Engineering Mathematics Material

SUBJECT NAME SUBJECT CODE MATERIAL NAME MATERIAL CODE : Probability & Queueing Theory : MA 2262 : Formula Material : JM08AM1007

2010

Name of the Student:

UNIT-I (RANDOM VARIABLES)

Branch:

Sl.No. Discrete random variable

1 2
3
i

ht tp

p( xi ) = 1

F ( x) = P [ X x]

:/ v /c ia se t

1) Discrete random variable: Arandomvariablewhosesetofpossiblevaluesiseitherfiniteorcountably infiniteiscalleddiscreterandomvariable. Eg:(i)LetXrepresentthesumofthenumbersonthe2dice,whentwo dicearethrown.InthiscasetherandomvariableXtakesthevalues2,3,4,5,6, 7,8,9,10,11and12.SoXisadiscreterandomvariable. (ii)Numberoftransmittedbitsreceivedinerror. 2) Continuous random variable: ArandomvariableXissaidtobecontinuousifittakesallpossiblevalues betweencertainlimits. Eg:Thelengthoftimeduringwhichavacuumtubeinstalledinacircuit functionsisacontinuousrandomvariable,numberofscratchesonasurface, proportionofdefectivepartsamong1000tested,numberoftransmittedin error. 3)

Continuous random variable

f ( x )dx = 1

ub

e.
2

F ( x) = P [ X x] = Mean = E [ X ] =

Mean = E [ X ] = xi p( xi )
i

xf ( x )dx

4 5 6 7

X2 = xi2 p( xi ) E
i
2

E X =
2

f ( x )dx
2

2 Var ( X ) = E ( X 2 ) E ( X ) Var ( X ) = E ( X ) E ( X )

r r Moment= E X = xi pi
i

r Moment= E X =

tk

M.G.F

M.G.F Page 1

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

/
r

f ( x )dx

f ( x )dx

Engineering Mathematics Material

tX tx MX (t ) = E e = e p( x ) x tX MX (t) = E e =

2010

tx

f ( x )dx

5) Var ( aX + b ) = a 2 Var ( X )

4) E ( aX + b ) = aE ( X ) + b

6) Var ( aX bY ) = a 2 Var ( X ) + b 2Var (Y )

7) Standard Deviation = Var ( X ) 8) f ( x ) = F ( x ) 9) p( X > a ) = 1 p( X a ) 10) p ( A / B ) =

p( A B)
p( B)

, p( B) 0

11) IfAandBareindependent,then p ( A B ) = p ( A ) p ( B ) . 12) 1stMomentaboutorigin= E [ X ] = M X ( t ) (Mean) t =0 2 2ndMomentaboutorigin= E X = M X ( t ) t = 0 tr th r Theco-efficientof = E X (r Momentabouttheorigin) r! 13) Limitation of M.G.F: i) ArandomvariableXmayhavenomomentsalthoughitsm.g.fexists. ii) ArandomvariableXcanhaveitsm.g.fandsomeorallmoments,yetthe m.g.fdoesnotgeneratethemoments. iii) ArandomvariableXcanhaveallorsomemoments,butm.g.fdoesnot existexceptperhapsatonepoint. 14) Properties of M.G.F: i) IfY=aX+b,then MY ( t ) = e bt M X ( at ) . ii) iii)
M cX ( t ) = M X ( ct ) ,wherecisconstant.

IfXandYaretwoindependentrandomvariablesthen M X +Y ( t ) = M X ( t ) M Y ( t ) .
Mean
t

15) P.D.F,M.G.F,MeanandVarianceofallthedistributions: Sl. Distributio M.G.F P.D.F ( P ( X = x ) )

No. 1 n Binomial

ht tp

:/ v /c ia se t

ub

e.
)

tk
n

/
np

Variance

nc x p x q n x
e x! q x 1 p (or) q x p

( q + pe )

e
e t 1

npq

Poisson

1 p

q p2
Page 2

Geometric

pe t 1 qe t

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

Engineering Mathematics Material

4 Negative Binomial

2010
p t 1 qe e bt e at ( b a )t

( x + k 1)C k 1 p k p x

kq p

kq p2

Uniform

Exponential

Gamma

1 , a< x<b f ( x) = b a otherwise 0, e x , x > 0, > 0 f ( x) = 0, otherwise e x x 1 , 0 < x < , > 0 f ( x) = ( )

a + b (b a )2 2 12 1

Weibull

1 (1 t ) f ( x ) = x 1e x , x > 0, , > 0

ij

f ( x , y )dxdy = 1 (Continuousrandomvariable)

:/ v /c ia se t

ub
P ( x, y ) P( y) P ( x, y ) P( x)

1)

= 1 (Discreterandomvariable)

ConditionalprobabilityfunctionYgivenX, P {Y = yi / X = xi } =
P { X < a / Y < b} =

ht tp

2) ConditionalprobabilityfunctionXgivenY, P { X = xi / Y = yi } =

e.
. .

UNIT-II (RANDOM VARIABLES)

P ( X < a,Y < b ) P (Y < b )

3) ConditionaldensityfunctionofXgivenY,

f ( x / y) = f ( y / x) =

f ( x, y) . f ( y) f ( x, y) . f ( x)

ConditionaldensityfunctionofYgivenX,

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

tk
Page 3

16) Memorylesspropertyofexponentialdistribution P( X > S + t / X > S) = P( X > t).

Engineering Mathematics Material

4) IfXandYareindependentrandomvariablesthen

2010

f ( x , y ) = f ( x ). f ( y )

(forcontinuousrandomvariable)

P ( X = x , Y = y ) = P ( X = x ) . P (Y = y ) (fordiscreterandomvariable)

5) Jointprobabilitydensityfunction P ( a X b, c Y d ) = f ( x , y )dxdy .
c a

d b

P ( X < a , Y < b ) = f ( x , y )dxdy

0 0

b a

6) MarginaldensityfunctionofX, f ( x ) = f X ( x ) = MarginaldensityfunctionofY, f ( y ) = fY ( y ) = 7) P ( X + Y 1) = 1 P ( X + Y < 1) 8) Correlation co efficient (Discrete): ( x , y ) =

Cov ( X , Y ) =

f ( x , y )dy

f ( x , y )dx

Cov ( X , Y )

9) Correlation co efficient (Continuous): ( x , y ) =

10) IfXandYareuncorrelatedrandomvariables,then Cov ( X , Y ) = 0 . 11) E ( X ) =

ht tp

Cov ( X , Y ) = E ( X , Y ) E ( X ) E (Y ) , X = Var ( X ) , Y = Var (Y )

xf ( x )dx , E (Y ) =

:/ v /c ia se t

1 1 XY XY , X = X 2 X 2 , Y = n n Cov ( X , Y )

ub
X Y

yf ( y )dy , E ( X , Y ) =

12) Regression for Discrete random variable: RegressionlineXonYis x x = bxy ( y y ) , RegressionlineYonXis y y = b yx ( x x ) ,

bxy =

b yx

( x x)( y y) ( y y) ( x x)( y y) = ( x x)
2 2

Correlationthroughtheregression, = bXY .bYX Note: ( x , y ) = r ( x , y )

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

e.
1 Y 2 Y 2 n

X Y

xyf ( x, y )dxdy .

tk
Page 4

Engineering Mathematics Material

13) Regression for Continuous random variable: RegressionlineXonYis x E ( x ) = bxy ( y E ( y ) ) ,
RegressionlineYonXis y E ( y ) = b yx ( x E ( x ) ) ,
bxy = r b yx = r

2010

x y y x

RegressioncurveXonYis

x = E ( x / y) = y = E ( y / x) =

x f ( x / y ) dx
y f ( y / x ) dy

RegressioncurveYonXis

14) Transformation Random Variables: fY ( y ) = f X ( x )

dx dy

u y v y

(Onedimensionalrandomvariable)

15) Central limit theorem (Liapounoffs form)

IfX1,X2,XnbeasequenceofindependentR.VswithE[Xi]=iandVar(Xi)=i2,i =1,2,nandifSn=X1+X2++Xnthenundercertaingeneralconditions,Sn

:/ v /c ia se t
n

ub
n i =1 i =1

u x fUV ( u, v ) = f XY ( x , y ) v x

n.

16) Central limit theorem (Lindberg Levys form) IfX1,X2,XnbeasequenceofindependentidenticallydistributedR.VswithE[Xi] =iandVar(Xi)=i2,i=1,2,nandifSn=X1+X2++Xnthenundercertain generalconditions,Snfollowsanormaldistributionwithmean n andvariance

n 2 as n .
Note: z =

Sn n X (fornvariables), z = (forsinglevariables) n n
Page 5

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

ht tp

followsanormaldistributionwithmean = i andvariance 2 = i2 as

e.

(Twodimensionalrandomvariable)

tk

Engineering Mathematics Material

2010

UNIT-III (MARKOV PROCESSES AND MARKOV CHAINS)

1) Random Process: Arandomprocessisacollectionofrandomvariables{X(s,t)}thatare functionsofarealvariable,namelytimetwheresSandtT. Classification of Random Processes: Wecanclassifytherandomprocessaccordingtothecharacteristicsoftimet andtherandomvariableX.WeshallconsideronlyfourcasesbasedontandX havingvaluesintheranges-<t<and-<x<. Continuousrandomprocess Continuousrandomsequence Discreterandomprocess

2)

Continuous random process: IfXandtarecontinuous,thenwecallX(t),aContinuousRandomProcess. Example: IfX(t)representsthemaximumtemperatureataplaceinthe interval(0,t),{X(t)}isaContinuousRandomProcess. Continuous Random Sequence: ArandomprocessforwhichXiscontinuousbuttimetakesonlydiscretevaluesis calledaContinuousRandomSequence. Example: IfXnrepresentsthetemperatureattheendofthenthhourofaday,then {Xn,1n24}isaContinuousRandomSequence. Discrete Random Process: IfXassumesonlydiscretevaluesandtiscontinuous,thenwecallsuchrandom process{X(t)}asDiscreteRandomProcess. Example: IfX(t)representsthenumberoftelephonecallsreceivedintheinterval (0,t)the{X(t)}isadiscreterandomprocesssinceS={0,1,2,3,...} Discrete Random Sequence: Arandomprocessinwhichboththerandomvariableandtimearediscreteis calledDiscreteRandomSequence. Example: IfXnrepresentstheoutcomeofthenthtossofafairdie,the{Xn:n1}isa discreterandomsequence.SinceT={1,2,3,...}andS={1,2,3,4,5,6} Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)
http://csetube.weebly.com/

ht tp

:/ v /c ia se t

ub

e.

tk

Discreterandomsequence

/
Page 6

Engineering Mathematics Material

3)

2010

Condition for Stationary Process: E [ X ( t )] = Constant , Var [ X ( t )] = constant .

Iftheprocessisnotstationarythenitiscalledevolutionary. Wide Sense Stationary (or) Weak Sense Stationary (or) Covariance Stationary: ArandomprocessissaidtobeWSSorCovarianceStationaryifitsatisfiesthe followingconditions. i) Themeanoftheprocessisconstant(i.e) E ( X ( t ) ) = constant . ii) Autocorrelationfunctiondependsonlyon (i.e) RXX ( ) = E [ X ( t ). X ( t + )]

4)

5)

Property of autocorrelation: (i) (ii)

RXX ( ) E ( X (t )) = lim
2

E ( X 2 ( t ) ) = RXX ( 0 )

6)

7)

thentheprocess { X n } , n = 0,1, 2, ... iscalledthemarkovchain.Where

9)

equaltoPn.(i.e) Pij( n ) = . Pij 10) Markov Chain property:If = ( 1 , 2 , 3 ) ,then P = and

1 + 2 + 3 = 1 . 11) Poisson process: If X ( t ) representsthenumberofoccurrencesofacertaineventin (0, t ) ,then

thediscreterandomprocess { X ( t )} iscalledthePoissonprocess,providedthe followingpostulatesaresatisfied.

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

ht tp

8)

a0 , a1 , a2 , ...an , ... arecalledthestatesofthemarkovchain. Transition Probability Matrix (tpm): WhentheMarkovChainishomogenous,theonesteptransitionprobabilityis denotedbyPij.ThematrixP={Pij}iscalledtransitionprobabilitymatrix. Chapman Kolmogorov theorem: IfPisthetpmofahomogeneousMarkovchain,thenthensteptpmP(n)is
n

:/ v /c ia se t

ub

= P X ( t n +1 ) xn+1 / X ( t n ) = xn Where t 0 t1 t 2 ... t n t n +1 Markov Chain: Ifforall n , P X n = an / X n 1 = an 1 , X n 2 = an 2 , ... X 0 = a0 = P X n = an / X n 1 = an 1

e.

tk

Markov process: Arandomprocessinwhichthefuturevaluedependsonlyonthepresentvalue andnotonthepastvalues,iscalledamarkovprocess.Itissymbolically representedby P X ( t n +1 ) xn + 1 / X ( t n ) = xn , X ( t n 1 ) = xn 1 ... X ( t 0 ) = x0

Page 7

Engineering Mathematics Material

(i)
(ii) (iii) (iv)
P [1 occurrence in ( t , t + t )] = t + O ( t )

2010

P [ 0 occurrence in ( t , t + t )] = 1 t + O ( t )

P [ 2 or more occurrences in ( t , t + t )] = O ( t )

X ( t ) isindependentofthenumberofoccurrencesoftheeventinany interval.
et ( t )
n

12) Probability law of Poisson process: P { X ( t ) = n} =

, n = 0,1, 2, ... n! 2 2 2 Mean E [ X ( t )] = t , E X ( t ) = t + t , Var [ X ( t )] = t .

UNIT-IV (QUEUEING THEORY)

n Numberofcustomersinthesystem.

Meanarrivalrate.

Meanservicerate.

Wq Averagewaitingtimepercustomerinthequeue.
W s Averagewaitingtimepercustomerinthesystem.

1) 2) 3)

ServerUtilization =
Pn = n ( 1 )

Ls =
Lq =

4)

2 1
1 (1 )

5)

Ws =

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

ht tp

Model I

(M / M / 1): ( / FIFO)

(P0nocustomersinthesystem)

:/ v /c ia se t

ub

Ls Averagenumberofcustomersinthesystem.

e.
Page 8

Lq Averagenumberofcustomersinthequeue.

tk

Pn SteadyStateprobabilityofexactlyncustomersinthesystem.

Engineering Mathematics Material

6)
7)
Wq =

2010

(1 )

Probabilitythatthewaitingtimeofacustomerinthesystemexceedstis

P ( w s > t ) = e ( ) t .
8) Probabilitythatthequuesizeexceedstis P ( N > n ) = n +1 where
n = t + 1 .

Model II
1) =

(M / M / C): ( / FIFO)
s
1

s s 1 ( s ) n s ) ( 2) P0 = + n! s !(1 ) n= 0

5) Wq = 6) Ws = 7)

Ls

ht tp

( s ) P Theprobabilitythatanarrivalhastowait: P ( N s ) = s !(1 ) 0
s

8) Theprobabilitythatanarrivalenterstheservicewithoutwaiting=1P(an arrivalhattowait)= 1 P ( N s ) 9) P ( w > t ) = e

t s 1 e t ( s 1 s ) ( s ) P 1 + 0 s !(1 )( s 1 s )

Model III
1) =

(M / M / 1): (K / FIFO)

Page 9

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

:/ v /c ia se t

Lq

ub

4) Ls = Lq + s

e.

1 ( s ) P 3) Lq = s.s ! ( 1 ) 2 0
s +1

tk

Engineering Mathematics Material

2) P0 =

2010

1 1 k +1

(Nocustomer) (effectivearrivalrate)

3) = ( 1 P0 ) 4) Ls =

( k + 1) k +1
1 k +1

5) Lq = Ls 6) Ws = 7) Wq =
Ls Lq

8) P [ a customer turned away ] = Pk = k P0

n s

n= s

s 1 4) Effectivearrivalrate: = s ( s n ) Pn n= 0

5) Lq =

( s ) =

( 1 k s ) ( k s ) k s +1 P0 s ! (1 )2 1
s

6) Ls = Lq + 7) Wq =
Lq

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

ht tp

P0, n s 3) Pn = n! n s ( ) s ! s n s P0 , s n k

(s )

:/ v /c ia se t

s 1 ( s ) n ( s ) s 2) P0 = + n! s! n= 0

ub

e.
Page 10

1) =

tk

Model IV

(M / M / C): (K / FIFO)

Engineering Mathematics Material

8) Ws =
Ls

2010

UNIT-V (NON MARKOVIAN & QUEUEING NETWORK)

1) Pollaczek Khintchine formula:
LS = E ( t ) +
2 2 Var ( t ) + ( E ( t ) )

2 [1 E ( t ) ]

(or)

LS = +

2 2 + 2 2 (1 ) 2 2 + 2 2 (1 )

2) Littles formulas:

LS = +

Wq =

3) Series queue (or) Tandem queue: Thebalanceequation

P00 = 2 P01

1 P10 = P00 + 2 P11

P01 + 2 P01 = 1 P10 + 2 Pb1 1 P11 + 2 P11 = P01 2 Pb1 = 1 P11

Condition P00 + P10 + P01 + P11 + Pb1 = 1 4) Open Jackson networks: i) Jacksonsflowbalanceequation j = rj + i Pij
i =1 k

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

ht tp

:/ v /c ia se t

Lq

ub

WS =

LS

e.
Page 11

tk

Lq = LS

Engineering Mathematics Material

Whereknumberofnodes,rjcustomersfromoutside ii) Jointsteadystateprobabilities
P ( n1 , n2 , ...nk ) = 1n1 ( 1 1 ) 2 n2 ( 1 2 ) ... k nk ( 1 k )

2010

iii) Averagenumberofcustomersinthesystem
LS =

1 2 k + ... + + 1 1 1 2 1 k

iv) Averagewaitingtimeofacustomersinthesystem
WS =

LS

where = r1 + r2 + ... + rk

5) Closed Jackson networks: Intheclosednetwork,therearenocustomersfromoutside,therefore rj = 0 then

nk P ( n1 , n2 , ... nk ) = C N 1n1 2n2 ... k

Where C N 1 =

n1 + n2 + ... + nk = N

ht tp

ii) Ifeachnodessingleserver

n 1n 2n ... k
1 2

iii) Ifeachnodeshasmultipleservers
P ( n1 , n2 , ...nk ) = C N
n 1n 2
1 2

a1 a2
1

...

Where C N 1 =

n1 + n2 + ... + nk = N

n 1n 2

a1 a2

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

:/ v /c ia se t
k

P11 P21 ( 1 2 ... k ) = ( 1 2 ... k ) P k1

P1k P22 ... P2 k Pk 2 ... Pkk P12 ...

kn
2

ak

...

kn

ak

ub

(or)

e.
Page 12

i =1

tk

i) TheJacksonsflowbalanceequation j = i Pij

rj = 0

Engineering Mathematics Material

, ni < si ni ! ai = ni si , ni si si ! si

2010

---- All the Best ----

Prepared by C.Ganesan, M.Sc., M.Phil., (Ph: 9841168917)

http://csetube.weebly.com/

ht tp

:/ v /c ia se t

ub

e.
Page 13

tk