MIT6_041F10_assn05_sol

Embed Size (px)

Citation preview

  • 8/12/2019 MIT6_041F10_assn05_sol

    1/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    ProblemSet5: Solutions1. (a)Because of the requirednormalization propertyofany joint PDF,

    2 2 2 23 13 21 = ax dy dx = ax(2 x) dx = a 22 12 + = a

    3 3 3x=1 y=x x=1soa= 3/2.

    (b)For1 y 2,y a 3

    fY (y) = axdx = (y2 1) = (y2 1),

    2 41andfY (y) = 0otherwise.

    (c) Firstnoticethatfor1 x 3/2,

    fX,Y (x, 3/2) (3/2)x 8xfX|Y (x| 3/2)= = 2 = .fY (3/2) 3 3 12 5

    4 2Therefore,

    3/2 1 8x

    E[1/X| Y= 3/2] = dx = 4/5.x 51

    2. (a) BydefinitionfX,Y (x, y) =fX(x)fY|X(y| x). fX(x) =axasshowninthegraph. Wehavethat

    401 = axdx = 800a.

    0SofX(x) =x/800.FromtheproblemstatementfY|X(y| x) =2

    1x fory [0, 2x].Therefore,

    1/1600, if0 x 4and0< y X.Thisoccurswithprobability

    40 2x 1 1

    P(Y> X) = fX,Y (x, y) dy dx = dydx = .1600 2y>x 0 x

    Wecould havealsoarrivedatthisanswer byrealizingthat foreach possiblevalueofX,thereisa 1/2probabilitythatY> X.

    (c)The joint density functionsatisfiesfX,Z(x, z) =fX(x) fZ|X(z|x). SinceZ isconditionally

    uniformlydistributedgivenX,fZ|X(z| x) = 21x forx z x. Therefore,fX,Z(x, z) =

    1/1600for0 x 40andx z x.Themarginaldensityfz(z) iscalculatedas

    40 1 40|z| , if|z|

  • 8/12/2019 MIT6_041F10_assn05_sol

    2/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    3. (a) In order forXandYto be independent,anyobservationofXshouldnot giveany informationonY.IfXisobservedtobeequalto0,thenYmustbe0.

    Inotherwords,fY|{X=0}(y| 0) fY (y).Therefore,XandY arenotindependent.=1.5

    1

    0, otherwise. 0.5

    0

    0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

    x

    2

    1.8

    1.6

    1.4

    x/2, if0 x 1,(b) fX(x) = 3x/2 + 3, if1< x 2,

    fX

    |Y(x |

    0.

    5)

    fY|

    X(y

    |

    0.

    5)

    fX

    (x)

    1.2

    1fY|X(y| 0.5)=2, if0 y 1/2,0, otherwise.

    1/2, if1/2 x 1,fX|Y (x| 0.5)= 3/2, if1< x 3/2,

    0, otherwise.

    0.8

    0.6

    0.4

    0.2

    00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

    y

    1.5

    0.5

    00 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

    x

    1

    (c)TheeventAleavesuswitharighttrianglewithaconstantheight.TheconditionalPDFisthen1/area= 8.The conditionalexpectation yields:

    E[R| A] = E[XY | A]0.5 0.5

    = 8xydxdy0 y

    = 1/16.

    (d)TheCDFofW isFW(w) =P(W w) =P(YX w) =P(Y X+w).P(Y X+w)canbecomputedbyintegratingtheareabelowthelineY=X+wforallpossiblevaluesofw.The linesY=X+wareshown belowforw= 0,w=1/2,w=1andw=3/2.TheprobabilitiesofinterestcanbecalculatedbytakingadvantageoftheuniformPDFoverthetwotriangles. Remember tomultiplytheareas bytheappropriate joint densityfX,Y (x, y)!Takenotethatthereare4regionsofinterest: w 0.

    Page 2of 7

  • 8/12/2019 MIT6_041F10_assn05_sol

    3/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    w= 0 y= 2 x1.21

    w (1, 0)y=x0.8

    y

    w=10.6

    w (2,1)0.4

    0.2

    0

    1 +w 2 +w0.2

    1+w2+w2

    ww2

    0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

    x

    TheCDFofW is0, ifw 0

    0, ifw 0.

    Asasanitycheck,FW() = 0 andFW(+) =1. Also,FW(w)iscontinuousatw=2andatw=1.

    4. (a) Ifthetransmittersendsthe0symbol,thereceivedsignalisanormalrandomvariablewithameanof2andavarianceof4.Inotherwords,fY|X(y| 2)=N(2, 4).

    Also,fY|X(y| 2)=N(2, 4)Theseconditionalpdfsareshowninthegraphbelow.

    Page 3of 7

  • 8/12/2019 MIT6_041F10_assn05_sol

    4/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    8 6 4 2 0 2 4 6 80

    0.05

    0.1

    0.15

    0.2

    0.25

    fY|X(y| 2)fY|X(y| 2)P(error| X=2) P(error| X=2)

    Theprobabilityoferrorcanbefoundusingthetotalprobabilitytheorem.

    P(error) = P(error | X=2)P(X=2) +P(error | X= 2)P(X= 2)1

    = (P(Y 0| X=2) + P(Y

  • 8/12/2019 MIT6_041F10_assn05_sol

    5/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    than 1/2.Another wayto prove that the two are not independent is to calculate the productoftheirexpectations,andshowthatthisisnotequaltoE[XY].

    (b)ApplyingthedefinitionofamarginalPDF,

    for0 x 1,

    fX(x) = fX,Y (x, y) dyyx+1

    = 1dyx

    = 1;

    for0 y 1,

    fY (y) = fX,Y (x, y) dxxy

    = 1dx0

    = y;

    andfor1 y 2,

    fY (y) = fX,Y (x, y) dxx

    1= 1dx

    y1= 2 y.

    1

    0 0.2 0.4 0.6 0.8 1

    fY

    (y)

    1

    0.80.8

    fX

    (x) 0.6

    0.6

    0.4

    0.2

    0

    0.4

    0.2

    00 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

    x y

    (c) Bylinearityofexpectation,theexpectedvalueofasumisthesumoftheexpectedvalues.Byinspection,E[X] = 1/2andE[Y] = 1.

    Thus,E[X+Y] =E[X] +E[Y] = 3/2.

    Page 5of 7

  • 8/12/2019 MIT6_041F10_assn05_sol

    6/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    (d)ThevarianceofX+Y is

    E[(X+Y)2]E[X+Y]2 =E[X2] + 2E[XY] +E[Y2] (E[X+Y])2. (1)In part (c),E[X+Y] wascomputed,soonly theother threeexpressionsneed to becalculated.First,theexpectedvalueofX2:

    1 x+1 1

    E[X2] = x2 1dydx = x2 dx = 1/3.0 x 0

    Also,theexpectedvalueofY2 is

    1x+1 1E[Y2] = y2 dydx = (3x2+ 3x+ 1)/3dx = 7/6.

    0 x 0Finally,theexpectedvalueofXY is

    1 x+1

    E[XY] = x y dy dx0 x

    1= (2x2+x)/2dxdy = 7/12.

    0Substitutingthese into (1),we getvar(X+Y) = 1/3 + 7/6 + 7/6 9/4= 5/12.

    *Alternative(shortcut)solution toparts (c)and(d)*GivenanyvalueofX (in([0,1]),weobservethatY X takesvaluesbetween0and1,

    andisuniformlydistributed. SincetheconditionaldistributionofYX isthesameforeveryvalueofX in [0,1],weseethatYX independentofX. Thus: (a)X isuniform,and(b)Y =X+U, whereU isalsouniformand independentofX. It follows thatE[X+Y] =E[2X+U] = 3/2.Furthermore,var(X+Y) = 4var(X) + var(U) = 5/12.

    6. (a)LetAbetheeventthatthefirstcointossresultedinheads. TocalculatetheprobabilityP(A),weusethecontinuousversionof thetotalprobabilitytheorem:

    1 1

    P(A) = P(A| P=p)fP(p) dp= p(1+ sin(2p)) dp,0 0

    whichaftersomecalculationyields

    1P(A) = .

    2

    (b)UsingBayesrule,

    P(A| P=p)fP(p)fP|A(p) = P(A) 2p(1 + sin(2p))

    , if0p 1,= 10, otherwise.

    Page 6of 7

  • 8/12/2019 MIT6_041F10_assn05_sol

    7/8

    Massachusetts InstituteofTechnologyDepartmentofElectricalEngineering&ComputerScience

    6.041/6.431: ProbabilisticSystemsAnalysis(Fall2010)

    (c) LetBbetheeventthatthesecondtossresultedinheads.Wehave

    1P(B| A) = P(B| P=p,A)fP|A(p) dp

    01

    = P(B| P=p)fP|A(p) dp02

    1

    = p2(1 + sin(2p)) dp. 1 0

    Aftersomecalculation,thisyields

    2 2 3 2 3P(B| A) = = 0.5110.

    1 6 3 3

    G1.Leta= (cos, sin ) andb= (bx, by).WewillshowthatnopointofRliesoutsideCifandonlyif

    |b| | sin |, and |a| | cos| (2)

    TheothertwoverticesofRare (cos, by) and (bx, sin ).If|bx| | cos | and|by| | sin |,theneachvertex (x, y) ofRsatisfiesx2+ y2 cos2+sin2= 1andno pointsofRcanlieoutsideofC.ConverselyifnopointsofRlieoutsideC,thenapplyingthistothetwoverticesotherthanaandb,wefind

    cos2 +b2 1, and a2+sin2 1.

    whichisequivalentto2.Theseconditions implythat (bx, by) liesinsideoronC,soforanygiven,theprobabilitythattherandompointb= (bx, by) satisfies (2)is

    2| cos| 2| sin | 2= | sin(2)|

    andtheoverallprobabilityis

    1

    2 2 4 /2 4| sin(2)|d= sin(2)d=

    2 0 2 0 2.

    Requiredfor6.431;optionalfor6.041 Page7of7

  • 8/12/2019 MIT6_041F10_assn05_sol

    8/8

    MIT OpenCourseWarehttp://ocw.mit.edu

    6.041 / 6.431Probabilistic Systems Analysis and Applied Probability

    Fall 2010

    For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

    http://ocw.mit.edu/http://ocw.mit.edu/termshttp://ocw.mit.edu/termshttp://ocw.mit.edu/