r

16 Pro b a b ilit y a n d Ra n d o m p r

Solution
Let X: the milage'in thousandsof miles of the twe.
Then X is a R.V with p.d.f

,x)0

x<0

(i) Probabilitythat one of thesetyres will last for afinost

10,000miles
=plx<roJ
=Jor(*)a*
=
| *llo
"['1"-**
=l-e-rl
Ll,
= 1- e{'5 = 0.3935
(ii) Probabilitythat one of thesetyres will rastany where
from 16,000to l
= p [1 6<x:2 4 ]
r24 1 --:-
=
lu ;i" 'od*
_ e_o.8_
e_1.2

=0.4493- 0.3012
= 0.1481
(iii) P[Oneof thesetyres will last atleast30,000miles]

=plx >:o]
f I "-ia*
= r3o
Z0
= g-1'5

=0.2231
,r Example13
J A random variable X has the distribution function
[0, ifxSl
F(x)=Lk(x-1)4 ifl<x<3
ifx>3
Find (i)P[2. X. 3J and(ii) MeanofX.
 Random Variables
L .L 7

Solution
rhe p.d.fof X is f (x)=*t(lr)
fo, ifx < 1
.'.r(x)= -')'
|;ntl< iil:i.,
=r = +rcf(* - 1)'dx=1
[,-t(4*
(*
ouf '')''l'='
L 4l, 1
1 6 k=1 = k= G

; r(x)=[i,'.-r' forl<xS3
otherwise

(i)p[2.x.3]=f r(x)ax
=l|r*-1)'d*
=*[t,.-'l']'
15
=-
t6
(ii) MeanE(x) = f,,.xr(x)ax

=if .(,.-r)'d"
rIx(*-r)o (*-t)''l'
+[ 4 20 .1,

=i[io')-rPa]
=!f t, -9.l
4L sl

l-!=z.a
Mean=
Example | 4
cableXis acontinuous RVwitfrn'a'f = tot(t - x)'0 < x < 1'
Thediameterofanelectric {(x)
a numberb suchthat
Findthe(i) valueof k. (ii) rhe c.d-fof x (iii)e[x =;l ;'*t;)(iv)
r (x < u)= r ( x > t )
F-

ProbabilitY and Random

Solution
(i)To findk
=t
f_r(x)ax
n X(,.-")a x=t
kfl-!'l=1
3l
L2
=!=1 .'.k=6
6

=[3"0-.)iffi;:"='
r(*)
(ii) c.d.fof X : F(x) = f o* (t - *) ax

=u[,4-4]
3
[2 )
=3x2-Zxt ,for0< x <l

l::L'I.,
...r(*)=[3;'-r.'

( o f l . * .!l
1t1
1ii,yr[x 5) =-$,-]
=il :. x. ?l
tL;'"';l

- fii(.-.')*
_
fii(,.-"')d*
 L .1 9
R an d o m V ariables

(iv) To find b such ttrat P (X b) = P (X > b)

1
'
=:
"'P(x<b)
r b/ r\. t
+6[ (x -x')dx=-

(rr'- z*')or=l
+6bz -4b3 -1=0

6 = ! satisfiesthe above equation'

;. ,lo=i

ExamPlel5
f(x)=3x2'0<x<1' Find a and b such that
A continuous R'v X has u p'd'f
0'05
(i)r(x < a)= r(x > a)and(ii)r(x > b)=
Solution
( i)P (x<a )=r(x>a )

.'.f,r*ta*=
fl*'a*1
...at=l_a3 + ut =1,

"'a=0'7937
(ii)n(x>b)=o'os
A
+ flx'a*=0'05
b3 = 0'95
"' =
= b 0'9830

*:rltt""
withc.d.fF(x) = p[x < x]. Showthat
variabre
lu:r,*.*r*" random
E(x)= f [t-ttlrl]*'
Solution
SinceX is non-negative,

f *(4*
E(x)=
 .3 9 P ro b a b ilit y a n d Ra n d o m p ro c e

1.3 Moment Generating Function and properties

MomentGeneratingFunction
\-//
The momentgeneratingfunctionof a r.vXis defined Mr=o(t)=81",,1 ,)
origin ) Similarlythe M.G.F aboutthe meanp is definedby M
*=r(t)= Ol",(x--
Notation

1.M.cF aboutoriginuoQ)= ul/,)

2' M'G.Faboutmeanlt u r(t)= tle'@-a)l
If Xis a continuous
R.V.then,
M,(t)=fa.y6ya*
If Xis a discreteR.V. then,
Mr(t)=le"r(x = x)
N o te
U ,(t)= e -' Mo(t).

Mo me n tsa b o u t th e o ri gin
r ... - - ) .,
M*(t)=nl"''l=slt*+
't-uL' J - " 1 ' l * . . . . +t'x'
- - , + . .IJ
u - ""a
* (t)=r+L,n(x1+ln@,)*.....*L
...u rl*,1*....
= M* (r)=t* t'
, l pi* ! p i* . .
l!., 21..
".+--;.1t"+...
Hence the rn momentaboutthe origin

It, = co- efficientor Lnu, (t)

N ote
r) p:=#lM-(,)1,=,
x =a)=nld6-)lana
2) M*(t)(about

u r (t)=r* +,....*\p,,*... where

f,ai.*r;
 Random Variables 1 .3 9

Pi = nl(x - o)' J= r'hmoment

-t
aboutx = a

3) t) = zl"'(*'llona
Mr(t)(aboutmean

Mx (t) = r + **o, *.....*\ o, *... where

f,.n

P, = El(x-t)'] =r'h centralmoment

Limitations of M.G.F
l. A random variable X mayhave no moments although its m.g.f exists.
2. A random variable X canhave m.g.f and some or all moments; yet the m.g.f does not
generatethe moments
3. A random variable X can have all or some moments, but m.g.f does not exist except
perhapsat one point.

Properties of M.G.F

I. M*=oQ)="-*M*(t)

Solution
M*=,(t\-- nl;v -"11= El"o)= e-''Mx (t)
"-"'
2. T\e M.GF of the sum of a number of independent random variables is equal to the
product of their respective moment generatingfunctions.

i.e.,If X1,X2,...,Xnarez independent

randomvariables,then
(t) = M *, (t)M
M *r+x2+.,......+xo ,^ (t)
"r(t).....at
3. If Mx (t) = zle'* f thenM o Q)= M* (Ct)
Solution

u u Q)=r[el*11= nle{"v
1=M* (ct\
4.If y= *+ b,thenMr(t)="*M*(al) whereU"(t)=M.G.Fof.X.
 L .4 0 and Random Pro

s.rr Y = thenM, (t) = oM change

oforiginand
+, "-"" r"u*t
"[f) "f
S o l u ti o n
u r(t)= u ["' "]
I rx-o\l
= sl " 'ol )l = ' --'^u1"til.1
LJ
T'"(;)
Uniqueness
Thoeremon M.GF
The M.G.F of a distribution, if it exists, uniquely determinesthe
M * (t\ = M, (t) = X and I areidenticallydistributed.

Solved Examples
Example I

fr
FindtheM.G.Fof a randomvariablehavingttrep.d.f ;(") = 3'-t < x <2
]
10,otherwise

S olution:
_rr , [ ^ , , -l,
Mr(t)=
' f ,J! " ' ' d ' =J:ll?t |I
ut
L Jx=_l

-"-'
)l

-"'' fort +o
3t
For/ = o,tt,t
* (o)=l | , * = t

f _"-t
...u r(t)=l "tt ,t*u
3t
Ll ,t =0
I
Example2
Find the M.GF of a r.v Xwhose momentsare given by pi =(r +t)tZ'
 Ra ndom V ariables 1 A1
Io.!I

Solution
MXQ)=f!,:
=i"L't'+t)t2'
=i(" +t't(zt)'
r=0

:.T\em.g.fur(t)=(t-zt)-' I
\ / -
(t-zt),

Example 3
Find the m.g.f of a r.v Xwhose p.d.f is defrnedby
,
v fx ,fo ro <xsl
f(x)=l z-x ,fo rl <xS2
L0 ,otherwise
Hence find mean and variance of X.

Solution
M*(t)=Lf f @)dx
=fxsdx+tr_,1**
f
=L'l.?j
| (a')
7),.1'-ai.;],
,"-l' l- " "-12

e' e' I e, e2, et

= _ _ _ + _ _ _ J- _ _ _
tt't'tt2tz

=---v^=[.,
l-2e' + e2' ( t-
\ J
",\'
h-"'l'
.-.M*(t)=Y
To find mean and variance:

M*(t)=gl=(r.?t.
l'
=M',(o)
u=l*,r.0)],=,
L. 42 P ro b a b ilit y a n d Ra n d o m P

M"(t)=z(r*+.+.
, 2 6
..]
l[+*l*..
)
[ )\23

=1
.'.Mk @)=rft)(i) = I + Mean

ui(t)=rll.l.
l[+.' I ^t' ' * lt+
' L 2 3 ) L z l*" l+ zl' + t* "6 lL 3
=l='F')=I
Mrp)=r[i.i]
x =n(x') - {a(x)\'=I-, =I
varianceor

HenceMean=t anavariance=*

Example 4
A randomvariableXhas probability function p (r) = * =1, 2,3,.... Find
|,
andVariance.

Solution
u*(t)=ir".*
r=l

=ir4l'
2t1l
x =t\ o
)

.l- . r ,:2
=1lr***[1]
21 2
. I
r \2 ) )

'.u*(t)=++=* 2

d=Mk(q=*l*),="
-r'(*')l
=L--pe-1,,
l1r-"'1"'
 Ra n d o m V ariables 1 .4 3

...p'r=l=a-1
'r' -2
LQ-"')'
),=,
(,)],=,
,t=l#""
Ip- n)'(2",
)- +",
(z- " )(-",
)f
LE],=,
.+""1
-lQ-"')za
L--P4-),=,
P L=z+ 4 =6
= p; - (A)' = 6- 4 = 2
...Variance
Hence Mean = 2 and Variance = 2

Example 5
Find the m.g.f for the distribution with

? atx =I
3
p(*)= I atx=2
J

0 otherwise

Also find Mean and Variance"

Solution
Xtakes the values I and2with probabilities and1 respectively.
f
u"(t)=le"r(x =x)

'3
*!"''
"'u*(t)=?"'
3
uk@=?"'*?"''
'3 3
uIQ)=?"'*1"''
'3 3
 1. P ro b a b ilit y a n d Ra n d o m P

:.pi=Mi p) =t *u n =Mi(o)=z
HenceMeanfi=l *OVariance= p'r-@)'

i.e., l r-f, =1
variance
Example 6
Find the m.g.f of a R.V with p.d.f f (r)= ae-*,x> 0 and =0, otherwi
first four moments about the origrn.

Solution
u *(t)= f e'*ae-*dx

' o tr "' ' a' x

t@ -(a-r\x

=ol-"-("4.1*
o- t
I lo
;.M* ( t\= *ir t <u

' =r*I.r.*.....
Arso u*(t)=[t-;)
The r"moment about the origin is given by
tr
p',=co-efficientof . in M*(t)
,rl=;i'
r =l'2'"'
"'P"
=
, r , = 2 , = 76 , , , ,F. ,o = 2
}
4
P r= -|lt z ^o
t,h

Example 7
Find the m.g.f of the distribution defined by' dF =!"-l*16*,..o0 < 'tr< €'
vi variance. 2

Solution
ur(t)= f*e" f (x\dx

=!
fe"e+ax
 Random Variables L. 45

=ilLo 'ddx+
fft."'a*)
*)
= Lrt'.')'d,+f e+'-'t'
i[
={ld9f
z[ r+t
*{-"t'-u}-l
l- t l* L J,]
--l
r [ -+r- t I|
2ll+t I-t.)

..M"(t)=| 7,V t t

Now M" (t) = (t-"1)-'

=I+ t2+ to+ .....

P i=o;P L=21=2
.'.Variance =). andMean=0
h= 6-Q4'

Example 8
p.d.fis givenbyf (')=1-l'xl'-1<x<1
Findthem.g.fof X whose
Solution
M"(t)=
fta'71'1a*
Xr'(t -r\e
= f,{ (l+x)dx+
T .- ,, - lo | ,, rrl l
=l4,1*.\-*l
' t'f_ ,
.l {1r-').+ |
t' ) o
L t' Ll'
=d\e''e'e'1
T - 7- 7- ; - 7 -7

"'u*(t)=t?
Example I
The random variable xassumes the value x with the probability
p(x = x)= q*-rp,x =I,2,3,...Find the M.GF ofxand find its Mean and variance.
 R andom V ariables
1 .5 7

Example I
Xis arandomvariablefollowingbinomial
distributionwithm eanz.4andvariance1.44.Find
P(x >s) andr(l . x < 4)
Solution
Let X - B(";p). Then p[x =xf=nC,p"q, *,
x=0,1,2,...,n
.'.Mean = np andVariance = npq .
+ np =2.4 andnp7=1.44

.'.q =nPq =l'4 =0.6

np 2.4
p=l-q+ p=0.4 and,=2.4 =6
0.4
Hence Pfx = xf = 6c,(0.q'(0.6)u-', x =0,1,2,..,6
p[x >s]=p[x =s]+r[x =o]
= 6Cs(0.4)5( 0.6) + 6C
u(0.4)6
= 0.0369
+ 0.0041
.'.P[x >5J=0.0410
P[t < X < 4]= p[X = 2]+p[X = 3]+e[x =
+]
= ACz(0.4)2(0.6)a
+ 6q @.q3e.6)3+ 6C (0.0),
o(0.4)o
= 0.3I 104+ 0.27648+ 0J3824
.'.p[r<x<4f=0.72sg
Example2
A perfectcube is thrown a large number
of times in setsof g. The occurenceof
;alled a success.In what proportion of 2 or 4 is
setswourd you expect 3 successes?
So l u t i o n
pf2or4l=3=+
'6 3
1)
H en ce O=;,q =f andn=8

Pfx=31
=8c,f1)'f3)' =0.2730
\3 i \.3 i
i.e.. In nearly 27.3 percentof cases,one
would expectsuccesses.
 P ro b a b ilit y a n d Ra n d o m

Example 3
A arri'B play agannein which their chancesof winning are in the ratio 3:?-
of winning at least 3 gamesout of 5 gamesplayed.

Solution
forl's *it"i"S =
Probability
J
Probabilityfor B's *irrrrirrg = |
IEt XNumber of gamesin #cn I wins.
PIX =rl=nC,P'qn-'

(?\*'
=sc-[1)"
'\s/ \s./
P[atleast3 gamesout of 5 forl's winning]
= plx>jl
= P(3)+ P(4)+P(s)
(
z\3 t\2 (
=r.,[;J . sc [1).(.?)*r*[1)'
[;J _\5/\si -\s/
= {1t0,. 27x 4 +5x 81x2+ 2a3l
5"'

= 0.6823
Example 4
Six dice are thrown 729 lines. How many times do you expectatleast3 dicc

Solution
LetX the number of times the dice shown 5 or 6.

P[5or6,=l*l=I
t663
12
;.p=1 ^ o 4 = 1
Heren=6
To evaluatethe frequencYof X >3.
By Binomal theorem,
/ r \r /,r \6-r
rlx =rl=6C,ttJ r =0,r,2,....,6
where
t;J
 Random Variables
1 .5g
p[x >3J= p(3)+ p@)+p(s) p(6)
+
/r\3 /e \3
=ac,l:
--'(r/I (:J
l: | +6cot:l t:l +oc
-^(t\o(Z\, ()s(2\ /r\6
\3/\3) 'li)l;J.uql.j
= 0.3196
.'. Expectednumber of times atleast3 dice
to show 5 or 6
= uxp[x>z]
= 729
x 0.3196= 233

Example S
In a precision bombing attack, there is a
50o/ochancethat any one bomb will strike
target'Two direct hits arerequiredto destroy the
the targetcompletely. How many bombs
droppedto give a99%ochanceof completely must be
destroyingthe target?
Solution
If z bombs are to dropped,then out of these
atleasttwo should hit the target to destroy
completely. Hencewe require it

p(2) + pe) +...+ p(n) >

0.99 where

P@ =(:)p' Qn-',x= o,t,2,...nwithp = - !

n
\r,/ 2
Hence 1- ^P(0)- p(l) > 0.99

_ /r), /r),
...r_t
;l_ ,1:l >o.r r
\- t \z/
n+l
:+ < 0.01
2',
:+ 100(z+t) < 2,
The leastvalueof z is 11.
Hence, 1l bombsshouldbe dropped,

Example 6
The probability of a man hitting a targeti.
. If he fires 7 times,what is the probabiliry
| co
of his hitting the targetat leasttwice? (ii) How
many times must he fire so that the probability
of hitting the targetat leastonce is greater tnan
lt
1. 60 and Random P

Solution
Probability of hitting the target p = 1
n4
:.q=l- P = j
Let Xdenote the numberof times of hitting the target.Here n =7

ByBinomial rlx = rl=7c,(i)' (;)'" ,r = 0,r,2,...,7

distribution,
(D P[hitting the target atleast twice]

=plx >21
=r-P[x <2]
=t-[p(o)+r(r)]
- [rr\' +71:ll
=l-ll:l -/ r \/:\u'l
:l I
L\4/ \ 4 i\ 4 i _ l
=l-toi3 ' = 0 . 5 5 5 1 (a p p \

(ir) Plhittingthe targetat leastonce]

=t- p lx = 0 l= l-q n

:.r-q'r?-t-(1\'
3
,?3
\4/
=r -?=[1)"
3 \4,/
/E \' <_I
+l_ I
\4,/ 3
)n-1,

Example 7
Commenton the following:"The meanof abinomialdistributionis 3 and
Solution
In binomialdisffibution,mean) variance.But variance< mean.
Sincevariance= 4 andmean= 3, the givenstatementis wrong.
 Random Variables
1 .6 1
Example 8
-. success*
If the probability or I
, how many bials are necessaryin order that probabilty
too
ofat least one successis greaterthan 1.
2
Solution
p, hobablity of success= -f-
100
g, hobabilityof failure= ,19 =0.99
100
kt therequirednumberof trials : z
By B.D. theprobabilityof atleastonesuccess
=t_eo = t _ (0 . 9 9 )'
Bythe givencondition,

t-(0.s9)"
>1
2
=+(o.r)".]
,2

:+ zlog(0.99) < log0.5

=n(T.sssa).T.oseo
< _t + o.ero
i.e.,z[-t +o.eeso]
i.e.,-o.oo44z<_o.3olo
+ \ o.3olo_.o t
tm=()6.4
't
...n=69
Example g

Ifxandrareindependentbinomial
variates
,[ti) finap[x+r=3].
\ z'/ ^anft,l),
\
sofution
X +Y is also abinomial variatewithparameters I
\ + n2 =12 and p =
2
...e[x+y=3]
J
=pc,r \fll'
2) f1)'
\2)

=51
2r0
1 .6 2 P ro b a b ilit y a n d Ra n d o m p

Example t O
Tencoinsare tossed1024times
and the following frequenciesare
frequglcres with the expectedfrequencies: obsened.

38 106 188 2s7 226 r28

5s
Solution
Here z = l0
p = the probability of getting
a head in one toss.
I
.'.p =; and,q =l_
O =:
Theexpectedfrequenciesaretherespective
termsofthe binomialdistribufim
The frequency of rheadsis ,oro.ro".( !)'*'f t )"
r =0'r'2'"''10
) l') '
Hence
wehave*"
""-o*;;-''"''lu
o | 2 3 4 s 6 --j-

106 188 2s7 226 t28 59

 Random Variables 1_.6g
Proof
Let the m.g.f of { be Mu(t). Then

Mx,(t)= ("' -r)],i =r,2,...,n

"*n[4
M*(t) = Mx,*xr*...*xn
(t)

= M*, (t)M", (t)......Mx,(t)

= -t) -t)....."t
("'-t
"a("' "u1"'
_
"(A+tz-.+U\ct-r)
="'("'-t="*r[,t(",-r)] where
n= 4+k+.....+4
This is the m.g.f of Poissondistibution with parameter).= xr+ 1, +.....+ Hence the
\.
result.

Solved Examples
ExampleI
If X and Y are independentPoissonvariatessuch that p[X=4=p[X=21 and
P[v =z]= r[v =3] findv[x-zv]
Solution
Let X be Poissondistibuted with paxameter2 and Y be Poissondistributed with parameter
p.

T hen n [ x = rl ] = p l x= 2 f= += " - r !
' l! 2l
:.2":2

E (X )=2andV (X )=2

P[Y=2]=ply = ! {:!
rr2 1 33l=r = "-1-l'
:.p=3.'.V (Y )=3
vlx-2",=111;,151
 L .70 P ro b a b it i and Random

Example2
If X is a poissor
thatefx = 21=sp[x=
4]+eor[v=o]'
E(x'). Arso
find;t"l! -ch
Solution
p [ x = rl= ' - ^ !'
' r!
.'.P[x = 2J=ep[x= a]+eop[x= o]
;!l-9e-AAa
= -2, *eo"-^tru
41 6r
= t=].t' + L
=,x"4+3)"2-4=0
:.(t",++)( * _ ) = o
:.22 = -4 or ),2=l

Hence )"=l (Since 12 carrrctbenegative)

Mean, ), =I .

E f*']= ).2+ ) .= 2

P[x>"=,-;t;;],rr,

=l- e- r [t+4= 1- 2
e
ExampreB
A manufacturerof cottonpins
knows thatS%oofhis product
is defective"
more
than4pinswlr be defective,
rrr
:T::::111,T_og*i""..:h1T,
mateprobability that a box wilr
fail to meet the guaranteedquality?
Solution
Here =
'o=j ,o; 'n loo
...A =np=5

r[x = =0,1,2,.... where

']=+,, X: No.of defectrr.c
 Random Variables L. 7|
r Probability that a box will fail to meet the guaranteedqualrty

[d =p [x>4 ]
= 1 -P [x < 4]
= t - [p(o)+ P(1)+ P(2)+ r(:) + r(+)]
l- .s
=1 -e -5 1 1+ 5+ 3+tzs 62s1
| 2 6 24)
= 0.5620
Ex a m p l e '4
A car-hire firm has two cars, which it hires out day by day. The number of demandsfor a
car on each day is disfibuted as a Poissondistribution with mean 1.5. Calculate the proportion
of days on which neither car is used and the proportion of days on which some demand is
refused.

Solution
x\No. of demandsfor a car.
Then X follows Poissondishibution with parametermean X.=1.5
Proportionof days on which neither car is used = p[X = 0]= e-r'5= 0.2231
Proportion of days on which some demand is refused
=p l x>2 ]
=1 -P [x<2]
= 1- [P(o)+ P(r)+ P(2)]
- 1- e-r'5
[t + t.s+ t.tzs]
= 0.1913
Example 5
Six coins are tossed 6400 times. Using the Poisson distribution, what is the approximate
probability of getting six headsx times?
Solution
1
Probability of getting one head with one coin :
2 (t\u r^
The probabilty of getting six headswith six coins = | : | -
\2) & r
Average number of 6 headswith six coins in 6400 tbrows = nP = 6400 x - = l9g

Mean of the Poissondistribution = ). =100

 P ro b a b ilit y a n d Ra n d o m P

By Poissondistribution, the approxirnateprobability of getting six heads

r[x=*]=+= gs11,x=0,r,2,....
Example 6
If X is a Poissonvariatewith mean - l, , show that n [x'?] = ,ae[x + r]
s(lx-rl)=2.
trrat\l l,f
e

Solution
^_t 2, =0,1,2,"'
PIX = r] =;,t
ByPoissondisffibution

Then E(X) = )'andB(X2) = ).2 + 1.

EIX+1]=E(x)+1
= 1+l
LE(X+l)= 717+l)= 72 a 7
"'
It follows that E(X') = 2E(X +1)

E(lx-tl)=i..
fr{t'-tt)
. l" -tl
=" - t ) l 't --t s i n c e 2 =I
vl
x =0

1 Z *....]
I+ o
= e_' l,[.
tt* a * zt I
i- -l
= -e r ll* t -
L fl--(n+1)!l
=![,*i (n+1)-1-l
el i* (n + 1 )! I

= ![,* i t-i l -l
el- 3.' ! ft( n+ 1) !l

= 1Jl*e- ( e- 1;l=Z
e' - e

Example 7
IfX and Y are independentPoissonvariates,show that the conditiond
given (X+Y) is binomial.
 Random Variables
L. 73
Solution
Let X be Poissondistributed with parameter 2 and Y be Poisson
distributed with parameter
p . Then X+Y follows poisson distribution with parameter
).+ p .

r[x=r/x+y=n]=t[l==:^f=n_-4
P[x+Y=n]

=.lI='Jp[y=n-rlsinceX andY areindependent

FfX + y =;]-

"-LAt "-pOn-t
_ r!(n-r)!
+
"_<t*r,)7,L lty
-i+ n!

= ot (2)'f o.)"-.
r! (n -r)! [ , ? + p )ln + p )
.'.f[X =r/X +Y =n] =n C, p ' q o wh
-, e re ,= q - I -p = - ! -
+Aa+npd
X+p
theconditionaldistributionis a Binomialdistributionwith parameters
'"';" n andp where
-v =-
- 1 + p'
il
I
Example 8
A randomvariablehas the m.s.f. . r'r
"*o[4, ' Findits mean'standard
deviation
and
p[rt-2o<x<p.r";;;r;:;"":t!1t;. ')J

Solution
Sincethe m.g.f. of a poissonrandomvariableX is exn[;(", _ r)],
we get ). = 4.
=
.'. Mean Variance= 4
* F=4ando=2
P[p - Zo <X < p + 2ol= efg. X. S]

= p ll< x< 7 J
= " -o*f+
* * + * !* !* !* !]
L zt 3! 4! sr ' at' tr l
= 0.931
 1 .74 Probability and Random p

Also,
Plr - 2o<x < ttl=p[0<X. 4J
= p[X= 1]+plx = z]+e[x = :J

="*[+*s
*'2f
L 3J
= 0.4131
.
Example g
In a book of 520 pages' 3g0 typographicalerrors
occur.Assuming poisr
numberof errorsper page,frnd the probability that
a randomsampleof 5 pae:
etTor.

Solution
Averagenumberof errorsper page = 390 =
0.75
s20
=
:. )" 0.75
Let X: Number of errors per page

.'.P[x= rf=92; =0,1,2,...

P[No enor in a page] = p[X = 0]

=g- 1 =" 4'75

=0.472
... probability
thatthereis no errorin 5 pages= (0.+lZ)t = 0.0235
Example I O
Supposethat the numberof telephonecalls an operator
receivesfrom 9:00
a day follows poisson distribution with mean --
3. Find the probability tha:
receivesno calls in that time interval next day.(ii) In
the next three daysthe ope
a total of one call in that interval.

Solution
Let X: The numberof phonecalls receivedby the
operatorduring th*.
Then X is poissondishibutedwith mean Z = 3

.'.P[x=r]= =0,t,2,...
+,r
 Random Variables 1. 75
(i) no calls] =P[X=0]="-'
P[operatorreceives =0.0497
(ii) Probabilityof receivinga call = 1- e-3
The operatorreceives a maximum of one call in the time interval is describedas below.

Day First Second Third

No. of I 0 0
Calls 0 1 0
0 0 I
.'. The requiredprobability = r("-' - t'
)' (t )
Example | |
Fit a Poisson distribution to the following data which gives the number of yeast cells per
sqrxre for 400 squares.
No. of cells 0 I 2 3 456 7 8 910
per souare (x)

No. of Squares(f) 103 143 98 42 842 0 0 00

Findtheexpected
frequencies.
Solution
te, {ro
_ "o. =1.32
Mean = 2L
Ir 400
The parameterof Poissondistribution )" =1.32 .

poissondistribution is N " '2' = 4goe-' " (1'32)'

r! r!

_ 106.96(r.32)'
r!
(1'3?)'
The expectedfrequenciesare the valuesof toe.se, when r = 0,1,2,.....,10
r!
The expectedfrequenciesare 107, l4l, 93, 4I, 14,4, l,0,0, 0, 0.
 Random Variables 1. 105
Proof
rhep.d.f,of X is f(x) =
{^\^
"o|#r"
..'plx
L
t t]r = ['r"-'rxdlx = [-"'rxl-
- -
Jt L Jt "-rt X>s+t

P[x t:?:+:t t]
-r =
.'.P[x > s+ t/X t r] ^-x?
e[x> s]

- "-1:;"=e-tt=r[x>tJ
I 'l
e-rs
Another Form
If x has an exponential distribution, then for every constant a)0,
P[Y < x/X >u]= plx < x] or all x, wherey = X_a.

Note
The converse of this result is also true. i.e., if
P[X t s + t/X > rJ= p[X > sl, then X fottows an exponential
distribution.

Solved Examples
ExannpleI
If X has an exponentialdistribution with mean2, find p[X . I/X <21.

Solution
If X hasparameter)", thenitsp.d.f.is f(x) ={1"-u x >0
I o other"wise
Mean=1 =2 = ).=l
12
Ir -+
" x>0
f(x)=11
otherwise
 and Random

p lx < r n x < z ]
P [x. u x< zl=
e[x < z]
a1 -I
p lx< r l I;" ' &
_ - "L

f[x<z] ez1
t" 'd*
I
/ -;l
* \r
|
t-e'l
_\ /._l-"''

[-"'J,
( - +' ) ' l- e- '

.'.P[x < l/x < 2l=9!2!= 0.5686

' 0.6321
Example 2
A continuousrandomvariableX has the p.d.f. f(x) given by
l -x
f(x)=]Ae5 x>0
I o otherwise'
Find the valueof A and compute(i) p[X <6/X> 3l (ii) p[X > 6]
Solution
@x
3d"=1
f,r1";c*=t+Af,e
=+-SA[O- t] = t

.' .5A=1= +A= 1

5
-:
=[t
...f(x) x>0

tt; x<0
 Random Variables
1. 107
36
. e s _e 5
=_
-: J

e'

e-{.5 _ e_r.2
et '

_ 0.5488_0.3012
-
o.sAt8
p[x<6/x>3f=o.qstz
P [ x ' 3 J - r - p l x< 3 J
=r_ [r1x;ax
&
= e{6 = 0.54gg
Example B
that thc life ofan industrial lamp, in thousands
. .t]o:*" ofhours, is exponentially dishibuted
wrtn rarlure rate )" =i (one failure every
3000 hours on the average).Find (D the probability
that the lamp will last lbnger than its mean
life of 3000 hours (ii) the probability that
the lamp will
last between 2000 and 3000 hours and (iii)
the probability that the lamp will last for
another
1000 hours given that it is operating after
2500 hours.
Solution
Let X: the life of the industial lamp in thousand
hours.
Then X is exponantiallydistributedwith parameter
). =!.
fr+ 3
.'. Th e p .d.f.
ofX is f(x)={l'e' x>0

I o x<o
(, P[The lamp will last more than 3000 hours]
= PIXt 3]=t -r[x <:J
=r-fr(r)o*
=1- f1"-i6*
{3
| 3l

=t-ll-"' l="-'=0.3679
t_J
 .1 0 Probabi and Random

(il) P[The lamp will last berween2000

and 3000 hours]
=p[2<x<3J
= F(3)-F(2)
( =\r -z)
=ll-e, l-l r_"TI ,u
( /( .,
-2

=e 3 -e-l
= 0.5134-0.3679
Requiredprobability = 0.
1455
(iii) Probabilitythat the ramp
w'l last for another1000hoursgiven
2500hoursis p[X >3.s/x>2.sf=p[XrrJ
oy*"_o!,rrr.
...p[X > 3.5/X> 2.sl=l_ plx < tJ
[ r
= t- L t- e-rl
,J= e -=I O3 r 6 s
Exampfe 4
The amount of time that a camera
wilr run with out having to be
rese
wrth 0 = S|days. Find the probabilitl.
li:::^.:r"""ltial,distributiol
(i) have to resetin ress than20 days. * t,t

(ii) not have to be resetin at

reas:
Solution
Let X: the time that the camerawill
run with out having to be reset.
Then X is a r'v' with exponentialry
distributed wit,. e= 50 days.
The:
f r + x>o
f(x)=
lA"'
0
L x<0
(i) P[camerawill have to be resetin
lessthan 20 days]
= pfcamerawill
run for less than20
=plx <2oJ
=Iezo 7 -x

,i" "a*
= l=
11-1e'
l=0.5297
LJ
 Random Variables
2-6
(ri) P[camerawill not have to be reset in at least 60
d.y=]
= p[camera will run for atleast
60 days]
= P[X t 60]

=f1"#a*
60 50

= = 0.3012
Example 5 "f
The daily consumption of milk in a city in excessof 10000 gallons
is approximately expo-
nentially distributed with parameter g=1000. The cify has
a daily stock of 20000 gallons.
What is the probability that the stock is insuflicient for both
days if two days are selectedat
random?
Solution
If the randomvariableX denotesthe daily consumptiono-fmilk (rn gallons)
in a city, then
Y = X - 10,000hasexponential distributionwith parameterB = 1000.
Thep.d.f.of y is g(y) = <y <a
#"*,0
sincethedaily stockof thecity is 20,000gallons,theprobabilitythat
ttrestockis insufficent
on a particulr My = p[X > 20,000]
= PIY> 1o,ooo]
= l.
lo.*s(r)dv'
&t4
= | '
.,ro,mo ermdv
1000
[ -r -l-
l-"'*
L
|
jo.ooo

= -[o - =
"-to] "-to
The probability^that the milk is insufficient for both the days if two days are
slecErl atwrdqn = (e-to =
)' "-zo
Example 6
Supposethat during rainy seasonon a fropical island the length of the
shower has an expo-
nential distribution with paf,ameter2' = 2 ,time being measuredin minutes.
What is the probablity
that a shower will last more than three minutes? If a shower has already
lasted for 2
what is the probability that it will last for at least three more minutes? -irr.rr"r,
 1. 1 Probability and Random

Solution
Let X: length of the shower.
Then X follows exponential distribution with parameter ,r = 2 .
t l+
of X is f1x; J;e '
P.d.f. = x >0
...
L 0 otherwise
P[ashowerwill lastmorethan2 minutes]
=f,r1*;a*
F2
=l-
I f(x)dx
=r-1 f .Ta*
2h

/ =r)'
= l- l- e ' I
tt
\ ,/o
= e-f = 0.3679

P[the shower will last for at least 3 more minutes given that
it
= P [ x> s t x > 2 7
= p lx > 3 ] by memoryless
l-X
a6 |
- | -:e2dx
,tl 2

rf@

=l*t I
I-xl

tt
LJ 3

-3
=ez =0.223I
Example7
If X is a continuousrandom variablewith p.d.f. f(x) = €-*,0 ( x < €,
that the roots of the equation t2 - x(4t - 5) + 2l= 0 may be real.
Solution
The equationax2+bx*c =0 hasrealrootsif b2 _4ac >0.
.'. For real root of the equationt2 - 4xt + 5x + 2r =0 , we musthave
t6x'z-4(5x+21)>0
i.e., 4x,-sx-zl> o
 Random Variables 1 .1 1 1

5t Jlot
solving 4x2-5x-21=0, we get * -
8
5+19
-x= 8
-I
...x=3 Or-4
.'. The roots of t' - 4xt+5x + 2r=0 maybe real for x >3 (x cannotbe negative)
Requiredprobability= f [X > :]
r@_".
=tr e'-{ t ) (f = L -J 3
"_*l-
= e-3

Example 8
A component has an exponential time to failure distribution with mean of 10, 000 hours.
(a) The componenthas alreadybeen in operation for its meanlife. What is the probability
that it will fail by 15,000hours?
(b)At 15,000hours the componentis still in operation.What is the probability that it will
operate for another 5000 hours.
Solution
[,et X denote the time to failure of the component.
Then X has exponentialdistibution with fir€an = 10,000hours.

.'.1=10.00O+)"
). 10,000

Thep.d.f.of X is f(x) = - L"ffi,x ) 0 and: 0 otherwise.

10,000
(a) Probabilitythat the componentu/ill fail by 15,000hoursgiven it has alreadybeenin
operationfor its meantife = P[X < 15,000/X> 10,000]
plto,ooo<x<ts,ooo]
_
P[x>ro,ooo]
rd5,000
f(x)dx
=- 1..
f,.*r1"P"
e-l _ e-1.5 0.3679-0.2231
e-r 03679
= 0.3936
F

1 .L 2 0 P ro b a b ilit y a n d Ra n d o m

DistributionFunction
F(x)= f,r1*P*
= apx\ae-o*r dx
f
Put u = axq apxq-tdx
-dv-
F(X)= f,./ e-"du
= =l - s-ax|
[-"-']0"'

=+ F(X) = P(X *) = l-e-o"e

'
Note
(l\ From distribution function, we have

P[x<ul=t- and PIXt

e-o^P af="-{

(2) A r.v. X follows Weibull distribution witk

a,F>0 if the r.v. Y =aXP follows
distribution with densityfunction f"(y)=e-'

(3) When f =1, Weibull distribution reducesr:

distribution with parameter d .

Solved Examples
Example l
Supposethat the lifetime of a certain kind of an emergencyback up
r.v.X havingtheWeibull disfibution with parametera = 0.1 and'B = 9.5
time of thesebatteries. (b) The probability that such abattery will last

Solution
For the p.d.f. of X in the form f(x) - dpxf-1"-"^p,X) 0

='7.l,!*t]
Mean
\.8 )

(a) Meanlifetime= r =(o.r;;i.(t.*)

 Random Variables
1 .1 2 1
=10.r)-'f(3)=
roox2
=299 hours
(b) probab'ity that
abatteryw'l rast rnore than
300 hours
= p[X > 3oo]
=
f;or1*;ax
=
f*apxo-te-o*f dx
=
f;oto'rxo'5)x{'5"-{o't)*os*
Put y-(0.')xos
Thendy = (0.l)(O.5)x{sdx

.'.PIX>300J=
f"-ay wherey, =(0.1X300)0.5
= g-Yt -
"-{o.txroo)os
= 0.177

Example 2
Supposethat the time to
failure of a certra:
asar.v.having
weiburr
distribution isconsidered
*,h oJ::,T:T:3:T"f':Tnutes
torast?
(b)what
is*,"p,ouuuifrylii,".,n;ffi;:::ff
;::ffi"liTfcted il1;
Solution
(a) The expectedtime
to last for the component

=E(X)
= r(; . r)
"r
= 1o.z)-'f(4) = t25x
6= 750mimrtes
(b) Probabilifythatsuch
a component
willr rqu
fail r'
in ress
lessrnan
than10hours
ho = PIX < 600J
we know plx = a!= 1-
"-,a
.'.r[x < 6001=
1-.*,'*,i
= l -s .r' 6 e = 0.8149
 ProbabilitY and Random P

ExamPle3
a r'v' nar
(in hours) of a semi-conductor is
Supposethat the service life a
p --0'5 ' What is the probabilrty that such
distribution with a = O.Ort and
4000 hours?
will be in working condition after

Solution
Let X: the service life of the conductor'
denst:1
with parameter aandp and whose
Then X has Weibull distribution
f(x) = aPxf-t s-o*P'x) 0'
=0'5
Here a --0.025andP
after 4'000 hours]
Pfconductorwill be in workng
= f r1x;ox=P[x>4ooo]

We know that P[X 2a]=s-"'P

- ozsx+ooo).s
.'. Requiredprobability "-{o
= s-rsr =0'2057

ExamPle4
of carn* U:TO*""
If thelife X (in vears)of a certaintvpe " Y"l:lt that the life
probability
value of the parametera ' giventhe
0 =2 ,find the variance
of a and p ' findthe mean and
5 vearsis e-0" . For theseuulot'

Solution
grven bY
The densitYfunction of X is
sinceP=2
f(x)= Zdxe-o",x>0

Now P[xt sf= f,zaxe-o*'dx

-"(t'I
=g

= g-25a

= e 4 . 6* o = #
. , . e -rro
PIXt5]="'"
Giventhat

FortheWeibulldistributionwithparametersaandp,

v(x)=
-rrli.r)*a
E(x)= "flr(X.')-{'
 Random Variables
1.L23
For a=# *0, =2,thevaruesof meanandvariance
ull areobtained
asfollows
lor / r \; /r\
=[r.);.l.r)
Mean
\ 1 0 0 / \2 )
= t 0 , t-ft)
.-rl /r \
u sinrgl :l= G
; l =sJr \z)
' :"'
=
variance -
[#)' |", {,[;)]'l
| / r:,2 1
=roolr-lgI I
I (2Jl
=toolr-al
L 4l
Example 5
Each of the 6 tubes of a radio set has the life length
(in years) which may be consideredas
a r.v. that follows a Weibull distribution with parame
ters a = 25 arrif = 2 . Ifthese tubes
function independently of one another, what is
sr the probability that no tube will have to be
replaced during the first 2 months of service?
ils
Solution
Let X: life length of the tube
Then the p.d.f. of X is givenby f(x) - apy!-t"-o,r,x
) 0
i.e., f(x) = 5Oxe-rt"',x> 0
P[a tube is not be replaced during the first 2 months]
I ll
=Pl Xt:
-i

ol l
since2 months=1u*
L 6r-*

f,soxe-rr",dx

=(-"-"*'):
\/f
="#
... P[all 6 tubesarenot Ue,eptacetduringfust 2 months]
( -zs\6 -2s

I t ' I = t T = o' o lss

\/
r.L.24 Probabilitv and Randorn P

Example 6
The life time x inhours of a componentis modeledbyaWeibull di
Startingwith a large number of components,it is observedthat 15 o/oot
that have lasted 90 hours fail before 100 hours. Find the pararrrctera

Solution

P.d. f of x , f (*)=aP xQ-t e-o" , x>0

=2a x e-o" ,x>o

Giventhat P(.r < 100I * r%)= 0'15

P (90<.r < 100)

. = 0.15
P(x > e0)

t00

lza * e-o" dx
+ eo' =0.15
€
lza x e-o'- dx
90

put u=d*2 du=2axdx

r . r l ( X)

l-"-""
\ l^^
rm o-Nro_ o-tu.zo
' =0.153' :o =Q.lJ
/ r \@
l-"-"-
\ /S O
: " - * 'o '. : : , ,, r::i ;:. r':. : -::

'
This gives a =0.00008554.
 - ProbabilitY and nendarn
D

= x is definedby
The conditionalmeanof lgiven X
E l Ytx= xl= L Yf( r tx) d t
given X = x is defined by
The conditional variance of lis
vLY I x = xl=EIY't x = x7-\rlYt x ='l\'
Independence
Stochastical of RandomVariables
TworandomvariablesXandlaresaidtobestochasticallyindependem
rlX =x,aY= Y;)=flX =*'l'elf = Y'f
i.e.,Pii = (A XP , ) for all i andT

Incaseofcontinuousrandomvariable,therandomvariablesXandY
f (*,r1=g!\h(t)
I
where f (*,Y)= johlP'd'f of Xand
X and
s(x) = MaryinA densitYof
n(l)=Marginal densitYof I'

Solved
- ExamPles

ample I
distibution obtain
Given the following bivariate probability
(i) Marginal distributions ofXand I
Y=2 and
(iD Conditionaldisfiibution ofXgiven
1'
(iii) The conditionaldisnibution of IgivenX=

-1 0 1
J rtx->
t21
0 15 t5 15
321
1 15 15 15
2t2
2 15 15 15
 Two Dimensional Random Variables

So l u t i o n

J T\,[-+ -1 0 I 2pQ,v)
121 4
0
15 15 15 15
321 6
I
15 15 15 15
2t2 5
2
15 15 ls 15
Zp@,v) 6 s4
I
v 15 15 15

(i) Marginal Distribution of XandY:

rfx = x]=ln(x,t)andp[r= yf =ZpQ,y)

v

r. I 0 a
_l

654
r(x): 15 15 ls

v. 0 I 2
465
n(t): 15 15 ls

(ii) ConditionalDistribution of Xgiven Y = 2:

rlx =xlY =2]='l*:=*rnt-='l where

.r=-1,0,1
' PLY= 21

efx=-r/Y
=r]=?',!.?
'
=?
5/15 5

elx=o/Y:r1=#=i*o rlx=r/y=21=?tr!=?
;/15 5

x 0 I

P (X /Y =2): 21?
1'!\

I
1

*
 P ro b a b ilit Ya n d Ra n d o m P r

(iii) ConditionalDistribution of fgivenX= |

=yrx =tS=W
PIY v =s,1,
for 2

Y: 0 1 2
112
P (Yt x = r ) : 444

Example 2
l-ntX andlbe two randomvariableshavingthej oint probability function I i
marpsnEl
where x andy can assumeonly the integer values 0,1 and 2. Find the
distributions.

Solution
IX \I -> 0 I 2 I n(*,t)

0 02k4k 6k
1 k3ksk 9k
2 2k 4k 6k tzk
3k 9k lsk 27k
ln(*'t\
ZZpQ,v)=r
yx
> 27k =l

:.t = L
27

(i) MarginalDistributionof XarrdY

x 0 1 2

69 1 2
p(,), 27
27 27

(ii) Conditional Disftibution of X gSvenY = y

p I y -x n y = t l
plx=xty=tl='ffi
 Two Dimensional Random Variables 2. 8

* I x\r+ 0 I 2
lj
24
0 o-9- l s

* I 115
1 391 5
J
2 z!6
391 5
I
(ii) Conditional Distribution of lfor given X = x
1
{ x\r-+ 0 I 2
24
0
66
I 35
I
9 99
2 46
2
12 L2 T2

Example3
Two discrete random variablesXand lhave the joint probability density frrnction:

, \ A,'e-x. pv (l- o\*-Y

Plx,Y)
11; , -, v=0,1,2,...x; x=0,1,2,....
where ),,p are constantswith 2 > 0and0 < p <I .
Find (i) the marginal proability density functions ofXand L
(ii) conditional distribution of rfor a givenxand ofxfor a given r.
Solution
(i) Marginal probability functions of X and y.

e lr - P)"'
Pr $) =t P@'')- f ')"yl P'(t-y)t
.v=o ,l "^

-e - ^ + ' ; " - x t _p , ( r _ p ) ' - ,

yl(x- y\(
x'! ,=4,
 Two Dimensional Random Variables 2. 10
4
,Example
J If X and f are random variables having joint p.d.f
To
11.
x- Y )'o <x< 2' 2<Y<4
f(*,Y )=];tu -
f0, Otherwise
Find(i) r[x<tnr<:] (ii)plx + r < 3]and (iii)PIx<r/Y<3f
Solution
(i)r[x <rny <t]= f,t f e,il dyd,c

=*f fru -x-v)avdx

=*l[t'
-.\r-+];,*
=!g l(u-'-t)*
"\ 2)
tlt
= - l- x --
*' l'| =- t
tlt___ r l|
8L2 2 )" 8L2 2)
;.r[x<tnr.:]=i
(ii)r[x+r <:]= f f ']tu-" - y)aya,
=*I[('-a,-lf)','
*
=*f[,.-')t,-4-Q:)' .rf*
=*{'-7 x +x 2 r y ]-
=!ls*-t*'
*4* (:-')'l'
8L 2 3 61.
=1[3_Z*1*1_2.l
8L 2 332J
=f;ltn1=a
 a n d Ra n d o m p

(iii)Plx < | / Y4l = P[x=;lr-'I,' 3]

ply 4l

p(y4)=
f fifr_ x_y)ayax
=*rlru--\-(2-r\1*
801. - ) ) *"
' \z
=* f ')0,=]tz- zt=
[1- |
. . . p [ x< t /y< r ]= # = ;
Example5
If thejoint p.d.f of (X,y) is givenby (r, y) = 2,0 <
f x< y < 1, find
O M arginaldensiqz f,urctbns of X and.y .
(ii) Conditionaldensities (x I y)andf
f (l / x'.
(iii) ConditionalVarianceof Xgpveny=2.

Solution
(i) Marginal
densiry
tunctions
fr(r) andf"(y,):
f*(*)=
Lf e,ia,
= !)ro,=z(r-x),os.r
s1

f'(Y)=ff (*,Y\a,

= ( r *= 2y,0<y< r
(ii) ConditionalDensityFunctions:

f (x/t)=!kl=Z=L,0.*.,

-r(trx)=Hl =ii= fr ,o3x<vsl

 Two Dimensional Random Variables
2. 12
(iii) Conditionalmeanof Xgiven y = y

zfxry]= fx|(x/y)dx
( t'\
= | rl - l trx= -lrf x'f' = !
" { \y ) -l
y L z l" 2
nlx'/y)= =4
(x,f(x/y)*=1f+l'
/L 3J. 3
Conditional
varianceVfx / yl= Elr' t yl-
{nfx t yll'
_y' _y t _ y '
341 2

varianceofXgiven y =z=tl
Conditional =1
l2l'=' 3
Exampre6
It f (*,y)-eQ*'l,o<x,y<oand=0, elsewhereisthejointp.d.fofrandomvariablesX
,/ andI, find(i)r[x<r] (ir)rlx >rl Qtlrlx+r<\
Solution
(t P[x.4=Xfrn"*idyd*
=f"-'lf, I
ro ta
"dl la'
=I'-.l-"-li*
=
[e -,d x=t- L

(ii) P[Xrr]= f f e-@t)ayax

=I'r-'l-"rli*
=f"-.lr_rb
= l[ -e --' + - "-t*1-
l
L 2 JO
,l t
=l--=-
22
2 rL Probabil and Random P

(iir) rlx +r <1l= ll-' e+rafix

=J"-'l-n'I^'*
- ."')a,
= X"--[,
"-t
=
fe*ax-at ldx
=r-?
=(-t")i-a'(')i
Example7
TWodimensionalR.V (X,Y) have thejoint p.d.f
f (r,Y)=8r/,0 < x < Y <l and = 0,elsewhere
I r rl
(i) Find Pl X .; y.; (ii) Find themarginalandconditional
I z^ +J|
(iii) Are Xand lzindependent?,

Solution
l- r tl et4ey -,
(D P l x .;nY
Z
.;)=
+J f, | f(*,t)dxdy
L

= l'^(uyarat
tatt

=Jf'osrl
0 '1 . )il |
a,
L-Jo
=4I'oy'dy=(r^)'l'
=1
256
(ir) MarginalDensities:

g(")=lf (*,flat,=f3*vav
= q x ll-r' ] , 0 < x < 1

ff (r,ila*=('uva*
n(v)=
=su[4.l'
' l- 2
l,
'=4Y3ro<.Y<1.
 Two Dimensional Random Variables 2. 14
ConditionalDensities:

! : ' ! ) = 4y'
f ( x /.,y l = fn(y) \.\ = 4 ,0
y2 ' -
.x<
"- '/y< r

2Y
=;44=fri,0<'x<Y<l'
' *)=
f(trx)='!:t()=-P ^"'.
s(4
(ur) h(y)=4x ' f (*,y)
g(r)' ' '
,ffi*
:. X andY are dependent.

Example 8
The joint p.d.f of the two dimensional RV (X, f) is given by

I ^, )=ll tts .'L1 < .<r y s z

T f(r,y
L0 ,elsewhere
(i) Findthemarginaldensities of XandY.
(iD Find the conditionaldensitytunctionsf (x / y)nd f (y / ,)
Solution
(i) MarginaldensityfunctionsofXand I:

=f,lro,=l,lt)',,
s(.r)
=*G- r ,) ,r < x< 2
9\

h(y)=
fl,*=Yltli,
=* b , -t),t<y<z
9\-
(ii) Conditional Density functions :
,

r (*ry)= = <xsy<2
# 4 U=h,,
8
-w
r(yr*)=H= = # ,r a x< y< 2 9-

!{+-.')
 2.tL Probability and Random

ExampleI :

, The joint p.d.f of two randomvariablesX and I is

/ ,. , \
r/,, 1ft"(r-y),' o<x<2,lyl<*
J\x,y)=1^ , ,
|.0, elsewhere

Evaluatethe constantfr. Find the marginal and conditional density of i'

Solution
lyl< * - -x < y < x .
(D To find fr

t f"u(* -w)dydx=t
?2^'?2
=r
t I x' (2x)dx- k I xlO)dx
- [-ol '
:.Z klLl =r
L4 1 .
.'.8k= 1+ fr = 1
8
(ii) Marginal densitiesof X arrdY :
P
g(") = lJ ^t\x,YYY
1
: -'Y)aY
8 J*(t'
=" r dv-!f vdv
gr' - gt" -

=t-lul,-{(o)=4
8 L 'r - x 8" 4

.'.The marginaldensityof Xis

1

g(t)= L'0t"2

r'(r)=Lf @,y)a*
:. h(y\
\r/
=l f G' - ry\d*(-z<.y
-,
<o)
gty\
1
=. ll(*' - *y)dr(o
andh(y) '/
. * .2)
8Jrt
 Two Dimensional Random Variables 2 .L 6

_,(*)1,
: n(y)=*[(+)",
,]
=lfg*Y'-zr*t1
813 3 2J

=:-t..ff-r"-2<y<o

andh(v)=*[[+),
,(+)")
=llg-v' -zu*Lf
8133 2)

=:-I*fi*o<y<2
.'. The marginal densrtyof Iis

l!-+.*in-2<y<a
o o:
r(y)=l
?
<v<2
L;-i*fri"0
Conditional density function

f etx)='!;:P
' glx)
I(r- u)
;.7 (t,/ x)=U;# = I ! 1x, A<x<2
#,-*
10
Examnple
Thejointp.d.f of R.V.sXand lis givenby

>o'Y>o
f (''Y)=lj"-'''"-';x
L0 ;otherwise
nina^e[X>t/Y = t]
 Two Dimensional Random Variables
2. 22

:.fr(*)=]:"-r,,1+
,l2z hf "+*lj
l.l2zr
ByAreapropertyof standard
normaldistribution,

* f e-"''dz-l
42r k
f- (') = l"-" '',-@ (.tr< cc.
"'
'J2z
Marginaldensityof I:
Writtingx' - 2pxyr y' = (* - pyf * (I - pr)y,

tt
-u2 -rl$-orfl'
r'(il=ffiL"'tGla*
/\2
r r l l '- PY l
=_!
,l2n"-r,,*JZr r*
f !"-ulE)
Jt- p'
a,
Usingz =#, weget
!r- p'
=_f,"_,,,t*1e",,,
f,(y) a"f
=ir"_r,,*
h fe-,,,d2=l
.'.Marginal densityof Z is

=,ftr"-"'',-@
f,(Y) (/ (o
p=o,f (,,y)=**l-VPl
when
t [ (x ' + v ' 1 1
Alsof,(-)f,(y)=i"*pl _r- =, ) l=f G,y\
zlTlz

.'.If p = 0, X andy areindependent.

.,,Example 17
/
' The joint distribution function of two random variables
X and,r is given by
F(*,y)=l-e-' - e-v+ e-6*t),x>0, y> 0 and=0, otherwise.
Findthemarginalandconditional
densitiesofxand r. Are xand rindependent?Also computep[X<
t,y < t] .
 P ro b a b il a n d Ra n d o m

Solution
Thejoint P'd.ff (x,Y)=W

'.f (*,y)=*[t -e-'-"-t *"-('*ilf

oxoy -
=*1"-'-"-(*r)]
=
"-('*t)

rhejointp.dif (x,yr=[;u.'
:ffi:*:"t
Marginal densitiesof X and Y:

g(")=f t@,iar= f,e1*idY

= = >0
"-'l-"-'f] "-',*
lf (*,fla*=/f,e-'dx
h(Y)=
="-rl_e-,fi
=e r ry>0

Conditional densitYfunctions:

f ( t lY ) = H= e-"x>0
nQ)

f ( y r x.) = ' ! : * ) = e - ' , v > a

s\x)
Sinceg(x)ft (t)=s-@'l = 7(x,y),Xancl'Iare independent'

= ll"-<*'tafi*
P[x<1;r<1]
=(f"'*X|o'or)
=0-""X,-"')=(t-J)'
=0.3996
 Two Dimensional Random Variables
2. 42
Note
(I) For finite numerical values of X and
I the formula for
the co-efficient of correlation is

l\-
:Lx"Y-xY
fo =
(or)

l;>,o_* ltlf _v,

f=

nZv'-(Ir)'
(2) The co*elation co'efficient between'
two variabres is
unaffected by a change of origin and scale.

(I =ffora,
i.e.,If =+,thenrn=r- andis givenby

ut(3\- Solved Examples

Example | ,/
Ifxand rare independentrandom variables, show that
they are uncorrelated. Is the con-
verse true? Justify your answer.
Solution
IfXand Iare independentvariables,then
cov(x,r) = E(xy) - E(x) E(y)
= a(x)n(y)_ E(X)E(Y)
-0
...correlationcoeffrcientr =cov(x,Y) -

+ x andyareuncorrelutec.
gffi converse
is nottrue.
*The
variables may be uncorrelatedbut they may be dependent.,,
 .4 3 Probabil and Rand

Example
,Kis a variablewith densiry
f (x) =1 ,1,1. I
Let Y = X2. Xandlare dependent
u"riait"..

E(x)=
I,i*=i*l'=o
- L+ J_r
.'.E(x)E(y) =lE(y) =s
FurtherE (Xy) = t(x,x,) = E
(X') = o
.'.cov(x,r) = E (xy) - E (x)
E(y) = o
...X andy are uncorrelated
but dependent.
Elampte2
,/
\'/ Find Karl pearson'scorreration
co-efiicient betweenxand
rfrom
x78 89 97 69 59 le
79 61
61 6
6t
Y, 125 n7 156 112 107 n6 n3 lOi
Solution:
IEtU=X-75andy=y_125.
Tlten rr=r*

Wehave
2u =-7;2, =4;Zu,=rrr@
Correlationco-efficient

i' f^ , =., 12236

-=o'917 Hence6=0.917
'
 Two Dimensional Random Variables
2. 44
Example 3 ,/
A computerwhile calculating r, from2lpairs
of observationsobtainedthe following
n = 25,1x = 125;Zx' = e's};Z = ioor2y, =
stants: con-
! ouo*i I r"!= 508
A recheck showedthat he had copieddown two pairs,(6,14),
(g,6) while the correctvarues
were (8,12), (6,8). obtain the correct
value of correlationco-effrcient.
Solution
Correctvalueof = 125_ 6: g + g + 6 =125
I"
Correctralueof = 100- 14_ 6 + 12+g = 100
I,
Correctvalueof = 650_ 36_ 64+ 64+36 = 650
Z*'
Correctvalueof = 46A_l-96_ 36+ 144+ 64
Zy,
= 436
Correctvalueof =50g- g4- 4g+ 96+ 4g
1i |ry
= 520
.'.Correctvalueof co_efficient
of correlation
NLry -
n2*'_(2.)'J"
Zv'-(Zi'
l25x1o0
2s)(6so)-(rzs)'2s)($6)-0oo),
r 0.667
Example 4
If o',,oj and,ol_" be the variancesof x, y andx _y
respectively, show that
o?+ o'zr__:1-
."y =-----
zoxay
Solution
ol_, = var(r - l) = vm(x)+ var(y) 2cov(x,y)
-
= o: + oj - 2ro,o,
,'.2ro,or=o:+o1-ol-,
or_+o?_o,
/ x-v
:= '
-=
2o,6,
 L45 Probability and
Random

,/ExamOIe 5
J The followin table
gives the number
weights lpouna.yg of studentshaving
different

50-55
55-60
60-65
65-70

Calculateco_efficie
._--"- vvrrvr<ruun
Derween
heightsand weights.
SOfUtiOn
Let x andy be the variates
f^r rh-
atesfor the ,r.^:^,-.
weightsand heights
respectively.
rnt u=x-705 and
v=Z:g
l0 2.5
From the table, we
have

2"t" =37;2fu'=t%;lfuv =2s;>.fv

=tq
Correladon
1fi,= 404andy
co_efficient

fr y= luu=

.' .t=
roo(tn) _e7), o(4oAE
,,1
/
:0.09
 Two Dimensional Random Variables
2. 46

6l (\. c- t*
I
c\ c.l

T> 6l trr ra) rt

\o e.l a.l € o
.<F

$ F.- .+
r)I ail a.| \o
I

oo c..
$*'
F
c.l c.l
c.- c.|
(r.l (\
=
a+)

o
C-l
c.l t\
e @ @ A
\:,
\o 6l
tfl
s
\o a"l
{;,
G'

o o)
(\ + c-. cO

U
c)
H

aH
6l
I
e o /A
\7
@)
\y
oo
a\
oo
6.1
oo
a.l
\o
I
6

n)
I
\o C)
E
o
l!)
A
t*
g a

(u I o o
\:7 @ tal

t\
O
E
q>
L) o c71
a+r tr- o (\ E
o a
o

o
o
o
I 6
e e e A
LY la:
a
!1,

an
|u
o\ cal
I

cal
L) (U'

o\
oo
€ C.l
I
o c o t? s € \c) oo
€6t

o
()
o
cit

iI m
E
t cf}

---+

!n q?
n n
z- to€ (\.l c- c.l
\o
tr
\o

iI o
X
I
\c)
I
\o
I

\c)
c-
I
r/)
\o
E*" €. \
I
.s

\
 Two Dimensional Random
Variables
2 .6 6
slopeof (4=*:where r is thecoefficientof correration
between.r uody.
Takingry=r2ond mr=9,
o, fo r'
frr-\
ta n a =
I + 4 mz

!o' - ro'
ol
=' =o,
:*t o,'
+tarrd-r-r2. o,o,
r 6 r' + o rt
f nterpretation
(i) Wtrenr -0, a= E
,inesorregre.,.?;:;':Til::T';;::H*:ffi1,"::::ffi
(ii) luren /=t1' t*.,,=o+a=O
or
T":"
z -Thetworegression linescoincideandthereis
perfectcorrelationbetween
thevariablesx andy.
Regression Gurves for
the Means
EIY / X = xl is called theregression
functionof y onX.
nlX t f = yJ is calledtheregressionfunction
.f*
Regression curveof lon Xis l, = E[y / X = "*,
x]
RegressioncurveofXon Iis .r = E[X tf = yl

Solved ExarnpGE
Example I

obtain the equationsof the rines

of regressionfor the fotowing
data
 Probabitity and R

Solution

Zu =o;Zr,=o;Zu,=36;Zv,
=++ I
Regr essio n coeffrcienE : -d
u..='ZY.-(zu)(1v\
_,"i?_'0",_:::,
8x3Eo
= o' 667

u-='ZY.-(Zu)(1r')
-(z')'
8x:?''
=g*44=0.545

Yr
X=L" =68andV-- rr -69
nn
Equationof line of regressionof y on X
Y-F = t * ( x - x )
.'.Y_69=0.re2(x _a\
=I=0.667X+23.64
Equationof line of regressionof X on y:

x _ x =r 9 z_ ( r _ t)
oy
+x_ 6 8 = 0.54s(_69)
y
X =0.545y
+30.395
 Two Dimensional. Random Variabtes
2. 69
Example 2
In a partially destroyed laboratory record of
an analysis of correlation data, the following
resultsonly are legible.
Varianceof x : g.,Regressionequations:gx_10y+66=
0 ; 4;x_lgy=2Id.
Find (D the meanvaluesof x andy.
(ii) the corelation coefficient betweenx and y
and
(iii) the.standarddeviationofy.

Solution
(i) Sinceboth regressionlines passthroughthe poiht
(i,r),we have
8t - 10t = -66 and 4Or -lB, = 214. Solving
we get x = 13 and ! =lT .
(ii) Let 8'r-10IF66 = 0 and,40x-l8y= 2l4bethe
lines of regressionof ronxan d x on y
respectively.
8 , 18
^" 214
-..66 and.r=
Then ./= j:t+:- y+ --'
t0 l0 40' 40
.r,=*=1-o
u,=#=*
Hence,t =4 .9 - 9
520 25
.'.t=t]=10.6
Since both *r" ,Jgr"..ion coefficients are positive,
we take r = 0.6
(iiD We have
614
f.4=_
615

4 3o..
. ._ =;.*since -x = 3
o-
5 5 3 -
)or=4

.'. Standarddeviation of v = 4
./
,/
Example3/
The equationsof two regressionlines obtainedin a qorrelation
analysisare as follows:
3x + l2y = 19;3.y+ 9x =46. Obtain
(i) the value of correlationcoefficient (ii) mean
value ofx and v.
 2 I Probability and

Solution
(i) Regression
line ofy on x: 3x +I2! =19

=+'yt 2_1_1r*12
2
Regression
line of.r ony: 3y +9x = 46

= ) . r -- ! r * {
3' 9
...bo=-* u"ab* =-!
12ry3
Hencert =bo.bo

=1
t2
r J t r=+-:-
=r=t;fti.e.,
Sinceboth regressioncoefficientsare negative
1
f=-
-2J3
(ii) Both regressionlinespassthroughttrepoint (I,y-)
3 7 + l2 y= 1 9 a n d9 7 + 3 ! = 4 6
Solving,we get f =5 andI= I
'3
Example4
I
/
Given that X =4Y +5 andy =lcX +4 arethe lines of
respectively,showthat O<4k< l. If k =L find.the meansof
16'
of correlationbetweenthem.

Solution
X =4y +5?b, =4
y = k X + 4 -b o = ls

: . r2= 4 k .B u t0 < r2< l+ 0 < 4 / r< l

If f r = a . , , = ! + r = 1-2
1
t6' 4
I
;., = smceregression
coefficients
arepositive
i,

L
 Tw o Dimensional Random V a ria b le s
2. V0
For fr = j, ,rr. two linesof regression
are
t6'
x-4y =5 andx-l6y=-6 4
Solving,V =5.75and-7-
=28
Example 5
If the linesof regression of y on x and)con y arerespectivery
arx+br!*cr=e
arx+ bry * c, = 0, provethat arb,S arb,. ^4
So lu ti on
a rx+ b ry+q= 0 ) b o= - f,| o.1

arx+brytcr=0+bo =_!26 ony)

.'.r'=boxb,
/ \/
( o,\( b, )
=t_+ !t _- |

t 4 , 1 [o
. ,)
arb,

4o,
.'.rt <l+arb,<bra,
Example6
For two random variables x andy with the same niean, the
two regression equations are
! = ax +band x = dy +B. Show that !=+. Find also the cornmonmeanand correla-
B l-a
tion coefficient.
Solution
Let the common meanbe m.
Then the two regressionequationsare
y _m= a(x_m)and
x _m= a(y _*\
But y = ax + b andx = d! + B aretheregressionlines.
...b=m(t-a)and f =m(t_a)
b' l-a
=- =.
B 1 '-a
 Probabilit and Random

ExamPle l
a-'r
lifetime of a certain kind of electric bulb may be considered
The that the a
250 hours' Find the probability
mean 1200hours and S'D
centrallimit theorem'
exceeds1250hours using

Solution
the life time of the bulbs'
Let X i (i =1,2,"'60) denote
ando'' =Y(X')=2502hrs
Then p' =E(X,)=l200hrs
=250"
= P=1200 ando'
X follosr
life time of 60 bulbs'By C'L'T'
I-et Ndenotethe average

tet Z =!a+ betheS.N.V

c t lln
=rlz >t'ssf
Plx'1250]
=0.5-P[0<z 4'551
= 0'5- 0'4394 I
= 0'0606
.'.Requred ProbabilitY

ExamPle2 4a
variateswith parameter
lf X t, X z,X t' " "I n arePoisson
so= x' + x' + " "' + x " n
rna r[izo< S"3160]where
Solution
S"=X, +Xr+""'+X"n=75
gtr"es
with parametet)'=2' This
EachX,is Poissondistribtrted
var(X,) = )"=2
p = 75 x 2 =
distribution wrth rnean n
By CtT, S, follows Normal
variance= no2 =t50
.'.,s'- N(t50'150)
Tofind Pllzo<s"<160I

r€tz =|ffibe theS.N.V

 Iwo Dimensional Random Variabtes
2 ,g g
plrzo<s,
<t6oJ=r[_2.4ssz<o.s2l
=p[_z.qs<z <o]+p[o<z <o.tz]
=rfo <z <2.+s]+
r[o <z <0.82]
=0.4929+0.2939
= o'7868
Exampre B
A dishibution with unknown
mean p has 1.5. Using CLI, find
should be taken from how large a sample
the disfribution ,r, or..o**ce
probablitv
s*pr",n"**r ;;;;;ffiffi"fiil::":':fl.-ur beatreast
o]gs,h"i
,h"
Solution
tet X denote the samnle haah
^r^ --,,
Given n takenfrom
thedistriburion.
E(x,)=, #t;?fr ::i]J,"',
By CLT,Xfo[ows normal ;o
disfibution with p and
variance o,
t.".,8ar( o,{) n
Wehaveto findr, rrr".) o#
r*pre, suchthat
Pla -0.s. On >o.es (or)pnx pl*0.5]>0.e5
_o,,*o.s] -
L€t =!(
Z
o /,/n

=;LX
t ln"-
- al
w h e nX - p =0 .5 ,2=-S (o .O a .^^- r -

:. rllP _a!. o.r]= eflz

k0.4082,,1;f
So we have to the least
value of z such that
nl z kot.+ol2J-nl=
o.e5
From the area table of
the norma-lcurve,
r[l z 1<t.sa]=
o.e5
.'.0-4082J;=1.96+ Jr - r-96
.'.Therequired 0-4082
sample
sizen = 24
J
P
ProbabilitY and Random

ExamPle4
of size100istakenfrom"t"tt:t:1,:-t:t^""T"3,1t"t3"Ti
A sample will
of thesample
thatthemean
canweassert
..r1:T;"ffffi;r*
than 4'
F=60 bY more
Solution
from the given
the samplevaluesdrawn
I'Et Xi,i= 1,2,3'"'100 be
:
.Then f(X,\=60 and f(X)=400'Herez='100"
with mean p =
mean 7 follows noTal distribution
;;;;:;sample
o' 4oo A
va ri a n ce ;=ffi =*
p = 60 by more
that the samplemean X wi-ll not differ from
Probabilif
- pl=A)
"[x .'. - r,l=4= ool<+]
"[x "ilt
= r [so <X < a + )
-

z =-I1-,
r.et =N
o l{n
tu'
.'.P[s6< x 364]=rl-z<z <21
=2P[o <z<21
=2x0.4'7'72
= 0'9544
.'.RequiredProbabilitY

ExamPle5
thatinacertain
Suppose circun' thattlre
il;';; whatistheprobabilitv
'o:ttlt'o,:,T:,:11"::l"Tl;nff
"r":"TJ:":,H';* '"'n'"tf"rv'
assumingindependence'
will exceed98 ohms'
the circuit

Solution
in the i'i resistor'
l-et X,rcptesent the resistance
distributedR'Vs
TlrenX,',X 2,"'X ro NQidentically
resistance
connected in series'the fotal
Since the resistors are
RI S,- Xr+ Xr+ ""+ X'o
=0'20for all f'
HereElx,l=5 andl"d(x')
,[
f;
{
Two Dimensional Random Variables
2. 100
By C.L.TS, - N(nyt,noz)
To find P(.1, > 98) :
E (S ")-np=20x5=1 0 0
var(5,) = oot = 20x 0.20= 4
lo-JD
I-r:t 7=

P[s, tt]=r[i >-t)

'
=0.5+p(-1.z . o )
= 05+ p(0 .2 . 1 )
= 0.5+ 0.3413
probability: 0.8413
.'.Required
Example 6
20 dicearethrown.Findapproximately
theprobabilitythatthesumobtainedis between65
and75 usingC.L.T.

So lu ti on
Let X, be the R.V representing
thenumbershownon the ih dje.
6, r1\ l-
rh e n E(x ,)=I;l i l =l tr +2+3+ 4 + 5 + 6 1 = 1

jr,,+22
ul*t)=li,l= +32
+42
*ri*1
=91
6
=
Then
"*(4) nlxll-ln(x,)f'
_9r _49 _35
64t2
I-et S, be the total score.

So= Xr + X, + ....+Xzo

t(s,) =2 0 *Z =7 9=(np)

=2 0 *f =\={ "C)
" *(s" )
2 .1 0 1 Probabi and Random P

By CLI, S, - N(n1t,no'\

i",+-{zof)
Let Z =s'-70 oetheS.N.v
lrTs
!3
When S, =65,2 = -0.65
S^=75,2= 0.65
< 75]= r1-0.t5< z < 0.65f
:. Pl65<,s"
=zPlo
<z <o.6sl
= 2x0.2422= 0.4844

.'. Requiredprobability = 0.4844

Example7
What is the probability that the averageof 150 random points from 6e
within 0.02 of the mid point of the interval?

Solution
Let Xi(i =1,2,...150) be the randompoints in the interval.
Then the random variable X, follows uniform distribution in the interval

1 11. for0<x<1
The P.d.fof x, is f (4=
t_O=l O.o***

o* b - o)'
Mean = *rdvariance -@ for turiform distributionrn
2t2

:.Elx,l=I
2
*a vlx,l=
't .L^
12
'
Let X denotethe averageof the randompointsX,, X2,X3,....X$,
Xr+ X, +......+Xrro
. v =-----36-
"^
r(X)=I^o*'(t)=#*
rnen
= 1800
 andom Variables 2. 104

ronndp[lx-0.51
PL -'aJ
I
./12x 150
.'.r[0.+s< * <o.sz]=p[-r,..
=zPlo'<z<0.8+.
=2Plo<z
so.ssl
=2(w023)=0.6046

...The requiredprobability : 0.6046

Example8
If I/i ,i =1,2,..,20areindependentnoisevoltagesreceivedin an adderand Zis the sumof the
voltagesreceived,find the probability that the total incoming voltage Z exceeds105, using
central limit theorem.Assumethat eachof the random variables ( is uniforrnly distributed over
I (0,10)
j

; Solution
The randomvariable ( follows uniform distributionover (0,10). Then E(v,)= 5 and
u/ = 1nn
( 10- n\ ' tuu
vU /,1=. = v fi
\'/r, \ =3T
12 12
Thetotalincoming
voltageis [/ =V, +V, + .....+yn
By CtjI, ZfollowsNormaldistributionwith mean=nF =20x5 = 100 and

variance= not =t?o

J

Tofindrltr >rcSl:

-1oo
I-etZ =l/ betheS.N.V
1500
r/;
2 .1 0 3 P ro b a b ilit a n d Ra n d o m P r

P fv1 1 0 5]=
P[z>a39]
=0.5- P[o< z <0:,9]
= 0.5- 0.15i7= 0.3483

.'.P [TotalincomingvoltageZexceeds105]:0.3483

Example 9
A coin is tossed10 times. What is the probability of getting 3 or 4 or 5 headsl
limittheorem.

Solution
ktXdenote the number of headsin 10 tosses.
Here z = I0, p= probability of gettingheadin a trial =
]
:.q=L-P
^2
=+
MeanofX =nD=1 0 x 1 = 5
-2

X = npq=19*1*1 = 2.5
Variance
22
:.p=san d o = f f = 1 . 5 8
To find Pf3< X =5] . SinceXis a discret€RV we find P12.5< X <551
t -" . -5 5 . 5 -5 1
:.Plz.s<x < s.5l= Pl <z <# |
' L 1.58 1.58I
-
=P[-1.58
<z <0327 wherez =
=P[-r.58
.z.ol+r[o< z <w2]
=P[o<z <r.587+r[o
<z <B2l
= 0.4429+ 0.1255

Hencerequired probability = 0.5684

Example I O
The buming time of a certain tlpe of lamp is an experimentalrandom variable
hours. What is the probability that 144 of theselamps will provide a total of rnmr
hours of burning time?
 Two Dimensional Random Variables 2. t 04
Solution
Let X, be the buming time of the i'h larcrp.

." X1,X2,X3,....Xtuareidenticallydistributedindependent
randomvariables.
For exponential distribution,

E(\ / )Xl=
"At
1 *d ,,-i*"" =I
.'.Mean of X, = 30 and var(X ) = eOO

Let S, = Xr I X, +....+ Xr* bethe total bumrngtime.

By CLT,S, - N(np,no'\

s,:nlt = 4 -(t++)ro
I-et2 =
Ji,o 30.,1144
= sn - 4i20
360
To find P(S, > 4500)
plz>o.sl
PIEt45oo]= ? _4500-4320
360
=0.5-P<[ 0z < o .sf = 0.5
= 0.5- 0.1915= 0.3085