I

ExRcrsEs

8l

Wc considcr thc discrctc distributior of Eramplc 25.3. Thc FMGF of X is

GAt)= MArtt)

I
= 2-t

t<2

Noticc that highcr dcrivativcs arc casily obtaincd for thc FMGF, which was
not thc c{s for thc MGF. In particular, thc tth dcrivativc is

cY\t)- rt(z-tl-''t
Conscqucntly, E(X) = CAI') = l(2- l)-r = l, attd EIX(X - lD = 6i(l)
=2(2-l)-!=2 tt follows that E(xr)=E(x)+2=3, and thus, var(x)

=3-12=L

SUMMARY
Thc purposc of this chaptcr was to dcvclop a mathcmatiqal structurc for cxprcssing a probability modcl for lhc possiblc outcomcs of an cxpcrimcnt whcn thcsc
outcomcs cannot bc prcdictcd dctcrmidstically. A random variablc, which is a
rcal-valucd function dcfincd on a samplc spacc, and ahc associatcd probability
dcnsity function (pd0 providc a rcasonablc apptoach to assigninS probabilitics
whcn thc outcomcs of an cxpcrimcnt can bc quantificd. Random variablcs oftcn
can bc classificd as cithcr discrctc o! conlinuous, and thc mcthod o[ assigning
probability to a rcal cvcnt ,/{ irvolvcs summing thc pdf ovcr valucs of ,.1 in thc
discrctc casc, and intcgrating thc pdl ovcr lhc sct ''l ilt ahc continuous casc. Thc
cumulativ. distribution function (CDF) providcs a unificd aPProach for crprcssing thc distribution [Link] to thc possiblc valucs of (hc random variablc.
Thc momcnts arc spcial cxpccicd valucs, which includc thc mcan and variaocc
as particular cascs, and also providc dcscriptivc mcasurcs for othcr charactcristi6
such as skcwncss of a

distribution.

Bounds lor thc probabilitics of ccrtain typi of cvcots can bc crprcsscd in

tcrms of cxpcctcd yalucs. An important bound of this sora is givcn by thc Chcbychcv incquality.

EXEFCISES

|,/ L.t. -

lt,f) [Link] 8n {rbitrary outcom. rcsulting [Link] lwo rolls of thc four-ridcd dic
of Eramplc 2.1.1, Tebulatc lhc discrctc pdfsnd skclch lhc [Link] of thc cDF lor thc
followinS random variablcsi

(a) Y(.)-l+./.

CHAPTEi

E1

(bl

(cl

2.

SANOOM VASIAOLES ANO THTIA OISTRTEUTIOXS

zlcl-t-J.
W(c) - (t - i)t,

A gamc consists of first tolljng an oadinary sir'sidcd dic oncc and thcn tossinS an
unLiascd coin oncc. Thc scorc, which consists ofadding thc numb.r ofspoli showinS on
lh. dic to lhc numbcr ofhcads showing on thc coin (0 or l), is a random variablc, say X.
Lisl thc possiblc vlluca ofX snd lsbulalc thc valu6 of:

(a) lh. discrctc pdf.

(b) lhc CDF st itt point5 ofdiscontinuity.
(c) stctch thc Sraph ofthc CDF.
(d) Find PIX > 3].
(e) Find lhc probability lhat thc scorc ii an odd inlctcr.

3.

A bag conlaint thrcc coins, one ofwhich ha.s a hcad on bolh sidcJ whilc thc othcr lwo
cointatc normal. A coin ii choscn at random from thc bag snd torscd [Link]'Th'
nl,mbcr olhcads is a random variablc, say X.
(a) Find lhc discrct. pdfofx. (Hir!: Usc lhc Law ofTolal [Link] with 8r
two-hcadcd coin.)
normal coin and Br
(b) stctch lhc discrcl. pdf and thc CDF of X.

=a

t.

A bor containr 6vc colotcd balls, two black and thrcc whilc. Balls arc drawn succtstivcly
s/ithout [Link] lf X is thc numbcr oldraws !ntil lhc lalt black ball is obtaincd, find
thc di56ctc

pdtl(r}

'.r'

A discrclc random variablc har pdfl(x).

lal lff(xl - Hll2Y for x - 1,2,3, and zcto othcrwisc, find r(
(b) Is a function oflh. foft I$l-kUIPY-ll2l fo. x-0,l,2aPdfforany,t?

6.

Dcnolc by

[x] thc Srcalcst intcSc. no!

cxcscding x. For thc Pdfin Eramplc 211, show

zcro ifi( < 0, and
([x]/12)t for 0 <.x <

lhat thc CDFcan bc tcPrcacnted as F(x)

onc

'(ry
-'

ifr > ll.

A [Link] random variabtc

and zaro otharwist-

ll'

Xlas I pdtoflhc form/(x)

d8

- t)

fot x

0' l' 2' 3' 4, J'

(a) Find thc conslant c.

(b) Find thc CDF, F(x!
(c) Find P[X > 2].
(d) Find E(xI

8.
'
9,

A nonnc8ativc inlcgc.-valucd [Link] variabl. X ha5 a CDF ot thc form

F(x)- I - (lPF'I for

x-0, l,1,,.andzcroif x<0.

(a) Find thc pdfofx.

(b) FindPtlo<X<201.
(c) Find P[X is cvcn].

Somc(irnca it is dcsirabla to assiSn numcrical'codc" valucs to crPcrimcntal rcsponscs that

arc not brsicalty ofnumcrical typa. Fot cxampl., in lcsting lhc color Prcfcrcnc.s of
crpcrimcntal subjccts, tupposa that thc colori bluc, Srccn, and rcd occur with probabilitica

85

E(ERCrsls

ll4, l/4, .nd ll1- [Link]. A [Link] iDt'8'r llrluc is t3!i8ncd to Gach color' md lbii
a."naorn [Link] X lhat c'n lake on onc o[ thclc tbr'c inlcScr velucr-arcaponda to
(2, CanI$l -UFFQI2')I-l'l for : - - l, l, O bc us'd as e pdf lo' lhis crp'rim'nl?

- (llr2)r for x = 0, l' 2 bc 'llcd?

(c) can/(x)=(l - xl/4 lot x' -l,q2bcusrd?
') lO.
kt X b. a disc-tctc .endom [Link] such that PtX - rl > 0 if :-l'2'3'o'thc v'luca
- -'
l,
PtX = xl - 0 othcrwisc. SuPPosc thc CDF is F(x) - 'oJx(l + x) at
""a
O1

c"n 1t'r

1,2,3, or 4'

(a) Skclch thc gtaPh ofthc CDF-

O)

Skctch thc [Link]

o[thc discrctc pdf /(x)'

(c) Find E(XI

tt.

to thc
A playcr rolls a sir'sidcd dic and t'c'ivca a numtrr o[dollars corrcsionding
playc. pay ro..olling
iJnru., otaor on thc facc that tumi up. what amount should thr
to makc this

't2.

. "tair-

Samc?

A continuout rsndom variablc

zafo olharwisc.
(a) Find thc [Link] c.

pdfgivcn by/(x)

has

-'6(t

- r)xt it0<x< t

rnd

(b) Find E(Xl

"

t3.

/(t)

A tudction

h&t thc following form:

.f(x)-lr-o'tt
and zcao othcatlaisc(e) For what valucs

off

is

<x<'o

/(x)

Pdn

'

(b) Fitd thc CDF ba3.d on (a).

(c) For *het values of t docs E(X)
r

t4.

Gxist?

Dcrcrminc whcthr cach of thc foltowing functions could bc a

of thc domain:

(.) F(r)-.-';0<r<co'
(b) F(:r)
(c) F(r)

' tt.

- al -cocr<0.
- I -.-'; -l<.t<co.

Find th. pdfcotrdponding to cach of(hc lollowing CDFs:

(a) f(i) - (xr + 2x + l)/16;
<x<1.

(b) F(x) - I

-.-rt - lr.-r';

-l

<x<

coil>

0'

r 16. ll l,(xl,t - l,2,...,n,.rcpdft,.howth.t

prllr)

is a Pdf whctc P,

;0

and !P1

cDF ovcr

thc [Link] pe.t

CHAPTER

E6

t7.

AANOOM VASIASLES ANO

fXEIfi

DISTRISLITIONS

A random yatiablc X has a COR such ttrat

r(x)

(rl2 0<r<l
- ln t<x<Jn

lx

(a) Graph F(x).

(b) Craph thc Pdl

/(tI

(c) Find P[X < l2]'

(d) Find Plx > l2l.
(c) Find P[X ( 1.251.
(0 whal is Ptx I.2t?

t8'

A continuous tandom variablc 'x has a pdlof lhc fonn/(ir)

2t/9 for

0<t<3'and

[Link] othcrwisc.

'

(a) Find Ihc CDF ofx.

(b) Find PIX < 2].

(c) FindPt-l<X<1.51.
(d) Find a nlmbc n such rhat Ptx < tnl
(c) Find E(xI

'

Ptx 7 nJ,

t' h-g), A r^ndom vltiablc X has thc pdf

fx' iIo<.r<t

JA)-|ry irl <x<2

(o

othcr*isc

(a) Find rbc mcdian olX.

(b) Skclcb thc Sraph of thc CDF and sho\P thc Position of $c hcdian on thc glaph'
,

d0.

A continuous random variablc X has CDF givcn by

ir*<t

fo
I

F(x)=

l2{x-2+ l/x) if I <x<2

il2<x
[l

(a) Find thc l0O x pth pcrccntilc of thc distribution eith P - l/3'
(b) Find rhc pdf of X.

tt.

Vcrify that thc followint function ha! thc four ptoFrticr of Thcotcd 223, rnd find
points of discontinuitl if anY:

[Link] i[-co<x<o

rr't-lo.s
''
I

ito<x<l

[l-c-'ill(x<o

22.

For rhc CDF, F(t), of [Link] 21, find

continuous tyPc, 4(x), and a numbcr

F(x)-aFlr)+(l -

o)4(x)

CDF of discrct' tyPc,

d < I such thal

<

Flr)'

and a CDF

of

lh'

t7

EtEictsEs

'e.

Lrt X b. a nndour [Link] witb dbdctc pdf/(:r)

- r/8 if x -

l,

2, 5.

.nd zc.o othcrwilc.

Find:
(a) E(x).
(b) v.(x}

(c) E(2x + l).

l zt-

Lct

(a)
(b)
(c)
(d)

bc cortinuout with

Pdfl(r)

3xr if 0 <

r < I, and z"o othcrwisc'

E(x).

va.(x).
E(x1.
Find E(3X

sxr + l).

x bc continuoul with Pdf/(x) - l/xr if l. < 'x <

(e) Doca E(X) cxist?
(b) Do.s E(l/X) cxist?
(c) Fo. what [Link] oft dos E(xt) criit?

25.

Lct

26.

At

Find:

<o, and zcro

othctwisc'

a compulcr stor., thc annual dcmand for a Panicxlar soflwaac Packegc is a discrctc
random variablc X. Thc storc owncr ordcrs fout copi6 of thc packeSc al sl0 pct copy
and cbarg.s crslomcrs 535 pcr copy. At thc cnd ofthc tcar thc plcka8c L obsolclc snd
thc oen.r los6 (hc invcstmcnt on unsold copi6. Thc pdfofX fu givcn by thc following

lablc:

(a) Find E(xI

(b) Find Vu(X).
(c) Exprcs thc owncr's nct profil y s!

va(rI

27.

Thc mcasurcd [Link] ofr circlc, R, har

(a) ihc [Link] valuc ofth. redius.
(b) th. cxpcctcd [Link].

(c) thc cxpactcd

'

pd

! [Link]
tl?)

funclion of X, and find

641

E(

l') rnd

- rl, o<r< I. Find:

eaar-

28.

Prov. Thcorcm 14.5 for thc conlinuou! c&tc. ,Iinl.' Us thc lranlformalion y
y/(/ + /) k sn odd function ofy.
thc [Link] notc (hat r(/)

29,

Conridcr lhc discrlc [Link] vrriablc X with pdtgivcn by thc following tsblc:

o22Ji
(6-3v6)/16 r/8
Thc [Link]

ofX

er2r6

is nol rymmctric Why? Show

thstA

0.

- .t -

F in

CH PIIN

8E

30.

NANDOM VAA|AAIES ANO THIIN OISTAISUTIONS

Lct X bc c nonncSalivc continuous random variablc with CDF

intcgrstion by parts to show that

E(X)

E(X)

<

co. Usc

-F(rDdx
E(lXl) < co,lhis rcsult

Norar For dny conlinuous random vsriablc with

lo
f.
f(x)- - | F(r)dx+ |
J-Jo
.

.id

r6

- |

Jo

gl

F(r)

-F(x)ldx

. lowcr bouDd on
this a usclul bound?
(b) Rcwork (a) tor thc probabilily
l? < X < ll.
(c) Comparc this boirnd to thc cxacr Probability.
(a) Usc Chcbychcv'r incquality to obtain
Excrcisc 24.

cxtcnds to

Ii

F[5/t < X < ?8] in

3?,

Considcr ihc random rariablc X ofEramplc 21.1, which rcprcscnB thc larScst of tt{o
nulnbcrs that occur on two rolls ola four-sidcd die.
(a) Find lhc cxFctcd Yalu. ofx.
O) Find thc variancc ofx,

3.

supposc

(^l

E(x)

/'

- lr

and var(x) = dr. Find thc [Link] mcan and variaocc of:

(b) l/X (assumin8t +

0).

(c) ln (E(assuming X > 0).

34.

X is a randorn variablc with MGF

Supposc that

Mlo

= (l/8)r' + U4l.1t + lslSret'.

(a) what is lhc distribution of X?

(b) what h PIX - 2]?

35.
i136f,

Provc Theorem 25.4 for a nonncgativc intctcr-valucd randorn variablc

Assumc that

X.

X is a continuous random variablc with pdf

.f(r)

cxp

[-(r

2[ if -2 < r

< co and zcro othcrwisc.

(a) Find thc [Link] gcncrating funclion ofX.

(b) Urc thc MCF of (a) to 6nd E(X) snd E(XrI

37.

Usc rhc FMGF of Erarnplc 25.5 to find

38.

ln [Link]

3.

Show that 01

E[x(X

IXX

2)], and th.n 6nd E(xt).

26, supposc inslcad ofordcairg foua copic! of thc ioftwarc paclaSq thc ttorc
owncr ordcB c copics (0 < c < 4). Thcn thc nurnbcr told, say S, is thc smallcr ofc or X.
(a) Erprcss thc nct profit y d a lincar funclion ofs.
(b) Find E(n for cach value ofc snd indicatc thc aolulion. that matimizls thc
[Link] Profit.

- E[X(X- lD-Ct- l]

thc c1lDoLol
iO. l/,t*Atl -lstM{r)L eb'!c 'l'll) ir I MGF Tbc funaioo flr) it c'lhd
-' iliill r-fui oi:jt, *a utJiiut of tb''rb dcdv iv' cY'lo1p6 11 1 - Q 4 - ry'!\o)'

rlt [Link] of X.
th.t/ - 9 0I
tlal tr - 9i(0I
Show
O)

ir [Link] thc

(.)

Show

(c) Us. *r{0 to 6td ,r lnd or for thc rendom varieblc of Ercrcisc 36'
(d) Utc **) to fird rr lnd tt for the Endom vsriablc of Erarnplc 25'5'