 JOE 2 211 = 204" “*(C) l e™(l1-e-™)"@; m,n) =1,2,3,... and we need to calculate E| Np |N = D | Let p;(2) be the marginal probability that N; =2 and let P2(m|2) = Pr| N> =nz|N,= zi = oe Pil4 be the conditional probability function Na for ny 4 dl all pL > -2(J-e-2 ym =| Al Aen 5 \21 2 no Wn | +elea (1-« —-e- ary =1,2,3,.... Then p(2,n2) 3( =].2 INEXt. 571 (2) <= zen) no=1 3 ] A 4 4 Je Ce geometric series with : > ratio l-e~* 1) 4x ee 3} Ze) AED, 2” <= n=! Thus, p2(n2|2) pP(2,m) = ——— Pi 0 (2) jl—-e77)m@-) = oie ia} - 5 —2 +— = e7*(1-e7)@71; 7 Giese«= l ea Ww =1,2,3,... ; = TL: 374 ) . . This is a geometric ‘ distribution in the form “number oftrials to the first success,” where the probability of . 9 ~ success 1s e ~. Therefore, the expected value, B| N2 | = 2 |, equals ——=e?. | aie (E) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 139 Section 7.3 Independence — Discrete Case 7-10 Method 1: We will complete the joint distribution table (using independence) and compute the expected value using all joint probabilities. Be ale 4 ta) 2). 2 Method 2: Since X and Y are independent, marginal probabilities. (5) (3) COTY 15 E[LX-Y] = ELX]-E[Y] We can use the ELX] = (-1)-(.4)+(0)-(.5)+(5)-C)) = .1. BLY] = ()-(.6)+(2)-(3)+)-CD = 15. E(X-Y] = ELX]-E[Y] = (.1)-(1.5) = .15. Section 7.4 Covariance and Correlation 7-11 Cov[X,aX +b] = E| X[aX+6]]-E[X]-£[(aX+5)| = aE[X?]+bELX]-E[X](aE[X]+5) aELX?]+bE[X]-a(ELX]) -bELX] = a(ELX?]-(ELX)) | = a-Var[X]. 7-12 E[F] -+3-4-—+3.5-—+3-6:—-— E[T) (2)-(4) = w|r 140. SOLUTIONS TO EXERCISES - CHAPTER 7 7-13 Var[X]=10(.2)(.8)=1.6. Cov(X,2X) = 2-Var[X] = 3.2. II 25Var[X]+81War[Y]+2(5)\(-9)Cov( 7-14 Var[SX —9Y] = X,Y) = 25(13— 9)+81(20)-90(7 -(-3)(4)) = 10. 7-15. 17,000 = Var[X+Y] = Var|X |+Var|[Y]+ 2Cov(X,Y) = 5000+10,000+2Cov(X,Y) Var{(X +100) +1.1¥] = Cov(xX,Y)=1000. Var X]+1.12Var[Y]+ 2(1)(1.1)Cov(x,y) 5000 +1.21(10,000) +2.2(1000) = 19,300. (© 7-16 8 = Var[X+Y] = Var[X]+Var[Y]+2Cov(X,Y) = (27.4-52) +(51.4-72)+2Cov(X,Y) => Cov(X,Y) = 1.6 1.6 = Cov[X,Y] = ELXY]-ELX]E[Y] = E[XY]-5-7 => E[XY] = 36.6 covariance formula Cov| Cy,Cz |= E[(X + Y)(X +1.2Y)]- E[X +Y]- EX +1.2Y] E(|X?]+2.2E[XY]+1.2E[Y?]-(£[X]+ E[Y])(ELX]+1.2£[Y)) = 274+ (2.2)(36.6) + (1.2)(51.4) —(5 + 7)(5 + (1.2)(7)) = 3.8: (A) 7-17 Cov(X,Y) = 1-0-(.17) +++-+3-3-(.08)=[1-(.37) +2-(.42)+3-(21)]-[1.45] ELXY] te aa SE = 2.8-1.84-1.45 = 132. t 718) Cov3Xx,—1Y |= E121 XY SE Rete —21{ELXY]- ELXJELY]} l| ~21Cow(X,Y). PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 7-19 (a) The joint probability distribution with replacement is: ELXY] = F(U+2+3+24446434649) EY Cov) (b) 4220) The joint probability distribution without replacement is: E[XY] = (24342464346) =F. conx,y) = ee 36 (|(1.2)2)5 30. (a) 6 5) 2 Tae aeWe = 0e Var[X er ]ee (b) Var|Y | == 02 fea B47 =e PEE COVEY.) = 2.=(1.2)(.8) ease (a) Var[X+¥] = E|(X+¥)?|-(ELX+¥]) 10 Alternate solution: Var[X+Y] 10 9 aa3 2 al a3 eis) > — 36 10 = 4-(2)? = 0. = Var|X]+Var[Y]+2Cov(X,Y) = 0. Also, note that X¥ + Y = 2 is aconstant random variable, so the variance 1s zero. @ 141 142 @ SOLUTIONS TO EXERCISES - CHAPTER 7 7-21 The joint distribution for X, Y, and Z is summarized: So ea (a) Cat ra Since the joint probabilities are the products of the marginals, it follows that Y and Y are independent. (b) Var[X+Y+Z] = E[(X¥+¥+zy |-(B[X+¥+z]) = 3-(3) ~ 75 () SEXY Zi m7 () gy =. ea AX]: EIY]-HZ] = ie5 a-5-= (e) They are pair-wise independent but not all three independent. Bae Te Sh 7.07 S 4 85 —(1.95)(.6) ee 2° 2-2 = Prix sl) Pry =1)- Pr ayy 4.35-1.952 -/3.4-.67 7-23 CovLX,Y] =Cov[_xX,-3X] Cov[X,Y] a For example, = -3-Cov[X,X] —3Var[X | \Var(X WarlY], 7 Var x1(9VartX)) = —3Var[X] | =i) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 7-24 Because daily sales for B are more spread out, Var(X) <Var(Y), and Var(X+Y) = Var[X]+Var[Y]+2Cov(X,Y) negative < Var(X)+Var(Y) (E) Section 7.5 Joint Distributions for Continuous Random Variables 7-25 ed a aii “2 dy de Prox Y <1)-= ( Ne 7-26 (ah shaded area total area = 1/2 1 6 ie f(x,y) 1s constant (because joint uniform). The constant equals l areaofregion (a) (b) ee ee 4/3 4° fx(x) = f° fvay ely) rt. ar s a Lem Bysymmetry, = a 3,/l—y = Sys E[X]=0. for -1<x <1, zero otherwise. : for 0< : y<l, zero otherwise. You should do the integration to check this. ia E|Y] = 0 } (11) -40'0? & = +f) y-(l-y)!"* dy = [iy Ir(y) vo |to substitute w=l—y fw? Ho me ie Ou = A. (—du) — # 143 144 SOLUTIONS TO EXERCISES - CHAPTER 7 7-27 The device fails if either component fails. The correct answer choice is (E). There are other correct ways to write this including: 1- ie [ fls.ds dt or lizff(s,t)ds dt + he lief(s,t)ds dt . 7-28 Pri(X+¥<l) = {,[ede dy= fl-ee?] * dy = [[cet+e)] dy = [-ye!-e"] . =l—2er*. (D) y=0 7-29 7-30 The joint density function is constant on the triangular region with vertices (0,0), (0.L) 2 and (L,L), whose area is =. Thus, f(t,,t2) == for Biel (C) p= { f,(e +13) = dt; dt) = 2E 0<4, <b <L. Both components are still functioning in the given triangular region. Mm 40 x=50-y Boundary Line Region of Integration: ~~ X,Y greater than 20 i) 10 20 30) {0 SO 6 30 -50—x pres Thus, the probability is the integral, 125,000 We iF IU)) Gy) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 7-31 Pr(X +Y 21) is found by integrating f(x,y) over the shaded area. (D) ee Pr(X+Y = 1) 1 x? eee ie eee 2 Seen lt17 aA: 0.75 0.5 y-values 0.25 x-values Note that Y takes on values from 0 to 1. g(y) = PY Pony) ax = JP IS yde = 1Syx my oy gly) = Jisy? *1—y") 0 otherwise | 7-33 7-34 No. f(1,2) = 1.3-.8—.8 20-30 for 0<y<! , = —.3. (E) Density functions cannot be negative. = 30%. ey. 50 9 7-35 The area of an ellipse of the form Pr(safe (safe rabbit)it) = (xh) i (4 1-20-30+-4-407] | 40-80 : lat bh? = 3973. 5 = 1 is Area = z-a-b. @ 145 146 ® SOLUTIONS TO EXERCISES - CHAPTER 7 7-36 The joint density is constant on the shaded region whose area is rae = 34 Thus, f(¢,f2)=—— on the shaded region. By symmetry, E\T, + T)] = 2£[T] 2E[T;] 148) 9 6 pl0—1 \45 8 2 | J,jptaran+ [fa asa = A34 (488 7-37 The joint loss lives on the region 0< x<1land 0<y<1. (x+y-—1) over the sub-region where E{max(X + Y -1,0)] = (Est Je 0@) x+ y21. _ = 512540) The policy has a payment of So, {,2x: (rt y=) dedy = + (A) leif(v)6e*%e?” dxdy II Flee” cof 0 dy II he{6ye?” .(-e-? +) dy J,(Sve? 6ye )dy Don’t get scared now. a) IF.2ye*” dy -2 il.3 ye"dy SE this is the expected value of an exponential ry. with mean 1/2 = 3(1/2)-2(1/3) 10 — this is the expected value of an exponential rv. with mean 1/3 Cees: 5 = —=-= ==. +i, 6 | lo (D) x? wh 7-39 ELX = [ fia 7 (10- xy’) dvdx = ali [10x- : Ja = 5.77. (C) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 7-40 The failure time, /, eas A # 147 is the maximum of Y and Y. if SY i ae inital = ake (35)dvde + [wee , (%) dxdy = 2.54+2.5 = 5.0. (D) I(x) f(x,y) if y>x if x>y Section 7.6 Conditional Distributions — The Continuous Case FAV fimyy=44n 0 (a) sets x*+y* <4 OO f otherwise fx (x) = | ‘ th iy ee 4-12 4a ~ 20 for —2 <x <2, zero otherwise. Similarly for fy(y). (Dyess (ilno=1) = * edhe AMESOY,og fx) aa, for=/3 2 y<./3. That is, given Tae et X =1, the conditional distribution for Y is uniform between —J/3 < y <3. 7-42 3 Bie8 = 2691. folx=Day = oe (c) Pr(¥>.8|X=1) = (a) f(x,y) = (b) eer l Given X =<, the conditional distribution for Y is uniform on o.3| : 1 Ib fry |X =x): fx(x) = aaa =i as x= 1. fomQ sve Therefore l | ahaa We nae amr” [x9 PY > An = 7 = vea5 148 @ SOLUTIONS TO EXERCISES - CHAPTER 7 (c) eee We have to find the marginal density function for Y, fy(v)= [= dx =—In(y), foreU=y =< i So fy (1/4) e= In(4). Sa (1s) al) F(i/4) ~ taf)” 4) - =f fx{v1¥= Then, i pesekal a iste ees) Ne ” Pe x>Hy=t atin hoe (d) ~~E(Net Loss] = ELY—Y]= i,[o--y)t PGi {de -+. (e) Since Y given X =3/4 is uniform on ne/4], elyix=3, _ 3/440 _ 3 2 (f) 7-43 8 “Off with her head!” Fy (y) = [fy dx = 4y—3y’, for O< y<1. ELX|Y=y] = [x-fe(xl¥=y)ak A {poe ; 1 (2—x-y % Y) ees es 4y—3y? Sy(y) for 0's yal. So, for example, HE]X |Y= l= sdk 2.2 Ce = 7-44 fr() = ibS(x,y)dk = hes a= Lye. 25—2y = , 4-—3y = 1622. 2 é 2s tor 0< pao E[X2|Y=y]= lice fr(x| ya = [ox ev/y y e = Alternatively, note that f(x,y) 1s a function of Y alone. This shows that f(x|Y = y) is uniform on the interval [Oy] , Hence, E[X?|Y=y] = Var[X|Y=y]+£[X|Y¥=y] Since the joint density function - eal aba is constant on the region bounded os. by y=0 and y=4-x’, we have that, given X =1, Y is uniformly distributed on 0< y < 4-1? =3., 3-0)? Lheny |Var|y 1 |x =| ce ca) - 3 >, and so oyj\y-(oh .866. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 7-46 # 149 The event that the device fails in the first hour is represented by the L-shaped region ys lf (Ose SOs The complement to this region is the square (1SxS 2))(iy 2), Therefore, the probability that the device DOES fail is ff 1A es — dxdy = 1- fp Gy = 625" (D) (Ole ake tr (y) = [24x = y +2 Bo) tor Gaye eo. yeesO f(x | y= Ly) yor Sy (y) lay Oye (E) Section 7.7 Independence and Covariance in the Continuous Case 7-48 By independence, we have f(x,vy)=e ~* <y on the rectangle [0<x<o)x[0<y <3]. We need to determine the region in this domain for which John and Mary meet. In order to meet we must have Y < X < Y+1( see diagram). Pr(shaded region) |= SF a |t — Se Q.a S [-e'(v+0] =5(I-e") Se ey = <(-e(I-4e") = 1125 using integration by parts x X denotes John’s arrival time, Y denotes Mary’s arrival time. 150 SOLUTIONS TO EXERCISES - CHAPTER 7 TAQ (ay LD1 ofc) =e 2 Porereniy ee Sa at cy 2 a ar sees 0 otherwise (6) fr(y)= [)fy) ar=1, for O< ysl. But we already knew that Y was uniformly distributed on the interval [0,1]. () 7-50 £E| VX -¥3] = f,foe pha = 6928. The region where the joint density lives looks like: The shape of the region makes is easier to conduct all integrals in the form i [C --) dydx . EX epee ee |=SEA 3 I,I. Sree X-V dyde = 3 i. 5% Vee as Meee 5 = 800, (eo EY Bed mall I xy* Se dyaxhe = Gel _ Cov[ X,Y] 8 pl ex. 3 j,i) Ayo 5 a3 iheeee OF1.244 Bats ri fica Oa Cal ieckaeaeta Se dvax = a liea= =1.037 —(0.800): (1.244) = .042.. a oT 7 1.037, (A) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 151 3 7-51 Weare asked to find Pr(1< ¥Y<3| ¥ =2) = I1: f(v|X =2) dy. fe2)= "fd = [Syed = 4 : yk 2) ibaa NE Vey Fx (2) Se | etry =|e 3 * Pr(l<¥<3|X=2) = f'2y3 dy = =. (E) 7-52 Given X =1, the conditional distribution is uniform for 0 < y <1. EY? |X=1]= [\y?- QIK =Ddy = fy? -ldy = +. 2 1 3) 1 3 I for 7-53 (a) f(xy)=4n 0 o: x- vege 9 ») ye < ro otherwise 2 ©) b Prl(X24¥?) >.5]=1- 203) =1205) _ (ei Pix<Y] =>. @ fe@=| se =~ fal j2 Ly (e) ? Ree 1 : The random variables are dependent since the joint density does not live ona rectangle. 152 @ SOLUTIONS TO EXERCISES - CHAPTER 7 7-54 Var[Z]| = Var[3X-Y-5] = 3? .Var[X]+(-1)?Var[Y]+ 2(3)(-l)Cov(X, Y) = 911) +1(2)-6(0) = 11. 7-55 (D) LetX be the time to the next Basic claim and let Y be the time until the next Deluxe claim. The joint density is given by f(x,y) = eot!2 ev/3 Des . The event Y< X has probability given by, = x2, PY <X) = i = —x/2 i} € : = TEN Ne 1 ¢l ew 5 bs (1—e-*3) ad = Hie \ [kx aeay mer Sie —¢- 6) dy a 2 (C) : implies k =2. ELXY] = {,[22x dx dy = 7 EL = {,[x 2x de dy = 2. 1 el Ele If[jy 2x de dy = > Covey) = s-25 S70, We could also observe that X and Y are independent because f(x,y) can be written of a function of x alone times a function of y alone, on a rectangle. Since independent, then Cove.),) = 0. (3) 7-57 VarlY|x] = E[Y? |x]-(ELY | x1) Ix (x) = [Fe of Ya tex ees x)dy = Finede fy mde f(x,y) is ax Ix (x) 2x == |, That is, given a particular value of x, the conditional distribution for Y is uniformly distributed on the interval [x,x. +1]. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT x+l SY see) 2 EL Pete] = (oy erence 3 eetnioers = 3 and ais E[Y |x] = [ l ydy — (x4)? =x? Z Vanhy \x= ely 2 3x? 43x41 =|aes al 4 ” se We could also have recognized this immediately since the variance of a uniform random variable on an interval of length lis 1/12. . Ese (A) l—x f(x) = Is f(x,y)dy = i, 24 xvdy = 12x(I-x)?, Jee ae i yO)= 9 r{ . WV ~~ Ix (x) l 2y= t) ee. = 24xya SS ay=aEe SoS Vege egy = Oy 2 ==>, O< 0<x<1. (l-x)* 4 <a ay 2 priy<ty=4] bl promt meadj(eenan Pr <4 sl ioe: x=4] = nein]i 7-59 (C) The joint density lives on the region shown, with iver aex) fy(xle lo b Of xsl ys ysrtl Since the joint density is uniform, we can use simple geometry to complete the problem. The event that Y is greater than 0.5 is the complement to the triangle with vertices (0,0), (0,1/2) and (1/2,1/2), whose area is 1/8. Hence, Pr| ¥ >.5|=1-1/8=7/8. (D) Xx # 153 154 ® SOLUTIONS TO EXERCISES - CHAPTER 7 7-60 F(x) = [,30-9? dt = 1-(1-x)3: O<x<I1. In this instance F(x)>x and |x—F(x)| = F(x)-x = (1-x)-(I-x) and G = 2[ [(1-x)-(-x)' Jar = 4-4 =5 Section 7.9 (©) Bivariate Normal Distributions 7-61 Var[X|¥ =y] = 02 (-p?) = (32)-.84 = 7.56. oyy-6 = V7.56 =2.75. 7-62 Given Y=-2, X is normally distributed with Pep 3 = 8+=(-3)(-2-(3)) = 7.82 Var[|X |Y =-2] = 9(.91) = 8.19. Pr(X <9Y¥ =-2)=Pr(Z < O78) ) =Pr(Z <.41) =.659. V8.19 7-63 (a) ox =1, uy =0, oy =2, wy =—1, and 1— p* =.36. So p=+.8. We know that p=.8 by looking at the sign of -1.6x et5 } \ 2 (b) Pr(¥ >.5)=Pr(Z > oS: ONS PNZ > 5) = 3085, 7-64 Let A~ N(w=10,000,o = 2,000) be the random claim amount from company A, if there is aclaim. Let B~ N(u=9,000,o = 2,000) be the random claim amount from company B, if there is a claim. If X = A-B, then E[X]=E[A]— E[B]=1,000 and Var(X] = Var[A]+Var[B] = 8,000,000 so that oy =2,828.43. 01,000 ) : Pr(X/ <0) = Pr(A<B) r( ) = Pr} i Z < <————_ 7898 43 | = Pr(Z ( <-.35) 5) = .362 6 given claims from A and B, Pr( d<B) ——_,, PrA<By=le G60)-(30) ————_ "= 33625 oa) G0) a probability no claim from probability claim from A A and claim from B and a claim from B = .2234. (D) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 155 Section 7.10 Moment Generating Function for a Joint Distribution 7-65 X = J+K+L. Since the losses are independent, Mx(t) = Mj()-Mx(@)-Mz() = (1-22)? .-28)?° - (1-28)49 = (1-24)-” M(t) = 2001-20)", M¥(t) =4400-20)-!2, and M%(t) = 10,560(1-21)-33. E(X3) =M% (0) =10,560. (E) defn 7-66 Myz(ttr) aa = Ele” +22| 2 El et(x+Y+n(¥—x) | E| eft +t.)X+(-t; YY | = EI ett +12) X o(ta—h yy | Since U and V independent SE ge aa II es Lei +f x |B ele yy| (t) +t)? = My(tit+h)-My(m-4) =e 2 (t2-t)? -e 2 = ef th’, (E) Section 7.11 Chapter 7 Sample Examination 1. 0 = E|(X-2)3] = ELX3-2X?+4x-8] = 9-2E[X7]+8-8> Gh + AP Zz. (a) ey TAG. = 3-4 = +. The area ofthe triangle is 8, so ; I(%y) = {t for lo otherwise O0<Y <2 and 2y-4<x<4-2y see Wy (b) By symmetry E[X]=0 and E[Y = [ ike g ddy = 2 (c) The event [Y >1] is a triangle (see drawing in text) of area 2. The subset consisting of [¥ <1/2][Y >1] has as its complement in [Y > 1] the triangle with vertices (1/2,1), ( 1/2.7/4) and (2,1). Its area is 9/16 and so the area of [X<1/2}A[Y >1] is 2—9/16 = 23/16. a) 73 Thus, PrL¥ <1/2|¥ >] = zone = = = 71875. (d) By symmetry, E[X |¥ =.3] = 0. 156 ® SOLUTIONS TO EXERCISES - CHAPTER 7 3. tt for Osa Sh 0 otherwise and fy(y) = yt+.5 for0<y<l 0 otherwise (a) fx(X) = b) ALX]=21Y]=5. (c) X and Y are dependent. The region is a simple rectangle, but f(x,y) cannot be written as a product of a function of x alone and a function of y alone. Or realize, for example, that f(1,1) = 2 ~ €.5)-(:5) = fy@:- AO. x+y (dq) f(x|y)=—— yt+.5 for0<x,y<l. (e) ee 9 EIX|¥=.7] = =... f Pay (f) E[7X -3Y]=<. ‘gap gia we1/144Re oe a naa (h) Varl7X -3Y]==>. a 1 rela) (b) 2 (c) ay 8 Sa (lee el7 so.and ee) y= 2s 33. (e) X and Y are dependent. p(x| y=3) (g) (h) (i) Gj) (k) (1) E[X|¥=2] = 19/11. E[_XY]=3.7575 =124/33. p= —.03108. V[X]=.19835and V[Y]=.6336. Cov(X,Y) =—4/363. Var[7X —3Y]=15.88. ae ee] PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT (A) rE | (b) ey # 157 a47 = 92EXE BY 4 bz pas: Var[2X-Y +4Z] = 4Var[X]+(-1)?Var[Y]+16Var[Z] —4Cov(X,Y) +16Cov(X,Z) —8Cov(Y,Z) = 4-(2)+1-(4)+16-(5)—4+16-(-1) -8(2) =58. (.5) Se Cover 2) time Covix4) OneION: 25, Cov(X,Y)=-6. mis on : ) Cov(X,Y)=82/3and p =2.768. Since the correlation coefficient must be between negative one and one, this made up problem can not exist. From the statement of the problem, rts)=|2ix+y +9) 0 for O45 0< Y< X¥ <1. Thus, leand 075 otherwise. fr) = [fe yav = [204+4 hen =) = jy) = 3x2 elon 2 21+ y) fx (1) 03 ~ 05 2 my 05 2G jpe Ye | = \, fr(y|x=.)dy = | cia Then Pr(¥ <.05 ¥ =.10) > ~ fy 6.1) = 03. 416. (D) The key step is to note that, = on the triangle 0< y<x and 0<x<12. whee y)= fr(y|x)- a) X Since X is uniform on [0,12], ELX]=6. You should do this without calculation since X ~Unif (0,12). Vp fi[,v Sl.) dvdk = ie[= ee: dydx = (45a oe = eh =[ j; a dyd. 12° =" >5 dydx= (E x de==e Cov(X,Y) = ELXY]-E[X]-ElY] = 24-63 =6. (©) 24.. 158 @ SOLUTIONS TO EXERCISES - CHAPTER 7 10. Let O denote the number of tornadoes in county Q. Let P denote the number of tornadoes in county P. First, Pr[P=0] = .12+.06+.05+.02 = .25. Then, jy Ve ee Se ee it Pry 22) eS 09, 32 SS (.88) =. 980. (D) 6 i hk (l-x-y) dydx _ 2 (l-x-y)? = 6[/ere l-x x dx _ 3 [ex de = - (ex) Bri 12. Let let 8a= 488 G~ Exp(A =6) denote the random wait time for a good driver to file a claim and B ~ Exp(A =3) denote the random wait time for a bad driver to file a claim. PG <3( B= 2) =P nG = 3) Pile2) I (een?) (le) (=e). DB syle"? 13. (C) d-e e's 7?) e-7 (C) Let P=premiums ~ Exp(A =2) and C=claims ~ Exp(A=1). joint density function is f(p,c)=+e-?’7e* for p>0,c>0. By independence, the We define X == We want to find the density function for this ratio. We will use the standard technique of finding the CDF first, and then taking its derivative. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 159 FGe)e =: Prea) = Pr S = | Rite < ae _ i, WP =f = ere i> —p(x+1/2 ))dp —ePl* (e-P/2 = pie (l-epsa ey ee ~plo 4 I [ee) —p(x+l/2) p=0 SS IS eecie 4 2 (x41? Th ell. be) ‘@) =F F(x) @)=———. The CDF for an exponential distribution with parameter 8 is F(x)=1—e~/. Then, Ene8) 0 0+5-[e% — e895] 476 [e48 _¢-#i| = 71223 (0) peeks ———— Sa N=0 Let 7; N=l N>I =X; + Y; denote the total number of hours watching movies or sports of the i" randomly selected person. E[T;]=50+20=70 and Var[T;] =Var{X;+Y;] = 50+30+2-10 =100. Let T = 7, +72 +-:-+Tjo9. Since the people are chose randomly, it implies that 7; are T; are pair-wise independent, for all i # /. E|T | = E{T; +T) +---+Tigo | =) 7,000) Var{T] = Var[T, +T2+---+Tioo] = Var[T]+---Var[Tioo] = 10,000, so that Or = 100. Pr(T <7100) = pr|Z< 7100 — 7000 00 = Pr Zi< |) = S413: (B) 160 16. SOLUTIONS TO EXERCISES - CHAPTER 7 Sketch the region where the instrument will fail. X fails in first hour Y fails in first hour Pr(fails during first hour) = 1—Pr(both instrument work after 1 hour) -= x+y gohka dx dywee= > = 407 (B) AI 17. Again, the key step is to find the region for which the total payment is less than 5. Since we have ajoint uniform distribution on 0 < x <10,0< y<10, we can calculate probabilities using the ratio of the area of the event divided by the area of the square. Region Ee where t al be efit is less t an 5. Pr(event) 2:61 = 10 Sess 5 ae = ; 529), (G) PROBABILITY AND STATISTICS WITH APPLICATIONS: 18. A PROBLEM SOLVING TEXT @ 161 Since the joint density is f(x, y)=2e-¢*?”) =(e*)(2e*”) on the rectangular region [0,00) x[0,00), it follows that XYand Y are independent. Thus, the value of X is immaterial and we need only consider Var[Y|Y>3]. If W =(Y —3)|Y>3, then by the memory-less property of the exponential, W is also exponential with a mean of 1/2 and a variance of 1/4. Then, since Y| Y >3=W +3, we have Ee Yi3) e735. o-and VarYY > 3) Var[W +3] Var[W] = 5” = 25, (A) cae atn aif Sao fr9rn Taw iy m Ose if CHAPTER 8: A PROBABILITY POTPOURRI Section 8-1 The Distribution of a Transformed Random Variable 8-1 For l<y<e, Iny2=0. [a dy = (Iny)? "= So ei 12-0? = 1, y=l See ror 0 =x <1, Fy (x) -=—> =x. = l<yse. Since Y=e*, the random variable Y lives on Fy(y) = Pr(¥<y) = Pr(e* <y) = Pr(X <In(y)) = In(y). (a) $0) =O) =— for lay e. (b) Method 1: z[Y] = ["y-tay = [~ e-1. 1 y Method 2: E[Y] = a Mere. Deer Ore. 0. 7(x)= ~ooince ) = ase. the random variable Y lives on 0< y<o, Fy(v) =Pr(¥ <y)=Pr(JX <y)=Pr(X < y2)=1-e (a) f(y)=Fi(y)=2ye” for (b) Method 2: The key to computing the value of the integral of the following type is to use Subsection 6.5.2, property (6) of the Gamma distribution with parameters = (c) SA a =1.5 and #=1. 1 \ Jx -e7*dx = Also, r[$|-vr 2/5 j,x22. a ody = 3 r(3| = X ~exp(l), E[Y7]=E[X]=1. Since Y2=X, and Ore.) 0< y<om. Y fy=pen" = nC) I ‘2 ] r(4) = aie ve Z — Aa x = ce" = 1 and a =e. ay (a) fyrQy) = fx (x) for y<o. OS (by eel | = [yee / a — 1 ete. 6d Pe rs /4\e |= ae a mee oe? Se4 mk Ody = 1.3409 (Using a graphing calculator). 163 164 SOLUTIONS TO EXERCISES - CHAPTER 8 8-5 CDF Method: Fy(yy Fy(x)=1l-e™~*, so = Pry) = Pri0- ee <-y) 1/8 SS Tt 1/8 a Pree ea Fy V (2 a = 1/8 / (yoy’® _ (y/10)'> 1.25 _(v/10 Saf (a l-e a 1.25 Foyt Ce 1352 eNO = 9.125(0.1y)25e OY)”. (B) Transformation Formula: y=10x8 = x=(1p)'” and & =1.25(.1y)” (1) =.125(.1y)". Then, -125(1y) 125). ly)” eth)” (B) LO) = feo} eS We sofia) Wee ee S for 1<x<5. Let the random variable Y = a (a) Fy(y) = PHY sy) = Pr[ps y}= [x24] =1-2 lore (b) te see aes 2 ye: fora ves: =p fO)=F ve =| () EY) = | y- = ~In(.2)/4 = 4026. > 8-7 (diem ely | = (é)° Noand no, | yd 2£[X|=3 and Var x |= For X, fx@)=51 <x<4. x=Iny and ae = i dy Fry) = fx (x): eee y “ = Atle oy melon (5-1)? 12 The random variable Y lives on Eaak with Then, -_ r ries - <a Te ay ee. iiss. ne So, fy (8) Q) = 5 Loe) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 165 8-8 Fy(y) = Pr(¥<y) = PXT2<y) = Pr(-Jy << Jy] = eipeirs ip) = SG 2 a Oj One =e for y>4. (A) Alternatively, Sr(t) = F'(t)=8t%; t>2. ) 8-9 at The transformation y=¢? for i) gid ee = 1 . F(v) = PrV<v) = Pr(10,000e% <v) ¢>2 means t = ys Then Pi) = fr =8(y"} ao =4y™. (A) = Pr| R<In—— r In--4, 10,000 —.04 .08 —.04 _ 8-10 ee (E) 10,000 R= ai customery eal customers per minute. T minutes dK Ferry = Po 10 Sr) = tPri <r di : ; # fir)=Fn=2>. oe 8-11 mad Pri T = ; ‘ab gerald = r ae== - (E) X is the profit of company I. Y =2X fr(v)= SOE [ fsa : is the profit of company II. l Since Y=2.X, we have ay: ebcme Sl and ios Then ivalently in terms ofx,fr(x)=+-f|oad Lx]. (A) fei V— hh Gi oe or equivalentlyin terms of x, fy (x)= pa 166 ® SOLUTIONS TO EXERCISES - CHAPTER 8 fs(s)=Ae~** and fr(t)=Ae™. 8-12 f(x) = [fs(s)-fre—s) ds = (Ap): [er 2 A) ds = A? ee 8-13 x ee -e ds = A*.x-e, The joint density function is f(s,t)=¢ on the rectangle [0,2] by [0,1]. t-axis 0 l 2 S-aXiS There are 3.cases, O< x <1, 1<x<2,and 2<x<3. The event [S+T <x] hasa different shape, and hence different limits of integration for each of the 3 dashed lines in the diagram correspond to the boundary line s +¢ = x for Casel: O<x<1, Fy(x)=Pr[S+7 <x]= igi 3tds dt = Case2: 1<x<2, Fy(x)=Pr[S+T <x]=(le tds dt = aa Case 3: 2<x<3, (integrate over the triangle in the upper right corner to obtain the complement of F'y (x)) 1—Fy(x) = La p2 [se ares dt = Fy (x) = Finally, differentiating, = who = —_a ; i)— | : A(x 2)+ Ltx-2)°. —- | S) ~~" PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 167 Note: This can also be solved using the convolution method for each of the 3 regions, but care must be taken to choose the correct interval of integration for s: Tx (x) [, BO): fries) ds for O2xs1 [_ fs(9): fr(x-s) ds for lex [_ fs) frlx-s) ds for 2<x<3 Bel iP=m-2(x-s) ds 4 [Olen [os 2(x—s) dseior 2 es <%= 2 za 2d X—S)-AS = 10F = 2 oS IL: Z = = fore Ose <1 Le 5 [Olea ce< AEG eae 8-14 (a) (Oj fx (x)=e** -e™* for x20. Fl Xe co ifie (oe \ ar (See Example 8.1-9.) 3. Of course, E[X] = LS +7) = 2/8 |+ £7] = 241 =:3. 8-15 Let J, ~ Exp(A =10) be the time to failure ofthe first electric generator. Let T) ~ Exp(A =10) be the time to failure ofthe first electric generator. Var[T, + T,] = Var[T,]+Var[T,]+2Cov[T,,T] = 200. (E) =0) Alternatively, Let_Y be the time until failure on the first generator and Ythe time until failure of the Ve eter ye then § =-1(2,10), $0 Var|S]=28* = 200, (BE) 168 ® SOLUTIONS TO EXERCISES - CHAPTER 8 8-16 If S denotes the sum of the roll of two fair dice. io 51362] =e15]sb TOR The random numbers .28918, .69578, .88231, .33276, and .70997 correspond to sums, S, of 6, 8, 10, 6, and 8 respectively. 8-17 Suppose X represents the outcome of the roll of a single loaded die. Pr(S) | Cumulative Probability 1/441 4/44] 1/441 =.00227 5/441=.01134 10/441 15/44 =.03401 35 / 441 =.07937 70/441 =.15873 126/441 =.28571 35/441 56/441 70/441 76/441 196/441 =.44444 272/441 =.61678 36/441 441/441 =1.0000 73/441 345/441 =.78231 60/441 | 405/441=.91837 The “random” numbers .28918, .69578, .88231, .33276, and .70997 correspond to simulated outcomes 8, 10, 10, 8, and 10. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 8-18 X ~ Uniform[—5,25]. Prix) # 169 We want to simulate outcomes x such that te () aeee 25-(-5) ~ 30 = de solving for Xyex =.30u —5) The random numbers .79936, .56865, .05859, .90106, and .31595 correspond to simulated outcomes 18.9808, 12.0595, -3.2423, 22.0318, and 4.4785. 8-19 B~ Binomial(n=10, p =.7). LB | Pr(B)__| Cumulative Probability ON]15:9049E-06) NG 5.9049E-06 5 | |6 |0.20012095 |__—0.35038928 |8 |0.23347444 | 0.85069165 |9 [012106082 | _—0.97175248_ = The random numbers .79936, .56865, .05859, .90106, and .31595 correspond to simulated outcomes 8, 7, 5, 9, and 6. 8-21 By the inverse transformation method, w= (6) 3.756 .690 ZZ 998 BM: (b) {—In(l care 6.248 2.083 14.39 2.093 The mean ofthe simulated outcomes Is fy = 5.2034 and the sample standard deviation is s, =3.939. 170 SOLUTIONS TO EXERCISES - CHAPTER 8 (c) The mean and variance of the Weibull distribution are given in Subsection 6.7.3: With parameters 6 =2 and a =.03, ELX] = par(is aL) = (.03)-"2 (3/2) = (.03)-/2(1/2)F(0/2) = (.03)-!2(1/2)Jz = 5.12 Var[X] = pl r(u+2}-r(: A | - «asy'|ra-r(3) |= 70 ear PAG, 8-22 (a) c=12. This is a Beta distribution with parameters ~@=2and £=3. (b) F(x) =6x? —8x3 +3x4. (c) Erie (d) (ec) Fora Beta distribution with p parameters Lx Se =.40 and oy = @=2 and fB=3, PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 8-23 @ 171 Here we had two different assumptions about Bert’s age of death. Both models (uniform and exponential) imply that Bert would live an average of 20 more years. We do not know when Bert will actually die, so we simulated his age at death. Note: Under the uniform assumption the future lifetime is uniform on [0,40]. Simulated years Simulated years Given “random”| numbers until death — Present value until death — Present value uniform assumption of $100,000 if die at age exponential assumption of $100,000 if die at age 16.8167] 36458.25 ae ee! x=-20In(I-u) | 100,000e~%* 100,000e~°* x=40-u 1.207531 46847.15 12.638 boa et Pees ) Sa] 9 95.2,946:72 a (d) Bice 93011.05 Under the uniform assumption, _ oer ORS PV (ea= 100,000 |. Clay Pataa 00 e7 06x 140 531.60). Under the exponential assumption, e Le = W0e dx ly. _ 5000 [-e! dx a PV &= 100,000 |eGo e (soe R245 (a) Lets = mint 5000 x. ,y-): F(t) = Pr(min(X,Y) <2) =! - es (b) P PriminCx,Y) <3} = FG) 10-3 10 (c) &{[min(X,Y)] = = 1- 10, 10-f 10 iF t- 50 nies mae : oe (d) E[min(X,Y)*] = j, t 50 ae at = o0 ra K So Omin( X,Y) — 2.397. 172 SOLUTIONS TO EXERCISES - CHAPTER 8 8-25 (a) SuncesY=2X, YS Xvand max(xX,Y)=yY. Primax(X,Y)>7)=Pr(Y >7)= ee= 65%, (b) min(X,20—Y)=min(XY,20—2X). This number is always < = =6.6. Therefore, Pr(min(X,20-—Y)>7)=0. 26543) we Get 7 minx.) 3Z,). We 30en: FO =~ 000 E[min(X.Y,Z)) = foe 3 at 10 |10-18 (10-14 ~ 1000 3 12 es ayy 4 ei using integration by parts 10, i 300-2) YZ) | 1 == [ (b)byeewminx, E[min(x,¥,Z? a __3 |_ 20-93 1000 _ 200-9 3 =10 2010-9) | iz 60 | t=0 using integration by parts 10m Then, (c) Omin(x.Y.Z)— Vat) = 2. 1.93649. _ -\n-l Let T=min| X,...,Xio |. Then f= E|T] = 1 © n(lOmt)"= i, area O<r<l0. *: 10 eo hal 107 tO =9)* eee Be n n(n+1) 10” using integration by parts a 10 n(n+l) nl PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 10 10-1)" ; pein |ef Urs £08 n 10 2t(10-1)""! e 200—7 n(n+l) n(n+l)(n+2) ‘=0 using integration by parts ke 10”) a eeAlte n(ntl)(n+2) — (n+1)(n+2) Then, 2-10 2 holt D é (n+1)(n+2) (n+l)? VariT] = - 2 -]- (n+l)|n+2 107 n+l] (n4l)2(n+2) n=) Next, let S=max| 10 EIS] X,...,Xi0 |. Then fs (t)= ae Os ns?! A es tO" 10n = -——dt = gers! care : eye = men ee 0 10” 10”. (n+2) n+2 n—| n+2 2 (These integrals don’t require integration by parts!) Then, Var{S] ae = 107n ~=107n? _ n+2 (n+1)? 3 = 10°n ] n = = |= n+2 (n+l) 107n 5 (n+1)*(n+2) Note that S and T have the same variance, which might have been anticipated from the symmetry between the two random variables with an underlying uniform distribution. 9) 8-27 (a) E[max(X,Y,Z)] Ke (b) E| max(X1,X2 = Sate ae .—Q 175 dt |= X Xa) )| ai —e| 3 Zs m See = iy =_=-- 5, (c) Pr| min(X,Y,Z)<3] = |—Pr[min(X,Y,Z) >3] 1=—Pr[X 23(\¥ 23(|223] = ]—PrLX >3]-Pr[Y¥ >3]-Pr[Z >3] = 1-(2] Sa # 173 174 ® SOLUTIONS TO EXERCISES - CHAPTER 8 8-28 Let X ~ N(70,37) and let “success” be the case that XY > 78. The probability of success is given by p = Pr[X >78] = plz> 78-70 =2.61 = 1—.9962. Let n be the number of trials. Then, 60" = Primax( 44,495", 4,) > 76) Pr[at least one success] = 1—(l—p)” THT O20 In 0.4 Soeee 0.006) = 1—(.9962)”. = 240.67, son=241. Ge ct(10-—t)dt == c | Gl [0r-1 2 |adte== c 5x pate Ey (x) Se ik = i a 1000: Then, 1= Fy (10) =c| 500- “36 Psy é =10: > and so c =.006 and Fy (x) = 0.03x? —0.002x?. Fy(y)= 453 0<y<10, and F7(z) = 1-e)?; z>0. (a) Pr[max(X,Y,Z)>6] = 1-Pr[max(X,Y,Z)<6| = 1-Pr{LX¥ <6]o[¥ <6] o[Z <6}! = 1—PrLX <6]-Pr[Y <6]-Pr[Z <6] = 1—|(.03\(62) -(.002)(6°) |Sha -e6/5} —[.648]-[.6]-[.6988] = 1-.2717 = .7283 (b) Pr[min(XY,Y,Z) <3] I| 1—Prl| min(Cee ut ese & Z)>3]| data ies 3] 1—Pr[X 23]-PrlY 251.2 1[Z >3] = ae a) 1-|1-((,03)(3%) -(.002)¢ ERA en] BS) = 1-784} = 10 aa i 5-3/5 |5488] -= 1.3012 = .6988. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 8-30 # 175 Let Wbe the maximum of the three future lifetimes. Then, Fir(w)=2 3 and EW] = [30 es 4 Jem2 0a 3 = +30 is. 8-31 Let Wbe the maximum of the three exponential random variables. Then, Fy (w) = [1-e-#3]-[1-e-4 ].[1-e-5 ] II 1-ew3 Jw (w) = Fy(w)= peut —e ws Sr ee wil2 = e 8w/l5 ai e 2 w/20 — e 47/60 sews tet ae fi 12° —Tw/12i ews 8 i —8w/15 y, eee eee 20 ~ —9w/20 47 — 60 —47 w/60 Note that each term in this sum 1s the density function for an exponential random Xx: variable. If XYis exponential with parameter # then f(x) = oe , and E| Xx, = 26-8 Now. es Variexd |= p and EX Ei be Vor 1a Le oe. eed i.Wwhale id wie ee iedw+e-+ i.why (w)Re dw i = j, a0 ° an = 34445 DB _ 9 8. VAT dw = 7.4651. Similarly, 2 E|W|:= 47 —47 ti w/ Wied dy , ee aie! we2 ——e we ae —w/33 dwt--+ I, \eew+ fy (w) dw = j, Sa mei rs | a 20-60 Pieilig ate? (2) (= ye = (2)(40.2370) = 80.474. Thus, Var[W] = E[W?]-E[WP? = 24.7465, Note: The CDF for the minimum VofX, Y, and Z is given by Fy(v) = 1-[1-Fr(v) [I-A 1-20) 1-[e-” ak [e-” cit Le calee eet Consequently, the minimum V is an exponential random variable, with par This would make the calculation of the mean and variance of V much easier. 60 Ay” 176 SOLUTIONS TO EXERCISES - CHAPTER 8 8-32 Let V=min[ X,Y]. Then the CDF is Fy(v) = 1-[1-FyQ)]1Fy ()] = 1-[e? ]-[e-r"’7] = 1-e 25; v0. Thus V is exponential with £ = =. (ayn (vy = Te 2) and fu asses v>0 (by) Pry s3i (c) py “By (d) Lear oy =e 1— et) = 164248) 28) Let W = max [x ; ia . Then the CDF and density are, (6) se ey (ie) = [l= a2] [Fe PT] =-1 ee os Peony Vell aeBese 35° e 12were belw/5 Ga+ze oe = tw lw) and 5 ee = i573 (fy GC) Meee Ole) (g) EW] = J,whi(ov)dw Se w—e 5 = 0 Bia th) 5 dw+ |aN w—e™ry 0 7 dw owt |oe w—eo-’-? 0 35 dy 45 ern O84 72 oe FEW?) = Ihew fw (w) dw = OO we] Mee cure ay ee } Wore e wi? dw+ I wze = aysPe eo, 04 Reet bs wil dy — j, wr a wI/S el2w/35 dw -(3) | = (2)(65.4931)= 130.986. Var| W|= E\W? |= E[w] = 48.4792 and ow =J/48.4792 = 6.9627. E[max(T,2)] = 2-(Pr(T-<2))+ [°r-—-e No = 2-(l=e-*\+2-¢ PROBABILITY AND STATISTICS WITH APPLICATIONS: 8-34 @ 177 Pr(max(storm, fire,theft) >3) = 1—Pr(max(storm, fire, theft) <3) II 1—Pr(storm < 3) - Pr( fire < 3) - Pr(theft <3) fe" 8-35 A PROBLEM SOLVING TEXT !") —e "5. (1)—e 3/24) = 414. (E) The pay-off time for the first product is PR, = X. The pay-off time for product 2 is P; = max(X,Y). ETR]=ELX]=20 and E[P)) = +40 _ Now, pee Xe Oey 7 ote aX = Y, Thus, P, =X on the region 0< y< 40, y< x< 40 and P, =Y on the region 0< y<40,0<x<y. Thus, 8-36 40 -40 40 EUR -Py|= i. il Ta * drdy + | I aaan wtedy = a+ = 600. Cov(B,P:) = E(B -P)]—-ELB]- EL] = 600—-(20)- (2.40]= 66.6. (C) cet) = =[F(y)}4 Pr(max {%, ¥>,¥3, Ys} < y) Uae (y) = Eas (y) a nl. [itsinay m-cosmy-(1+sin wry) 4 E{max|= an [ieycoszy(1+sin zy)> dy (E) Let_X greater than 50 be a success. Letp be the probability of asuccess and let S'be the number of successes in 10 trials. Then, p= Prix > 20)= 1 Fy (50)= Ce eh] BO The second smallest claim is greater than 50 if and only ifthere are at least 9 successes. The probability that the 2" smallest claim is greater than 50 = Pr[S =9,10] = p'°+10p2q = 3152. (B) 178 ® SOLUTIONS TO EXERCISES - CHAPTER 8 8-38 Let Y less than 1/3 be a success. Then p=1/3 is the probability of success. the number of successes in 3 trials. Then Y) <1/3 if and only if S =2 Let S be or 3. Rentoe Pr[S =2,3] = p?+3pq = (+) +35) ee: (2) = 2593. (D) 8-39 700-300 = 0154. Lens ve Let X be aclaim amount and let p = PrLX¥ > 300] = 700-50 the number of claims that exceed 300 in three trials. Then Pr[S <2] = 1—Pr[S =3] = 1- p? = .7670. (E) 8-40 Let Xbe uniform on [0,10]. Then the expected value of the 4" order statistic out of a sample of size 5 is given by wd = 6.67. (B) Alternatively, the density function for Y4 is, S05) i/peut ead js eevivieeryahb: f 1 at arte ~ 105 [do-y)], and ElYs] = 20“~ [“poy*-y5]ay = 2 (B) 0 10° 8-41 Fora single observation Y, Fy(y)=l-—e7” and p=Pr[Y >1]=e7! of success. Pr[Y5 >1]=Pr[at least one success] = 1—qg? = 1—(1—e7'!)> = .8991. (E) 2 se fetta eT 8.2 The Moment-Gencrating Function Method n =AD = n y i=l | Qa; = | Q, +Q2+---+Qy _ ot M(t) = [[“« (t) = I Je 2 och ; |) j= 1 8-43 . = ev : This is the MGF of a normally distributed random variable with mean is the probability n n f= i=l "1; and variance }'o? . Y" PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 179 8.3. Covariance Formulas 8-44 CovfU,V] = Cov[2X-Y,-X43Y] = E|(2X-Y)\-X+3Y)]- £[2X-Y]- E[-X+3Y] = E[-2X?+7XY-3Y?]-(2E[X]- ElY])-(-E[X]+32[¥)) = —2E[X?]+ 7ELXY]—3E[¥2]+ 2(ELX]) —7ELX]BLY] +3(E(Y]) = -2(ELX?]-(ELX])?) +7(ELXY]-£LX]- ELY]) -3(21Y?]-(E1Y))?) —2Var|X |+ 7Cov_X, Y]—3Var[Y] —2(2)-3(3)+7(-1) 8-45 Cov[2X -4Y,X+7Y] = —20. 2Var{ X]—28Var[Y]+10Cov(X,Y) = 2(5)—28(2)+10(-1) = —56. 8.4 The Conditioning Formulas 8-46 Using the conditioning formulas directly yields, E(B] = Eg {E[B|G]} = .4(50) +.3(70) +.2(65) +.1(24) = 56.4. Var[B] = Eg \Var[B| G]} +Varg {EB |G]} are ae }:2(652) +.1(0) +|.4(50-56.4)? +.3(70-56.4)? +.2(65-56.4)? +.1(24-56.4)?| Here is the complete calculation in tabular form, modeled on Examples 8.4-1 and 8.4-2, where_X is the sporting event attended and Y is the quantity of beer consumed: Pr{X] | E[X|¥] | Varly|X1 | ELEY X]] | £[veri¥ |X1] | Eley xP] 6.4 20 Cubs aa or = — 1470 845 845 E(Y]= a X]]=56.4 E[Y|X] = ELey | XP |- ely) = 3372.6—(56.4) = 191.64, Var E|Var[Y |X]]}=1101.4, and Var[Y] = E[ Var[¥| X]]+ Var| E[Y | X]] = 191.64+1101.4 = 1293.04 = Oy =55.96 1000 250 180 ® SOLUTIONS TO EXERCISES - CHAPTER 8 8-47 (a) ELX] =Ep{E[x|6]} = Ey 2 2-5 mifhion: (b) Var[X] = Eg {Var X |6]} + Varg {ELX |0} Fo 62 0 D |Yar {8) I 3h h+ qo To Bet? I Seer 2252 nis: Soi Fa (million)° => = Prev > he) = poe 18} Me oy = 3.323 million 0 fig Ot ta Pees 6 Pree ee joe d0+ , x -e-5d@ = 5742. 5 8-48 (a) E[S] = Ey {£[S|N} = E[X,+X2+---+Xy] EL X,]+ E[X2]+---+ E[Xy] = N- py = un: Mx. (b) Var[S] = Ey Var{S |N}}+ Vary {E[S| N]} = Ey{N -Var[X]} + Vary {N-ELX]} Var[X]-ELN]+(ELX]) -E(N] = E[X?]-E[N]. Note: 8-49 See also property (2) for random sums. If we let T denote the total losses, then Sern l 12,500 Pp E(T |N]= ux N =12,500N 15,0007 /12 and Var[T|N]=o2N= 15 ,0007 12 E(T]= E[E(T |N)]= EluyN]= wx ELN]=12,500p Var[T] = En \Var{T |N}} +Vary {E[T |N]} 2 “ é) 7 =Var{T |=Ey \o} N}+Vary {uN} =03 pt ui pq = Ss= +12,500? - pg lis )2 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 181 8-50 In the case of partial damage the expected payment in thousands is given by, by parts with w=x-1 and dv=e-*/?dx reoOF J, @-1)-[.5003-e~*?dr] = .5003 f“ete ak 15 5003[-2e"(x+1)] II 15 = 1.2049 ] Exe ReymeeelDI 1.2049 pele E| Payment] = E [£(Payment |Damage) | = .94-0+.04-1.2049+.02-14 8-51 = .3282 thousand This is an exercise in detail. As an example, we calculate (B) Pr(S =4). Think about all of the ways that the sum could equal four: Pr(S =4) = Pr(X¥ =0,Y =0,Z =4) + Pr(0,3,1) +Pr(1,1,2) —. FH denoted Pr(0,0,4) + Pr(1,3,0) + Pr(2,0,2) + Pr(2,1,1) = (.4)(.25)(.55) + C4)C4)GC15) + (3)¢35)(.2)4+ (3). 491) + (.3)(.25)(.2) + .3)(.35)C15) = .14275 S=X+4+Y+Z 0.0365 0.07025 0.0965 | 0.14275 0.20125 0. 141 | N]rRe |] BI] NH} DA] CO}N] OO} EUS] = 0-.01+---+9-.066 = 5.2 0.16975 0.066 0.066 182 SOLUTIONS TO EXERCISES - CHAPTER 8 8-52 EY] = y-2-(l-y) dy = 7 Let N be the number of claims, a binomial random {, variable with n=32 and p=1/6. = Ven Let X,...,Yy be the policies with claims and let Ae [S] = E[ES|N]| =wy -sy 8-53 Let S=¥,+¥%)+---+Yy 8-54 ae 68 16 denote the random sum. RN oD= 32, p=). ee N ~ Binomial(n Val S\=00 Leal mite sy =+ and o; ==. _16 and oy Fa NIELS 37.1 eS We Prarie 37 a =32 2h a: ale Opal On iy + Lino far = LORE E eae The random variable T|.X ~ Unif[X,2X]. Soper fe a) ee 0x ef 0 otherwise 2x-3 yp Pr(T (i) = — : ——_ 5 dxeS id = 64 x>15 ere Sal. Let N be the number of people hospitalized. The unconditional probabilities for N are, Pr[L <1| (A (A) Then N is binomial with »=2 and N=0] = @ and Pr[L <1|N =1] =. p=.3. Given N =2, let.X be the driver’s loss and let Y be the passenger’s loss. The joint distribution of XYand Y is uniform on the square [0,1]x[0,1] andso PriL=4+Y<1|/N=2] = 0.5..Now, Pr[L <1] = E[Pr(L<1|N)| = (.49)(1) +(.42)(1) +(.09)(.5) = .49+.42+.045 = .955. We use Bayes’ formula to calculate the conditional probabilities for NV given that the loss L 1s less than 1. Pr{N =0|L <1j=-%, 955 Then, 2 2V |ie<1| 1 <1] = Oh. By =l|L<1]= err yand Pri Vi=2 20: $e (B) = +1--92.+2--983 534 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 183 Section 8.5 Chapter 8 Sample Examination I: Ha Vinee x. for Osa 61, 0 otherwise Alternatively, using the transformation formula: Se V=e 2x = ] dx 1 X=— = x 5 In ny and —ose SO, trf (y)p= 2. 1 3 ai =i y| ) , ay c 3(In y)? = 5 << <p) San : Ieyece = : (a) F(z) =Pr(Z <z)=1-Pr(X =z NY =z) =1-(e-25) (e773) = 1-825, (b) Z~ Exp(B =>). jes (c) pz <r note using independence 3 242 (2 =e (Gr 2-e 142507 (a) Mot) = Myzy() = Mx(t)-My() se a sea a (b) The MGF for S is that of anormally distributed random variable with mean variance 34. (CyAPiCs —4 ett Z =< 7 Veer 214) =.097: 4. Let Ny represent the number ofclaims from class 4, N, ~ Binomial(n = 500,p=.01) so that E[N4]=5 and Var[N4]=500-.01-.99. Similarly, Ng represent the number of claims from class B, Nz ~ Binomial(n = 300,p =.05) so that E[Ng]=15 and Var[Ng]=300-.05-.95. | and 184 SOLUTIONS TO EXERCISES - CHAPTER 8 The claim amounts, X 4 and Xz are constants, so wy, =500 and wy, =300, with ce =O7, =0. Thus, if S4 represents the total claims from class A and Sz represents the total claims from class B, then, E[S.s] = Mw, «Hx, = (5)(200) and E[Sg]= pv, x, =(15)(100), with, of, =o} Hy =(500-.01-.99)(200)? and 6}, =02, - u2, =(300-.05-.95)(100)’. Let T=S4+Sg represent total claims. E(T] = (5)(200) +(15)(100) = 2500. Var[T] = (500-.01-.99)(200)? +(300-.05-.95)(100)? = 340,500 = of = 583.52. The 95" percentile is 2500 +1.645-583.52 = 3460. _ k 3460 =< =1.38. (EB) 5. Let f; denote the number of pensions provided by the city for female recruit i. The probability distribution for f; 1s 0 l 6 (.4)(.25) =.1 E[ fi]=.7 and Varf f;]=.81. The total number of pensions is T= E[T]=70 fitfre+fot+-++fiooe and Var[T]=81. Using a normal approximation of the actual distribution of 7 (with correction for continuity), we find Prt 390) = Pr{Z . 90.5-—70 N ——— os — = Pr(Z $2.28) = 99. (B) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT (aye Pt] Max(XY) @ 185 <-6)'—=" Prix <6 <6] II i]s Prey )26] (ietemye(3) 1354 (b) Let Mbe the maximum of X and Y. Fy (m)-Fy(m)=(l-em |, 0<m<2 Fy (m) = Fy (m)-Fy(m)=(I-e")();,_ 2<m<oo E[M] = lp[1- Fay(m) |] dm . an (2-m+me") dm+ |e” dm : [4-24 d+m| er = 5[2+(1-3e%)]+e? = $G-e%) = 143 Var[N] = E[Var(N |A)]+VarlE(N |A)] fee, ee = BA| =), = Vari A| = 15> (3-0)? = D5 rv) 5000 Fy(x) = ‘vy(x) = 1-| —— We Pr[.X > 2s25,000] and d so Peixe= 000} Pr[Y; > 25,000] = i Pe (BE) 12 5000 |) — ) ea = (.2)!2, Ss4r (Ey Let XYbe a single claim severity in units of one thousand. Fy(x) = 1-e~* forx>0. Py Gj E[W) = Let Wbe the larger of two claims. a l= ge te [, U-Fr() oboe Then [2a 4 Then iorx> 0. ee" | Hie Dn + =1.5 thousand =1,500. (C) 2 186 ® SOLUTIONS TO EXERCISES - CHAPTER 8 10. Let X be the smaller of two observations and let Y be the larger. Then the range is Y —X. The joint density for X and Yis given by, fii via cere tor0aycoe Usrays and the expected value of the range is given by, 2[, [deter drdy = 2], Lyffexae— fjxetarle dy 2D:i |y(l-e” )+te* es | e dy = 2f [pre-e ay l 2= 21+ 5-1) (D) CHAPTER 9: STATISTICAL DISTRIBUTIONS AND ESTIMATION Section 9.1 The Sample Mean as an Estimator 9-1 (a) Weare to find v—Z05° te. Ly <W +2Z05 | Vn Dre GAS. Vn 2350-1.645- 2 < pay <2350-+1.64522 |=(2144.4 <py < 2555.6). J16 (b) 2==— ia jz= J16 S Jn ro? Ao mek h sarees J16 9-2. Let H ~ N( uy =70,0H =5) denote the size of arandomly selected hamburger. a= stim att denotes the sample mean. 4 Pr(ff <77) = PZ < a Ber 8) = 9974, Sih) 4 9-3 The confidence interval for the population mean is [ghee poss Ke (a) The sample mean must be in the center of the confidence interval, so X =11.00. (b) Theerroris ¢=2.50=Zo4 ero rar MESas IRY . The sample size was n=49. oar 187 188 @ SOLUTIONS TO EXERCISES - CHAPTER 9 9-4 yw=100 and o=16. (a) With asample of size n =100, ro7 = 100" Te | 167/100 Pr(97 < X <103)= 103-100 16/./100 = Pr(-1.875 <Z <1.875) = 93.9% (b) With a sample of size n = 200, : Pr(97 < X <103) =P (c) 97-100 16/./200 r(—-2. 65 2 <Ze 103-100 16 //200 2.03) = 99275 The likelihood that the sample mean is within 3 of the population mean increases as the sample size increases. cae leit So Section 9.2 Estimating the Population Variance 9-6 X= L434 9,.67 15,04 1254106 ; : 5 cd _ = 12.36 is the unbiased estimate for su. ie): * Zz S? = (14.3 12.36)" +--+ (10.6 =12.36)" 5-1 = 5.363 is the unbiased estimate for the population variance. 9-7 CERO Xe (n-1) ; v Gas : Via 4) 3239 se or eranly Leas So the 95% confidence interval for the population standard deviation is HERE vey es VIAL PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 9-8 n=10, so there are 9 degrees of freedom for the chi-square distribution. We are to find the value @ so that .025 = Pr(S <@). OZ 5e— 9 PTGS <a) = POS <a .§2 = Pr a OY an }= Pr(9-. SS? <9 -@ WAL: ae O| = Pr| 729)< Ais OY 9: 7) a Oy o = 72 »s(9) = 2.7 41.67- ere So =, jeTaT _ 92.92, 9-9 Begin by calculat ing the sample mean and sample variance. X =147.5 and S? =629.1. a Hohl) Lean) (5)-629.1 11.07 (5) -629.1 ipl 16.86 < ay <52.30. 9-10 .95 = Pr{S2 <c} y (19)-c 5 as Wiel) “9. woe Om = rr Ain Bade = ——— - O- =3.17 ie) Ou Zs @ 189 oy =41.67. 190 SOLUTIONS TO EXERCISES - CHAPTER 9 Section 9-10 The Student ¢-Distribution = 9-11 Frente a | Sy$C]eee Ftgate-| Sy $e] Vn Vn 92-2650 17.9 12 Bee -.2+2650/ =A. 22a y= 12 Ji4 43.2 +127 —55.9 < uy < =—30:5. 9-12 (a) OQ 298+ 2.93 $027 +2,50 X = alee eee 2.619. si Se +---oeee + (2.50—2.619) = 2 (2.58-2.619)* Gee — 092353. (11-1) (ey eX =F (n-1)- DIT S Wie <= X #1, (nl: oe Vn 2 vn 2619-2, 298:|022293.) 2 i < 2 6192,098 4oe | 28! vi, 2.619—.062 2.0) 9-13 (a) < tty aye S Ji < 2.619+.062. Obs 1 fy(y) = ——— yp te Bo -I(a@) / ) ana .r(2) u=,[2-. Then, y=nu? and ole n ] WAY = —_____. du > yl"2)-1g-) Nw 9 V5 0, Let ae SO, dy VO) ee OY)du z = = = I _ ylriD)-1p-(W2)¥ aera a I renee 9 NE 2 )\(n/2)—1 e ,-(1/2)(nu*) -2nu oe ste ie wee an? F(a) apy n y” 2 ‘yew 2)(nu? ) = 2) : Sax 2 (4) yl l enn 2 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT (Hea ae ae? for so Aes AG Then, f7(t) = iL.fu(u): fz(tu)-u-du of = Oy n n/2 4 3 tel ete ae 2 IP n/2 ») = (+) : gt ] : 5) v=ure - > U=vV _ nu 2 fda 20 4, and 2 yy) 2 -@ at 2 aa yn -eM2Y(nte"\u* hy n/2 fr r i) zh 5) 1 je 1 2B n/2 gi = ) lI N|s a Ee, eS 9-14 Sx : S? Foatmeln—-l) ls~~ ( 1 | >; —o<f{<0o y 81 a eee ne oya _ —8i - F'95 (12,7) ee tA) ce era a Ep foae2. ere OF et 357 Ips a) erin i)y Ox od Bas -Fyj2(n-1, m-l) oy 2/n © ee (a) ). = eon eg yey IF Ben =b set Ly | 2 ee . [ave ue eoW/2)y(nte)v . hy 0 r( )Noe (p r( orl = fell 2 1 + — 2 Z y4 : n © nl, p-(l/2)(n4t?) [2 n—| 2 — -y-du a ) the 2 eae) -u-du B(1/2 l - [) ie 2 du= svn ) dv n/2 (c) _tu (Qo [ Uy'-@ V2 py) 1 ie 1.86 < CX < 19.28. OF 5 ee @ 191 192 ® SOLUTIONS TO EXERCISES - CHAPTER 9 9-15 5.42Bl)reel EG meen1 5,42 ee of ROai ia AGH Taye Ae TEEN ce or 2 2 eb 2 lS) ee -3.80 5,07 4 05m ioe 5,97 Taking square roots, 9-16 Ue y SP hs < 1.7846. Oy For your bowling scores denoted as Y, n=6, X =147.5 and Sy =25.082. Your sister’s bowling scores denoted as Y, n=8, Y =194.375, and Sy =14.272. 25.0822 | oe ee ee (a7 Fe) 25.0822 1 (OT << 25.0822 of. Ones <A Ge i428 082" 44272" 5 4.88 BR) = 10m 3 889. Sister 9-17 (a) U~y2(m) -7($.2) Soleene ae 2), using Property (5) of the 2: . - m 2 m Gamma distribution in Subsection 6.5.2. Similarly, 1SCy eapoy Y emma Teer) ky ed Ds n ZR 5) fs(s) = Foror he SS URIS sean 2 or (tee S W =—, + (4) Dae m/2 n/2 (4) 2 \ (ip(n/2) =I oe = ti2 9 (wt) 2).e = mw £)'=] 22 4.dy aC PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT a m/2 - 2 ete 5 2 a ; ae Aes ws |n 3 "2 iia we ~9(n+m)/2 2 n+m 5 (4) Awe m/2 B= Bo PB ~nt+mw_ 1+(m/n)w —— (2)"" ay qantm and n/2 _(nt+mw)t eS eee See dt (m+n)/2 | (2 (n+m)/2 , on n for 0<w<o Section 9.5 Estimating Proportions 9-18 (a) ~ P= (0) = 2070 aangh 7, 69-31 REG) ih, (EES 205-3999 (c) .676 = .69—.01389< p<.69+.01389 = .704. “ Gk pee eer! .65 —1.96 a Aas Pt Zai24{ pine Seep. 65 +1.96,|°>. 57.5% < p< oe @ 193 194 SOLUTIONS TO EXERCISES - CHAPTER 9 ; AD (1—f = Zain} pee aTRAS = Za/2-0156, which 9-20 p=.57 and 03 =€=Z 92°, /eO—P) implies that Zy/2 = aren =1.92. Therefore the level of confidence is 94.52%. Because of rounding, the newspaper poll likely had a 95% level of confidence. 9-21 We have p =X, where the population X is a Bernoulli trial with probability of successp. From basic principles E[ p]= ELX]= p and Vari pj=— Ee) = PA Then E(pQ.— p)) = Elp)-ELp*] = p-(Varlp]+ ELpY) = p- [At pei~ lh Ee = p=p)-=*n = pq| = fo At q S =| fl 8| S = l (1 1g 3 As o x1] FH See aaa Section 9-6 Estimating the Difference Between Means 9-22 (a) (X-Y)-Za/2 en Civ=ity) = (A =F)ae Sadnd a8 a m m n (14.3-13.8) +1.96- n 2 (igsue) fi.g2 1.62 < (14.3-13.8)+1.96- Jaret os (AC) 59 te, (= ee ST end —.09 seconds (b) < (4#y—Hy) < 1.09 seconds Since 0 is within the confidence interval we cannot conclude that the estimated difference in population means is statistically significant at the 95% confidence level. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 9-23 $2 = i (m-1)S? + (n-l)S3 (m+n-2) eo eI a 26 ke = 227,052— 15,0887) For a 98% confidence interval with (m+n—2)=26 degrees of freedom, £o1(26) =2.479. (X-¥) -taya(m+n-2)-Sp,f+ +2 < (uy-uy) m n i l < (X-Y)+tg/2(m+n-2)-Sp-,/—+— on m (104 —97) — 2.479 -15.088 ota 2 Une) oe 9-24 14 (Ti —1s) ee DS. Begin by computing the sample statistics. Aig =10 = ee The pooled sample variance is Sp = i -,6377 +(9)-.9757 eee = (3.287—2.468) —1.721-.80 a+ 2 inane 9 05950=",00 EOD) 240 < (gp —Lpy ) < 1.398 Section 9.7 Estimating the Sample Size = [25-233 = 212.07, so sample size is at least 213. : @ 195 196 ® SOLUTIONS TO EXERCISES - CHAPTER 9 926 Fe [2] 2 2 zs 2s = 917.56 be SHOTS. ce 7) n= [eee 9.8 = 467.86, so sample at least 468 new professionals. Chapter 9 Sample Examination Ve aca ee ee 3 Sy =1.904. (b) Since we are using a small sample and the sample standard deviation, we need to use the Student-t distribution. Fre) D S|<i = X+taAUG 1)- S Al ey 3 204) Js Shean ade: [12 64 < wy <5.36 hours. Se, We are given 112 < wy <120 1s a 95% confidence interval and oy =16.07. We can use “= (4) 21 oe 6 pines to, ee and the margin of erroris ¢ =4. (hyn {|Ox *Za/2 |-|iE) z + he: PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT (Xna-w? = a ix)? = SE we - = 2 2 = <7 (nc, Titus (cet YX) ky 0B) i= (a) X = oo aie 64.5)? +. 645)? Cae [(72CER EON (b) Pte 5]Sine 2 = 5.345, Field n Se D 58.83 = 64.5-—2.998. toi(7) n ie |Nee c4s.2908/| S388= 07 B v8 error (c) The confidence interval for the population standard deviation, oy, is (n-1)-S? (n-1)- Se Le (n-l Hain? =i) 75.345 mae aa Women < Oye X57) S 5 X95(7) so that we are 90% confident that 3.77<oay <9.6. (a) X =21.6 and the margin oferror is € =0.7. eee 281.75. \, © n= See) -( 28275) ==) || ee alt =\| —_—_—— = 49. 2 @ 197 198 ® SOLUTIONS TO EXERCISES - CHAPTER 9 yx) -kY = S1(X?2-2.8-X;4¥2) i i! x we <S X; +nX? =l i=l = \\X?l -2X(nX)+nX? 10 nh LX; hs y _ i=l \e— ee oe 10 = [lil giel So 710) 2 = -2 | n a me {Ea a = n-| = 1,282 —-1-(110)?(110) 282 — 75 9 5.95 : ; =1).S ee = rea= 1575(9) S — < X5259) = 902 = 3.78 = oe 9S2 eas <o* Di 9S- S s 48 < oO. = 26.67 is a 95% confidence interval for a7. (2) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Also, cae a v7(9) => whe Pe[3.3= 350) Dus OS = <a : Pro? <A = mee (9. OS oe = ao = 21.62 1s a 95% upper confidence interval. (1) Finally, es. 90) = n—| S? 9 (n-1)S? _ OS hare yee Pre =16.92) r ete(9) at = Pnico) <o?a Thus, 4.26 = ‘oa <a? <@ isa 95% lower confidence interval for a7. (3). The answer is (1), (2) and (3). SSeS OE (E) ee J2o 2. yy eee Lae <i ; pAy 52 Thus, It follows that k=l andr=2 (B) a # 199 200 # SOLUTIONS TO EXERCISES - CHAPTER 9 6 y DX? Since the means are zero, DY =) ~ 77 (6) and 6 Dee 6 a 5104 (oy: es Using independence, F(6,9) =-="_——_=1.5-=!_—_ = 1.5W . Then, 95 = PrlW <a] = Prfl.sW <1.5a] = Pr{F(6,9)<1.5a] (A) Oars en ga OU) ) = OO Se S¥ 5S 992 aLX OF. A ~ y7(5) and ¥ 2 G2 Sy Sy —+ ~ y*(9). Thus, F(5,9)= Ox oe OF PrBSe Sy|=| <3= 2k= F(5,9)< 6~.99, Yo since F’9;(5,9) = 6.06. to25(9) =2.262. Y (B) Hence the answer is (A). to25(8) = 2.306 and the margin of error is f925(8)- 1 i) S S ox a Ws) The confidence interval is Let S=) 4 5+4.61 (D) Xj ~ 77(8). i=] Then, “ws =8 and of =16 > ws —205 = 8-(2)(4) = 0 (A) CHAPTER 10: HYPOTHESIS TESTING Section 10.1 Hypothesis Testing Framework 10-1 (a) Ay: God does not exist (b) A Type I error means rejecting the null hypothesis that God exists even though God does exist. (c) A Type II error means accepting the null hypothesis that God exists even though God does not exist. 10-2. “Answer (A) 10-3. Y ~ Poisson (Ay =(25)(.1) = 2.5) (see Subsection 4.6.4). Let @ be the significance level [hen ee es 10-4 Pres | 25 yee |e = .758 (D) ~ l The null hypothesis is that 44 =288.9 with the alternative being su < 288.9 X =286.9 and p-value = PrLX <286.9| 4=288.9] = Ve 286.9 —288.9 3 ma =,,057 Je There is only an 8% chance of observing a sample mean of 286.9 or less given that the population mean is 288.9. Section 10.2 Hypothesis Testing for Population Means 10-5 ¢(24) = 42-45 8/5 ="—1.875. From the table for a two-tail test, we have, ¢49/2(24) =1.711 and ¢95/2(24) = 2.064. Thus, do not reject at the 5% level, but reject at the 10% level. 10-6 #(20) = SS = 2.92. Pr[1(20)|>2.92]=.008461 (D) Using the function TINV in Excel shows that, (E) 201 202 10-7 SOLUTIONS TO EXERCISES - CHAPTER 10 Since the surveyor wishes to demonstrate a lower average price we use the left onetail Student-r test: . Ho: =A 21 Aye 4921 From the table, ¢19/2(26—1) = 1.708, so we reject the null hypothesis if f is less than —1.46, so we do not reject the null =157038. Then, 4(25)= eee 6.42 / (26 hypothesis at the 5% level of significance. The surveyor’s claim is not substantiated by the data. 10-8 Since the surveyor wishes to demonstrate a lower average price we use the left onetail Student-t test: Hg b—49 71 Hy up <49.21 From the table, ¢o2/2(24—1) = 2.500, so we reject the null hypothesis if ¢ is less than —2.500. Then, (23) = AT37—~ 1-21 _ 4 636, so we do reject the null hypothesis at the 1% 3.42/24 level of significance. The surveyor’s claim is substantiated by the data. 10-9 Since we wish to substantiate the claim of higher starting salaries we use the right one-tail Student-? test: Hae P= 952,000 Heil oa.000 From the table, ¢2/2(9—1) = 2.896, so we reject the null hypothesis if¢is greater than 2.896. 5 Then 7(8)= PRA —52 : Uae = 2.250 , so we do not reject the null hypothesis at the 4000//9 1% level of significance. The salary claim is not substantiated by the data. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 203 Section 10.3 Hypothesis Testing for Population Variance 10-10 Ser ae -1)S? The test statistic is a Seemv7 (n-1). Take Hyp :0* =50, with, H,:07 >50, so that we reject the null hypothesis with a sufficiently large value of the test statistic. This leads to a right one-tail test. Calculate ¥ =13 and S? =119.192. 11971995 =11.92. From the table, 77,(5) =11.07 and Then, ¥7(5)= be 50 Maps) 1 60. S0R025 <p <,05, (D) Section 10.4 Hypothesis Testing for Proportions Mei MeTH tests tans ieey ee ace Dlowb. ole) n sy. 100 With a@ =.05 we reject the null hypothesis for Z = 2.95 =1.645. Thus, the rejection region is 20( p-.5)>1.645 If x is the number of “good” then p= saa n => Te; 100 so reject the null hypothesis for x =100p =58.2. 10-12 p2>.582. (E) Presumably, Aaron is claiming he is an at least an 80% shooter, so this is a left one25 tail test and we reject for Z <—zo5 P- Po — .694-.8 =—-1.645. = p evan 694, and —1.58, so we cannot reject the null hypothesis (ee — Po) ap 8: cE 8) that he rene at least i 10-13. Since the claim is at least 60% skip, this is a left one-tail test and we reject for tS, | vA P-P0 p= 2 = 5375, and, 80 i SERat Popo) 67 (1,6) n 80 ee 114. so we cannot reject the null hypothesis ofat least 60% skipping class. 204 SOLUTIONS TO EXERCISES - CHAPTER 10 Section 10.5 Hypothesis Testing for Differences in Population Means 10-14, — 10-44+12-5 S2-= (m-1)S% +(n-l)S? je P (m+n—2) 22 Let 5 phere meer W=X-Y . This is a right one-tail test and the test statistic is 1(22) = —27=15___. 2.132,/4 + +b = 572. From the table, £/7(22) =1.717, so the p-value is at least 5% (using TDIST in Excel gives a p-value of 28.6%). (E) 10-15 +(n-1)S7 _ 11-4 +11-12 ee §2 _ (m-N)S¥ ee . Chee 20 Let W=X-Y. This isa oo F066. sien? left one-tail test with critical value T= 1 ,,(20) 221725. The test statistics 120) =e ene. 2.966,/4, +45 11 l It-T|=|791 + 1.725| = 2.52 (E) 10-16 Assume the female populationX and the male population Y are independent, normal, and of equal variance. Then we can check, Ho 2x — My =5 versus Ho: uty IC ee ete) 3.84,{4, + a — uy >5. Sp= SE From the table, ¢)/2(40) =1.684, and ¢(46) <t1/2(40). Thus, 1.800 = ¢(46) >t (46), and we can reject the null hypothesis. S2 10-17 Under the null hypothesis f = F(7,8)= ! OR Sy 9 te From the table, F = F'o5(7,8)=3.5. Then |f-F| = |1.778-3.5| = 1.722. (C) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 10-18 Calculate S% =155.6 and S? =28 For L., the test statistic is y*(n-l) = (n-1)S% _ (5)(155.6) 1556" BET. 50 This is a right one-tail test with y/;(5) =11.07, with test statistic (n-l)S¥ _ (4)(28) =2.24. Ge eed 0 so I. is true. II. is a left one-tail test But, 73,(4)=.71 and therefore we cannot reject the null hypothesis, and so II. is false. For III., the test statistic is So ; : F(6,4)= = = ae =5.56. This is a right one-tail test with Fs (5,4) = 6.26, and y SO we cannot reject the null hypothesis. Thus, III. is false. The answer is (A). Section 10.6 Chi-Square Tests 10-19 The test statistic is T = y?((r-1)-(c-1)) = 77((4-1)-(4-D) = 779) (A) 10-20 Under the null hypothesis the expected value e for each cell is 24. There is no parameter, so the degrees of freedom 1s 4. 2 (Aye= (32-24) _ G2- 24)? mh 24)? ple 24)" , 20- 20) 24 24 24 24 24 X25 (4) = 9.49, so the p-value is at least .05S. (E) 10-21 # 205 This is a 2 by 2 table, so the degrees of freedom is 1. The fvalues are: No Claim Claim Group | 40 10 50 Group 2 120 30 150 160 40 |200 The table of e values is identical, so y7(1)=0. 72 (1)- 735(0)| =|0-3.84 =3.84 (D) yj;(1) =3.84. 2 206 SOLUTIONS TO EXERCISES - CHAPTER 10 —— a0 [100-50= 30 [2000] aebard alsa ea 72 |e 5 waa RE aC ees | BEST otaleey [eet1008 2]eter 0ore eapawanraseny 773) =4.133 (B) 10-23 The given 3 by 4 table of actuals (/): Number 4 i x Z é = = , 1000 ; a ya(O) — (97ai ace RU ee Soe ee oe ———————————————E + 1o4 37 3 9 = 11.91, Then, |v2(6)— 735] = |11.91-12.59] = .68 é —|])2 yand 75;(6)=1 ¢ (E) Section 10.7 Chapter 10 Sample Examination Is The null hypothesis is p=.3, The test statistic is, Z = hae sn ee (.3)(.7) a 9 with the alternative p<.3. a = —2.06. , and we reject the null hypothesis for This is equivalent to p <.2056. 100 Then xs np=100-2056=20,56, (D) 59 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT a Let S = X,;+X2+X3 ~ N(0,307). — # 207 Then, —_So = 205 =1.645=> 0'=12.16 (C) (@F a ee a [aa te (2ye (74-67.03)* 67.03 eo ar | =.6703 ~.2681 =.0616 (16-26.81)? F (10-6.16) 26.81 6.16 12 Since the mean of X is zero, we have Oy oS 67.03 26.81 = 7:49. (D) ? ~ y7(12). Under the null hypothesis this means 5DY? ~ y*(12). This is a left one-tail test, so we reject the null hypothesis if v7 (23272 -.((12)=4.4. Thus, ae <4425 0° ¥7 <44 (A) i=) We reject the null hypothesis if 77(5) > v6) (5) =15.09. (m,-10)° n (m2—-10)10 10 yaad&) Also, ny = 60—m, —(144+8+10+8)=20-—m i Ay eames 10 OO = (14—-10)° i (8—10)- is(10-10)- f. (8-10)10 10 10 10 (nm-—l0)* >63.45 (A) = eA m-10 = > 15.09 10-n,. Then, 208 # SOLUTIONS TO EXERCISES - CHAPTER 10 6. This is a two-tail ¢ test in which we reject the null hypothesis if t(3) ( ) > 15.84-10 ¢(3) = 2.97. ( ) =———— 4/f4 025to25(3) ( ) =3.182. a5 Donot : J reject. (D) Frequency (f) | 40-Prob (e) E-. 20 De > pia = = a eed (11-8)? (11-8)? goey £112 —_ (5-4)? (10-16)? 4 TerrmenTG = 95 24 Sy (Oy = oom Db) 8. This is a two-tail ¢ test with rejection region (2) 2 ot995(2) t= 4.303 Given that t(2) =4.10, we do not reject the null hypothesis. 9. (E) This is a right one-tail chi-squared test in which the test statistic is ican = ~ y7 (14). Under the null hypothesis o* =10. Hence reject the null hypothesis for E 2 P To Wns (14) = 23,69 eel A236, ew () 10. This 1s a right one-tail test using the normal random variable W. Since the population variances are both given as 100, TAA EAE CaeDe 25 ee eee ee Ow Js Using the standard normal table, the p-value of this outcome is, Pr[Z =1.4 = 1—.9192 = .0808. Hence, reject the null hypothesis at the .1 level but not at the .075 level. ge (D) This is a right one-tail test using the test statistic, Vee \ON= Sian 102 75/10 (9)= OAKS AOS 12 120072 2.79. tyns(O) =Bea and fyi(9) =a Rae Hence, we can reject at the .025 level but not at the .01 level. (B) CHAPTER 11: THEORY OF ESTIMATION AND HYPOTHESIS NN TESTING ee Section 11.1 The Bias of an Estimator 1-1 The CAS Exam 3 tables list the Gamma distribution using @ in place of f. Thus, ELX] = a6 = 46 and E[Y] = (100)(46) = 4000. Then, e= se =.0025. (A) 11-2. Let X be exponential with mean @. Then the median m is given by 5 = F(x) | J) =1-e"? = m=86I1n(2). The density function for Y3-is, le = 30[1-—e 7 | |edtel: 6 |) eres | talile 30) |= a +e 0 -—2e 9a 9 |, allBe In general, {,xe * dx =? (using the constant fora ["(2,) density). Thus, 6?9 Ae 11-3. 11-4 a. 3 es 1 —6@. 60 55 The bias is = 9-6 1n(2) = 098 (C) The distribution function for the minimum random variable is 20 F(y) = Pr(Yoo) < y) = 1-LeA] = 1-e OA, Thus, Then, B = Y29) 1s exponential with mean i The population has exponential distribution with mean E[W]=6 +4. Therefore, E[W] = 0+4 and so 6 = W —4 isan unbiased estimator for 6. Section 11.2 Building Estimators 11-5 Let X be the claim size population. From the distribution type, pe Kee = ere _—_— oes = ee From the data, X= ”. (x) =1-— (42) for x >10, and , ye = x10" = 100/6 = 100/.6=10 —_— ae re Sa ©) E 209 210 # SOLUTIONS TO EXERCISES - CHAPTER | I 11-6 From the data, n =20 and ¥ = =-[0-141-34--+8-1]=3.1, and, BU espe ee 21 24a= 13 oe 1—32)-23.49.. de at er1+1°--3+---+8 i eaepiee re Equate. EIN |=rp=X =3.1 and Var[N|=rAd By=f- =3.49. 3.49 Then 1+8 =" = B=.126 (A) 4 11-7 From the data (in thousands), Y =75 and OG i=l =18,458.5. From the Exam 3 20° ines Gopala-\ ehap ei (a—-1)(a@—2) Then, #=75 and © Sas eo) a-l 26? ee 11-8 = 18,458.5. To solve, take the square of E| xX] to calculate, : 37 Oa 2nd ap 2(a-1) aie 18,458.5 = ee helnd(O) 4 From the data, Y =325 and roe =119,/50 se7> =) st a25e 8 tan i=] E[X]= af =X =325 and Var[X]=af? =T? =8,125 a= (OB) =—— =13 ap $125 11-9 => (E oe The density function for an exponential distribution with mean £ is ee lL -x/B [te enwis 0 i x20 eeu So the likelihood function is ib (P) l smootthy PScael po 2oy B Se 76 Bs [ —— Next, find the value of # that maximizes In{L(B)] = —nIn(P) (X; + X2 +°°++X,,) B Bore —N Xp ter tx Setting the derivative equal to zero, we have 0 = / += a which implies 2 = Neta Ossie Hl te PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 11-10 The maximum likelihood estimate for an exponential distribution is the sample mean, = ¥ = £50+1123 +67+1257 +435 +1896 = 904.6 hours. G 11-11 The likelihood function for this discrete distribution is L(p) = [p-—py* |_p-d- py ]--[p-d— p)* J=p"-d-pystet. In[L(p)]=n-In(p) +(x1+.x2+++ xy):In(1=p). d inf gp LM See = =nm (my +%2=+::'+%n) = 0, which implies the maximum likelihood estimate is p= Tr + 11-12 The likelihood function is L(@) = Axo Oxy" | ea ese -Ox,?7 = g” “(X] 23%) o Soe Sea oe In[L(6)] = n-In(O) +(0-1): In(xy -x9 ++ Xp). Take “ and set equal to zero to find the critical value to be De 11-13 =n In(x, -X2°X3° L(O) =(6 °°" ee) +1)" (x; +X_). and In L(A) =nIn(9 +1) + Oln(x)---x,). Take ise and set equal to zero to find the critical value to be dé @ 211 212 11-14 SOLUTIONS TO EXERCISES - CHAPTER | 1 Gye Yorn By end ¥-L-r(n,2). Se (x) = sees te xt le(’/P)x for x>0. (4) @-p: (b) ~ ETA] = El = ess fe (x)dx ie fon on = on _ptle(nlBx by = = Gh 11-15 F( = : i ie n-1 = LDL ao i nl (n—2)! = (= n THA = bee x) == lee —(x/0) } Bee eee n — eg een—-1 3 me toe a eo — p—(c/8) = —— = ¢€ = ini 12 3 We take e=mn( $8. The MLE for 6 =.Y =3.78. Thus, the MLE for ¢ is inl$3.78) =1,087. (B) Note: This is an application ofthe invariance property of MLEs under transformations. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 11-16 # 213 Let N be negative binomial, the number of failures before the r”” success and ie B+ltp Then, para \ iP k B lank Pk =Pr| [N=k|/= | | (ga) 4 ay: . The likelihood function L(r, 8) = po: p3- ps, where po -sa) : py SaaNeIne | )( B )ad 3! P35 Broly Resi 5] su] baa = i hen, L(y, )= po: ps ps = ps =) 2 2 Mel] 9 bes 3r Let Y= y,.. = We) — Ml Themey Cn 8y= ‘5 eve for O< y;; for O> y; n Prieretoren 1 (0) = ee =ere Gas 0 for O mint yi, Va) for #>min(j,..-,¥n) ' ‘ ia=min(j,,-..-, y,) L(@)=0 theta values As the graph of L(@) shows, L(@) is an increasing exponential function of @ up to G=min(Vj,--- y,), and then drops to zero for 8 > min()j,..-, V7 )2 a Dus, the maximum of L(@) is reached at = MM eee! ple) 214 11-18 SOLUTIONS TO EXERCISES - CHAPTER | | From Exercise 11-13 the MLE is, . = ee a ee a5 Ces n(x) toe =) In(.92 -.79-.90-.65 -.86) 11-19 f(x) = F'(x) = (6+1)x°. “ C= = Z ee Oe ©) From Exercise 11-13 the MLE is, = Sige In(x; 20,59) F8 Se UEC Sa s/s eae In(.56-.83-.74-.68-.75) 18) ( ) -f Lt 11-20 L(@) = O"e xX, xX} Xn =— => InL(0) DOP, 5) = 10.67 11-21 thousand (C) Let p be the probability of winning a bet (success). Each bet can be considered a Bernoulli trial random variable X with parameter p. The probability function is fy (x) = p*(l-p)'~* for x =0,1 and 0< p<1. S = s be the observed number of successes. Let be the number of trials and let Then, u i - ys L(p) = [[p*(-p)'™ = pa i=l InL(p) = slnp+(n-s)Ini-p) F the MLE for pis A AY => sl =x) -0-p)a = OnE [In L(p)] oe - p on : l-p a eat « es . p=—. Note that this is the same estimate used in standard n binomial proportion problems. In the problem at hand there are (20)(3) = 60 trials e and 7+(2)(3)=13 successes. Thus, ee 60 n (A) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 215 11-22 Directly from the data, ¥ =5 and the unbiased sample variance S? =5.5. Let p be the probability of a claim on an individual policy. From the solution to Exercise 11-21 above the MLE estimate for pis p = we = .05. The annual claim count is binomial with m=100 and probability p. The ‘variance of the annual claim count is mpg =100 p(1— p). Therefore, the MLE of the variance in the annual claim count is, 100- p-(1— p) =(100)(.05)(.95) = 4.75. The difference in the two methods is 5.5-—4.75=0.75 11-23 (D) ForX exponential both the MLE and the Method of Moments estimator result in X as the estimate for 9. Hence, 6, 0; = (a (CG) 11-24 a2 11-25 At From Exercise 11-12 we have 9=————_>» Leap e9 a = In(.25-.50-.40-.80-.65) = 137 2 ee) . Thus, (B) Let gq be the probability of a$500 loss. nent pa giao 2 peel — gs | > pss. Since there are 3 outcomes on each ofthe 3 trials, the probability of observing the particular outcomes $0, $0, and $1000, using the multinomial formula, is, 3 2 Since p< peg p. Setpie eel this probability is maximized for p= 216 SOLUTIONS TO EXERCISES - CHAPTER | | Section 11.3 Properties of Estimators 11-26 (a) Foreach x; we have E[x?]=E[X?]. nm 4.2 Then, bs 2 | = PEL] n n i=l = E[X?], and so the bias equals zero. 2 2 PT. y2(n-1), and SO vay| |= Var| 72(n-1)| = 2(n-1). o o (b) On the other hand, 2 rar| o 2 —l)o* oy n = —Var[T*] = Var{T*] = aioe =ee —> 0, and hence J? is consistent. Note: There is a minor complication here since T? is only asymptotically unbiased. This is because T* = 415 2 where S? is the unbiased estimator of o*. Thus, the argument showing that the variance of T” converging to zero implies consistency requires a slight technical modification. In f(x) =In3 + InO+2Inx-0x3 => LACES 00 See ge The CRLB is given by — oO) : : na]ome Un f2)] =— 0 = — : — = Ce is git (E) 06? et ae nel U. s 1 ee In Example 11.3-1 we showed that Var[@,] = n(n+2) Since this converges to zero as n goes to infinity, the estimator is consistent. 11-29 Since B is binomial, E[B]=np and Var[B] = np(l— p). Let p= ity n Ronn Joes | p(l-p) Then, Z| p|=p and Var[p] = ——-np(l-p) = ———-. (a ai Since this converges to zero as n goes to infinity, the estimator is consistent. 11-30 In f(x) = In A — Ax = The CRLB is given by A Vise sul A)2 hy: a ete ee aA2 ng|Slnfe:2) | = RA ary i ) So n (E) * 97, PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 217 Section 11.4 Hypothesis Testing Theory 11-31 (a) We use a standard normal z test based on the sample proportion p= in where X is the number of made shots in 100 tries. The rejection region is A — 80 z <-1.645, or equivalently, 2H < -1.645 > -80-.20 100 X¥<73.42. We reject the null hypothesis if Grant makes 73 or fewer shots out of 100. (b) Under the alternative hypothesis XY is binomial with n =100 and p=.5. Thus, “x =50 and oy =/100-.8-.2 =4. Wisk yer = Pr| X >73| p=.5] = Pr) Z> 11-32 esas V100-.5-.5 | mee Use a right one-tailed rejection region of the form X > 4, where, 05 = Pr x2 4|u=10|)= Prizes | > A=" 1.645. Vl/n lin 06 = LF <ajqett] = mz<a 4 Aah 15558 81564571. 050 bhen Vl/n Subtract these two equations to get, fel) eee Scie a Pi mee NOE 7 Jl/n n> (1.645+1.555) =10.24n211 (B) a peg : Under the null hypothesis 2=1, Let NV be Poisson. Then, Pr|NV= k] =m a we have, (ras) Reject the null hypothesis if N >k , where k is the smallest integer for which, Pr[N >k] < .10. From the table, we have k =3, with Pr[N >k] = 1-.92 = .08 < .10. Then, the power equals, 1-8 = Pr[N>3|A=2] = 1-Pr[N =0,1,2|/ =2] 50 = 51 e722 52 — 7 = ]—5e- = 325 (B) 218 ® SOLUTIONS TO EXERCISES - CHAPTER | | 11-34 11-35 The significance level is the probability of rejecting Hy when Ho false. II. is true, and III. is false. (B) is true, so I. is likelihood ratio critical region is of the form, | and so, the L(A) =An"eAMte+"+%) e(nitx E+E) EX, ) L(1) LQ) e(™ +X |” el Ftp Q*etarmetm) 4++-+Xp_) =< 2° k =e k' on <k. Thus, xX] a X2 =f-se6 Ny = Ink’ , < or equivalently, X <nink'=K. The population X is exponential, meaning that, wy under the null hypothesis. Using the central limit theorem, 05 =PrLX < K]~Pr z7< Aa 1/./100 eae G45 1/,/100 =o =1 = Tiles (B) Notes: iN, since Uy = + , the hypotheses can be restated as: Ho: uy =1 and Hy: yy =.5, a left one-tail test, so that the form ofthe rejection region X < K could be deduced without working the NeymanPearson inequality. 2. The statement of the problem asks for the lower limit of the critical region, although the solution, 1.1645, is, in fact, an upper limit. Se 11-36 The fact that Y =1.5 is not used in the solution. With a significance level of .05, the only possible rejection regions are: {0}, {0,1}, {2} with @ values of .01, .05, and .05, respectively. The power ofa rejection region is the probability of the region given 6. Thus, the powers ofthese regions are .04, .08, and .09, respectively. Therefore, the most powerful region is {2}. (B) 11-37 The Neyman-Pearson lemma will result in a right one-tailed rejection region of the form X >A. The calculations are similar to Exercise | 1-32: 05=Pr[¥2 d|q=10] =P z2 ae Bem S250 nH | 95=Pr|X > A] #=1|=Pt] = Z> = V25/n 3645. WZOtn = “teh =—1.645. Subtract these two equations to get, ANAS ed a Nie 7 n> (25)(2-1.645)? =270.6 oases =*2*}.645. 2oun =>n>271 (D) Then, PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 219 11-38 Let U be the maximum of the 5 observations: Under the null hypothesis the population X is uniform on the interval [0,1] and so, Fy (x)=x> for0<x<1. Now, 5| —— 11-39 > Pr/ Uc] == ji— =1-Fy(c) Peo) =1-8 =n i eailSiiees a e-(1 = = Sg = The calculations are similar to Exercise 11-32 and 11-37: 05 = Pr[¥>C|u=100] = Pe aca so at V9/n V9/n et eel Os PryZee ee ee eG V9/n V9/n ; Subtract these two equations to get, Ee) ea eS rife saan ee J9/n Pies V9/n Then, n> (9)(1.645)? =[3(1.645)] (C) Section 11.5 11-40 More General Likelihood Ratio Tests 025=Pr[X>k|4=L0=15]=Pr|Z>4a! = AH 1.96 = k=3.94. B = PrfX<k|w=5,0=15] = Pr] Z<22 2 =-71 Ps Sea 11-41 2 oe Olle sel ° Ex ‘0 09, a X| The likelihood function 10)=[]| 4 +|-() e weg ort (CO) XS Under Hp; T= 8=10 and L(Ho)=100)=(4) e l (47 10°? The denominator is the maximum of L(@) over the parameter space, which by definition, occurs at the MLE for 6. The MLE for @ is the sample mean, X= a =o 4. _ Thus; L(H,)=L(X) =L(.4) (4) e 94°” and the likelihood ratio is, CEG) ety 100% ON 44D e CP8 4) 9. = L(A) =..991 (B) 220 # SOLUTIONS TO EXERCISES - CHAPTER | 1 0 Nowe ee => (ae 2 ok Now, under the null hypothesis, for, Osx <1 for x<Oandx>1 > x<K. R(x)=0 for x>1. Thus, .05 = Pr| X <.05 or X >1] (D) SE n I ECA) (a, B) Nd fa prercemerees a-l (xj-x2-+ Ese ++ os Xn) n ] (x; Xp Xp, ) e i The numerator of the likelihood ratio is L(1,1) = ree css) and the denominator is, Z(2,1) = (x, -x2+ ++: -Xp ye ia ae a LT) ~ h the likelihood Thus, ratio is, R(x) — Oe— 1 ee . The rejection region takes the form, xy X2 hp Ske Six xg xy Si 11-44 Ss Es k te In— s Dinx, >In—-=C (B) The numerator of the likelihood ratio is f(x;0)= 1 m[l+x- denominator is the maximum of /(x;@), which is — | for —0<x<o. The (occurring when x=@). 4 Thus, the likelihood ratio R(x) = es eee + => 1x] 2 ssak, fia form ofthe rejection region. To calculate K, we have, 05 = Pri|X|>K]=\|1- le s t = Then, tout tan-l(x) 1 x=-K .05 = 1——tan=+(K) Wa = |-— (tan” (K)—tan7! (-K)] 1 . = 1-—tan“!(K) 1 K =tan( %..95}=tan(4752) 2 (E) is toe PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 1145: eh [le a Sea n XG i=l it Xt Xp, Bai aceePea er # 221 onmeniarer ihe likelihood ratio 1s; AoX19 "Xp Dives ae Tie . The denominator of the likelihood ratio is the maximum of Xp! L(A), which occurs at the MLE for 2. By Example 11.2-6, the MLE for 2 is X . Thus, the denominator of the likelihood ratio is, L(Y) =e" +. Xy!++X,_! Hens aaer rejection takes the form, eae e7"4o oe Mbetn! es a rari Cn Np Ao vr nX _g-ndy =e , epnX | 4) 70. =e Litts| eXy+ nX nX 1S equivalent to Ga xX The = Ae) n¢ Skene = Ks (A) Section 11.6 Bayesian Estimation 11-46 Let 6=1,2,3,4 represent the risk category and let Y =0,1 be the Bernoulli trial random variable representing whether or not an accident occurs (XY =1 means an accident occurs). In the solution table below: Pr[X =1| 6] denotes the conditional probability of an accident for a given 6, oia(G) denotes the prior distribution of risk categories, Prion PrX =1) => Trew(P) = 0 Pr xX =—1) 6 )7sq(@), the total of the Prior PrL_X =1] column, Pr[X =1| O]Zo14 (9) , the posterior distribution of @ using Bayes’ > Pr X =1| A]zo1a (P) Formula, Posterior PrLX =1]= > Prf] XA =1\| Ore, (@), the total of the Posterior ‘XY=1] column. Prior Pr[X ==1 |0] 7 old (A) Pr[X =1] Posterior ao Pr[X = 1] =0095 (a) Privat X all = Aew(1). 21584 (b)- > Prior Pri X =1.= .0303 (cy 29 Posterior Pr X = 1] =".0352 222 SOLUTIONS TO EXERCISES - CHAPTER | 1 Notes: 11-47 (1) Since there was an accident, the posterior probability of an accident (in the next period) increases. (2) Since X is a Bernoulli trial, Pr_X =1] = p = E[X]. Thus, the answers to (b) and (c) can also be interpreted as the prior and posterior expected values of X. The joint distribution of @ and - is discrete in 0 ot continuous in x. The prior distribution of 6 is Pr[@=1] = —5 and Pr[@ =3]= —. The conditional density for Ais f(eie=1)= aoa . The joint distribution evaluated at x =5 and @=1 is, x+ FGM (OOS Piet = [ (541)? fA@e)) = : " (5+3)* Pr[@=1| X =5]= 17 Ie) = = Stiritaclyt 5) = ee J\.2 Next, the posterior probabilities for 6 are, 1/72 = ask 3721 ne Bie =4 be eS 3/128 = 6279. a3798 The conditional CDF for X given @ (using the nea distribution) is, F(x|6)=1-— Then, Pr X>8|0=1]=4+ and PrL¥>8|0=3]=>. Using the posterior probabilities for 6, PrLX >8] 11-48 PrLX >8]6=1]Pr[@ =1]+PrLX >8|6 =3]Pr[0 =3] S$(5}orn+( 4 |e279 = 2126. Let X; and X> represent the number of claims in year one and year two, respectively. Then, the conditional probability f(2,.2|q) X; =land X2 =0 is f(1,0|g)=q(1-q). cs evaluated at probability function evaluated at X, =land X> =0 is, f(1,0,q)=q(1-q)a(q) =—— ~—(a" —q°) for. 6<q<.8. The marginal probability for X, and X> ne: at X,; =l and X> =0 , f (1,0) = Pr[ X) slot os= ap FU; 0, q) dq i = oe is(q* i “OF Then, the posterior density for g is given by, 4 f0,0¢) _ 974-7) Tr | ant 2 5 = . Then, Pr{q>.7] = J 8 (at=4°) dg = 156 “ ame 014 (q* =2> 7 <.6 )10l.0< is given by, 2) ig= PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 11-49 # 223 Let @ be the class of policyholders. Then the prior distribution is given by, Pr[@=/] = 2/3 and Pr[@=J/] = 1/3. Now, E[X|@=1] = 12,875, and E| X|0=1]=6,675. Given X = 250, the posterior probabilities for @ are given by, Pr/@=1|.¥ r[9=1| BPR r[9=1| =250] = (.5)(2/3) eal X=250] (52/3) 00/3) > 17° 28% eA er (.7)(1/3) X =250] Gio senda a ye Posterior ELY] = 12, 875(72)+6 6754] = 10,322. 11-50 Let @ be the class of policyholders. Then the prior distribution is given by, Pr[9 =1] = 1/2, Pr[9=2] = 1/3 and Pr[6=3] = 1/6. Let Xbe the number of claims. Fromithe table sh((O=1) = 1, E[X|\e@=2| = 2,and Aix 10 =3)= 3. Given X =1, the posterior probabilities for 6 are given by, iy ESL Pr[9@ eel ag) esee 6)(1/3)+(0)1/6) (1/3)(1/2)+(1/ 3 (1/6)(1/3) =1] = X Ms =2| ame ch Mat2) ea nOnediayeOliey (iam he Oey (1/3)(1/2)+(1/6)(1/3) +(0)(1/6) Posterior ELX] = (1)(3/4) +(2)(1/ 4) +(3)(0) = 1.25. 11-51 The joint density is feima=(4 } 0<x,x%2 <O=>O> max(x,x2). Then, with Y; =400 and > 500 Q? 4’ * i 4° toe =U eeLhe, 5000+. Note that Let f(x,,x2) be the marginal density for 11, X2. =600, (400,600)= [a 5000+d6 = Thus, the posterior density for @ is given by, f (400,600, 0) =(3 £(400,600) 500 (3)(600)° (3)(600) 39-4. cpr Since X is uniform on [0,9], we have ELX |6]= £ , and the Posterior x ELX] = (2 (3)(600)30-+ dé = (3)(600)' 63d0 = 450. p) 600 600 ey Note: In some ofthe tabular calculations that follow we have rounded numbers, typically to 4 decimal places. This may result in some columns not summing exactly to the number in the totals row. Students attempting to replicate the calculations should avoid rounding numbers in intermediate calculations, either by using exact fractional arithmetic, or by using storage cells on a calculator or spreadsheet. 224 11-52 SOLUTIONS TO EXERCISES - CHAPTER | | Let @ be the class of coin, with 6 = A corresponding to coins 1-4, 9=B corresponding to coin 5, and 9=C corresponding to coin 6. Since a coin is chosen at random the prior probabilities of A, B, and C are 2/3, 1/6, and 1/6, respectively. The given sequence of outcomes is HHTH and Pr| HHT.fee P| = p(1-p). Prior Moa (O) |Pr{H |0] |PrLHHTH |@]| Pr[HATH] | tye (A) (4)4 1/2 A | 22 .0625)(2/3) | |4 ~o613 (.0625)(2/3) ae =.0417 ~ 6809 Posterior Pr[H]| (1/ 2)(.6809) = 3404 (1/.4)3(3/4) | (.0117)(1/6) | R222 | (1/4)(.0319) =.0117 = .0020 ~ 0319 (3/4)°(1/4) (1055)(176) ere. =,1055 = .0176 .0176 FTRSETEY ESCs [Citic Gs IC = .0080 |GL4\2872) / = 2154 ets Cok Uy Me The probability of heads on the 5" toss is .5638. 11-53 The prior distribution of @ is @ =1,2,3 with equal probabilities of 1/3 each. Let Xbe the claim amount. The conditional density f(x|@)= per F(x|q@) =1- ( | a . Given the observation probability function /(20,q@) = A aloe x+10)@+! With CDF given by X = 20, anda value of @, the joint \¢2 The problem requires calculating P| X > 30) using the posterior distribution of @, given Y =20. We will use P[X >30|a] =1- F(30|a) = (<2+10 i a (5) Joint @ 1 Posterior | Xoa(P) | f[20|a] | f(20,a) | tney(P) | Pr[X >30|a@] | Pr[X>30] he AUS LO. 302 0037 = 5000 ay 0025 5) 3 } ; = 1250 PN eS 1/3 200 0074 0625 (.0625)(.3333) } se =3333 vail = 0208 1/3 3000 SS 000R es e002 0156 (.0156)(.1667) = 00124 = 16en = 0026 0074 1.000 1484 ie Totals} Foe (.25)(.5000) 5 .0074 3 1.000 30 3 «| 0074 The revised (posterior) probability that the next claim exceeds 30 is .1484. A PROBABILITY AND STATISTICS WITH APPLICATIONS: 11-54 The prior density for 6 is given by (0) =2;1/2<6<1 A PROBLEM SOLVING TEXT ® 225 and the conditional probability function for X given @ is f(x|6)= i; Jore-GO) 20h =0,1,2. The joint probability function evaluated at Y =1 is f0,A) = 200-6) 2(@) = 40(1-0); 1/2<@<1. The marginal for X evaluated at 1 X=lis4 f(0,0)= |, 400-0) do = + rsa (Gye! ;ue =120(1-6);1/2<@<1. 11 The Bayes’ estimate for squared error loss is E[@]= [,1207-6) ater 16. 11-55 E[,] = 5000- a ; # 5000 = 6,, is biased. I. is false. Since @, doesn’t depend on any observations of X, Var[ 6, | =(, and therefore 6, iS consistent. II. is true. The mean square error is given by, 5000) Jean |_ E|(Ala - 0) | = Eso 10 = _ 5000}- £|(2~ 612 so III. is true. (D) 11-56 The mean square error of an estimate c of X is given by E |(X-c)? |. As shown in the text the mean square error is minimized when c = E[X]=a@@=300. 11-57 A f) 2 (D) 9 The mean square error is given by E|(6-0)? |= i (V0) ie Je.where values of @ are denoted by y in the integral. Using tabular integration by parts, we have, 10. ( = Bia OO y? YY 10 \| | 2 29) EL8 at iy ede cs oe pal (9-0) 5-20-98) Tony oT a st | pees - (10 fee29" “X90 J 4410-11-12 For @=100, the mean square error is 1007 _ 66 \=2 226 11-58 SOLUTIONS TO EXERCISES - CHAPTER | | The bias of a point estimator 1s Bias = E [6] — 0. alG - E{6}) +(z14]- ay MSE(6) = E[(6-6)] II £|(6~ £[6]) + ae AG : £6) |i 2E|Bias8 = Var[6]+2- £{6))|+ E[Bias?] Bias -(0)+ E[Bias?] = Var[@]+ Bias? Section 11.7 Simple Linear Regression: Basic Formulas 11-59 Wehave N=9,X =2,Y =10,Syy =164,Syy =100. From formula (5), S2, Ey aya aes SPAT S2, ee wa EAT IIOG enamel y = T164)(100) aa? => Syy =—112.695 (negative since the correlation is negative). be a Pond 0 AY ORY =1137 oY Sao Then, tt oor, XX For X¥=6,Y 11-60 =d+BX =11.37-.69-6=7.251. Using the given data for_X and Y, we have, N =5,X =3,Y =13,Syy =10,Syy =34,Syy =17. Then, R? ee eee(or)5. (D) Sixx Sy Note: Since the values of Y, are also given, one could calculate ESS and use formula (8): ESS = Syy -(1—-R?). 11-61 Syy =10,Sxy =195=> berSxy =) 719.5, and Sixx @=¥ — BX =45-(19.5)3) =-13.5 > Y =@+BX =-13.5419.5-1.2=9.9. (B) 11-62 N=d,X s45.Y =1.8286- Sp =700,Syy 0804s Sap = nS p= 2X = 0336 and @ =Y—-BX =.3179 > ¥ =@+B-75 =2.8357. (A) 11-63 N=12,X =12,¥ =145.1667,S yy =572, Syy =59,793.6667,S xy =5792=> ie =10.13 @=Y-BX =23.66 = Y =23.66+10.13.X (E) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT 227 11-64 We have InY =Ina+ BX , and, Ce Rea ee ree ase NS|B 9119.01 7.030" |Om 49A 221,098) Prowisp is] [3523 |55|248.28|106.23 N=5,X =3,In¥ = 7.0463,Sxv=10,S(inyynv) =-0294,Sxiny =-5314 => B =23+4 =.0531 and In@ = 7.0463-.0531-3 = 6.8869 => InY = 6.8869+.0531-6 = 7.2057 =>. Y = e797 = 13475. (D) 39, OT 11-65 11-66 =g _ 23 ee 7ee _ BF OT RC ers s N=5,X =7.1,Y =1.06,Syy =.2,Syy =.332,Sxy = 18> (a) p= Sa =.9 and @=Y - BX =-5.33 (b) Y=-5.33+.9-7=.97 (c) 2_ Ra (d) ESS =Syy -(I-R?)=.17 (ec) TSS = ESS+RSS Sey ae _= 488 => RSS = TSS—ESS = .332-—.17 = 162 11-67 wehave Y = a@+ PX , so that the regression line passes through Since @=Y —BX Dah Minit eH G@+2f. Also, 3.3 = a+3B. Eliminating @ between these two equations gives, B=3.3-Y. 228 @ SOLUTIONS TO EXERCISES - CHAPTER | | Next, note that Y = fo" = "w = 5Y —12° Then, Sxy = (0)(5)+()G.5) +(2)3) +3w+t(4)(5)- S)(2)Y =129 5 3(57 -12)_-10Y = 5Y —65, Now, £= 22158 fete OO Combine this with the previous expression, Sixx 10 i= ieee aver aS 282 Sy ey 19 ae eee, This implies Y = 30 and i Sef 11-68 Weused Excel to generate the following chart. Actuary Income § 60000 ” o 50000 E[e) S 40000 y = 2884.9x + 43759 R?ae= 0.9119 Se #® 30000 0 1 ie 3 4 ) Number of Examinations Passed (a) Notice that the regression equation y = 2884.9x + 43,759 should be interpreted as follows: The intercept of 43,759 should be thought of as a base salary. (b) The slope 2,884.9 implies for each examination passed, companies compensate their employees an additional $2885. So if Hilary had passed five examinations while in school, we would predict that her starting salary would be near y = 2884.9(5)+ 43759 = $58,184. Note: The accuracy of the regression equation is measured by R-. However, ifthe data is not chosen randomly or disobeys our assumptions, then our work and associated statistics are meaningless. PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT # 229 Procedure to Use Excel® 11-69 li Enter the data in two adjacent columns, x first, then y. 2. Highlight the data and make an XY scatter plot. You should label the axes and the chart. 3. Left-click on one of the diamonds then right-click to add a trend-line. Choose a linear trend-line for simple linear regression. If a different model fits the data more closely, then another model may be preferable. Such a discussion is beyond the scope of this text. 4. Click on the trend-line to format it. Under options, add the R-squared value and the equation. (a) Salary and Weight 400 350 y = 9.7297x - 189.05 R? = 0.997 Ena | | 2 | 2g | | = 55 Salary (b) The correlation is ./.997 cenit =.998 (positive, since the slope is positive). _39 99. ¥ =200 then ¥ = 220+187:09 Oey 5 230 SOLUTIONS TO EXERCISES - CHAPTER | | 11-70 (a) Children and Income y = -2.9173x + 24.898 20 4 R? = 0.8329 10 | (1000) Income Disposable af > 0) 0 2 4 6 8 10 Number of Children (b) The correlation is —/.8329 =-—.9126 (negative, since the slope is negative). OMT at 4hen 9 0173 04 RoR FI29F ee 11-71 ACT and Grade in Stat 400 . 2 2 90 8 S a * y = -2.5691x + 140.58 R2=0.9723 80 & ® PRS 6 | £ 60 [tS | a 50 | 40 15 20 25 ACT Score Correlation is —4/.9723 =—.986. 30 35 40 PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Section 11.8 Simple Linear Regression: Estimation of Parameters 9) 11-72 Given Var[o7]=4. From formula (16), Var[ B= ca . From the data, DOG Seat co Var p\== ==. GE CE (n-2) B oi gee S yy 11-74 a0 Sie; Syyv ara as i ge fe ae ar Ve GsYe A | Sixx Formula (26) gives, Pe eeSS Te (N=D)RSS= (N= 2)RSS Ts , ESS /(N—2) ESS TSS. ESS” From formula (15), TSS = RSS+ESS. Also, R? = = $0 ESS TSS : 1 Th es RSS ESS ype 5 ESS! TSS’ TSS ame ns Paks iy Le He oo S89) TSS ESS 2 1— R? From the data, N =20,¥ =5,Y =6,Syy =1600,Syy = 6400, Syy =1920=> peek, SAC Saeed 7 asm 100086400 (18)(.36) Mares CV=2 RSS Ber SS BB) i S516 0 105. ® 231 232 SOLUTIONS TO EXERCISES - CHAPTER | | a elit i el at de et Ce ; 1-75 pe Ory ee 1okU = = i Sxx 1600 eS no eleee =1.2 and 6 = Y-BX = 6-1.2-5 = Then, Y =1.2X. For X¥ =10,Y =12. ) 6 = S?= ESS. z ee ee =409 = (1-.36)(6400 ESS = (1-R?)JTSS tos (18) = 1.734. (a) Using formula (23), Y% + ts/2(N—-2) ; ; rv is, 12 ++1.734 interval |2, the confidence 1 (10-5) 2 pee 1600 Jexr56)= (5.3,18.7). (b) Using formula (24), Y + ts;2(N-2) interval is, 12+1.734 11-76 i eH N Sixx fe 1 +—+ eB) S* , the confidence OSE |ae SO) We use the relationship, (1, N—2)= ies — R? , = (—15,39), with Nos and ho 67 36 hen es) -= at 212.942 From the F table ae) Fos (1,23) =4.28 , and /'¢, (1, 23) = 7.88. rit reject in each instance. 11-77 ESS =7SS —RSS =43—37 =6.. Then; F(,N-2) = RSS. —lg~ = aitSA Al oes aay Rl 41.83 =11.91. (n-2) Section 11.9 Chapter 11 Sample Examination ie Maximum Likelihood: L(@) = O°(x-+-x5)?! eGae. => InZ(@) = 5In@+(O-1)In(x:--xs) ae rae => {InL(O)) = F+ini%s)=0 Ale > Method of Moments: ELX] = Gives §=5 =1.07, Then, R—S = Ox® dx = R i osae in(21-43-56-.67.72) aa and X = .27. (A) =13 Nn >| =.518. Equating these PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT @ 233 Ay: fo(x) = 1/10 for 0<x<10, and0 elsewhere, Ay: fi(x) = 1/5 and 0elsewhere. for 5<x<10, undefined for x<0 00 ie) Msp sea Ai See fOvaees LO) undefined for The likelihood ratio is {1/19 x>10 This means the rejection region C must be a subset of [5,10]. _ length of C level is .05, we have .05 = Pr[C] => C can be any subset of[5,10] of 10 total length equal to .5. Then, length of C 5 Pu etersSel Ve Reny =ial) + ky » [hen ELY] = Since the significance ee AK E(Y )i + kxE|Y> | ey = lS hak, since Y,¥,,¥2 are all unbiased estimates and therefore have the same expectation. Then, Var[Y] = k?Var[Y%,]+k3Var[¥o] = (447 +k? War[Y] = (4k? + (1-1)? )Var[ Ya|. To minimize this expression, differentiate with respect to A, and set equal to zero: (Var{Y]) = [8k -20-m)|Var[y] = 0 > h=02 (B) au ies ~ F(12,11). Under the null hypothesis, S7 ony wh edUe SO, S050, e on _ F(12.11). KS, From the alternative hypothesis, this is a right one-tail rejection region. The null hypothesis is rejected if F(12,11) > Fos(12,11)=2.79. 05 = Pr i2 - == ave i Se o FQ IN) 2 a IRB O.5o) Oe PL = OK SZ, 22.79 (E) Then, 234 @ SOLUTIONS TO EXERCISES - CHAPTER | | 5: Let X ~ N(20,27) be the side of the square. £ Z Then, Y ~ N [20,22 ; The estimate for the area is X? and E[X?] = Var[X]+E[XP? = This will overestimate the true area by .8 meter. (D) N =7,W =1.86,Y =6,Sww =18.6,Syy =138,Swy =50 p= ae =2.65 and @ = Y-BW =108 => => Y = 1.08+265/X. AX ab Let_Xbe lognormal, with XY =e”, where W ~ N(,07). From the properties of the lognormal distribution in Section 6.3.4 we have, E[X]=e4*(/)>" ELX?]=e2#+2" Then, InELX] = io o? = InE[X*]—2InE[X]. 8. and and InELX?] = 24+207. Thus, The method of moments estimator for o? is In 4.4817 —21n1.8682 =.25 , and so the estimator for o =./.25 =.5. Since X is exponential, E[X]=6 and Var[X]=6?. Thus, XY = y(o2| Pr[Type II error] = Pr[_¥ <11,000|6 \|| d (B) 12,500] ‘ oe sl _ 11,000 12,500 | Pridé <~-1.2] = 1-@(1.2) =.115 (8) PROBABILITY AND STATISTICS WITH APPLICATIONS: A PROBLEM SOLVING TEXT Cy @ 235 Reject the null hypothesis if Sa mene S¥ 4 = F(n-1,n-1) = — iNas: = 3,077 = Fos(n—1,n-1). From the F table F'(9,9) =3.1789 and F(10,10) = 2.9782, so the minimum sample size to reject.occurs fom —1 = 10 =S"e7=11, (EB) 10. Reject the null hypothesis for X >C, C sufficiently large. Equivalently, reject for a ee ee se C= 81942 (C) n-1| ie The density function for Y, is f(y)=n x ee ana So; 212, |= n n ar! Then the bias is | Z 1a = eee n+l n+] Thus, we need —_T +] <a) a= a 20 eat ol So AW es eae Therefore, 1 must be at least 20. (B) D x= 2 Sohne = a Bs 625) ee) and G30 NOS KIS) 193. Fess tocar 78-11032 = 50.4, (504541103 50 aa yee Thus, 2009 is the first year during which premiums exceed 150 and 2010 is the first year premiums exceed 150 for the whole year. Either (C) or (D) is correct. ) Sakae eine 7 oT eS a —— oer Pa: 7 * d -‘ ; 7 a ~ et it aed ahs . mee ao a: eee ae * of <2°s 7 : : am ae So 4 oe eer a a * he— f wan ro es eat 7 ia < s Pa: ¢ a iy | bree : or ms ‘. - val — e 7 a . ‘the a an ISBN: 978-1-56698 il it81566 987226 Probability and Statistics with An= | TU 7 cf] Ws-AAV-