1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Density Function for Sums of Random Variables

  1. May 12, 2010 #1
    1. The problem statement, all variables and given/known data

    Given the joint density, f(x,y), derive the probability density function for Z = X + Y and V = Y - X.

    2. Relevant equations

    f(x,y) = 2 for 0 < x < y < 1
    f(x,y) = 0 otherwise.

    3. The attempt at a solution

    For Z = X + Y, I can derive the fact that,

    [tex]f_Z(z) = \int_{-\infty}^{\infty} f(x,z-x)dx [/tex]

    The support should be 0 < x < z - x < 1? But I am kind of lost from here.

    0 < x < 1 and 0 < y < 1, so 0 < z < 2? The book I am using tell me there are two cases but I have no idea how they deduced the two cases. From my very limited understanding, f(x,y) = 2 for all x,y in its support. So why are there two cases?

    Thanks!
     
  2. jcsd
  3. May 12, 2010 #2

    lanedance

    User Avatar
    Homework Helper

    try thinking about it geometrically, [itex] f_{X,Y}(x,y) [/itex] is uniform on the triangle within the unit square and below y=x

    now think about lines of constant z = x+y, and compare with your intgeral
     
  4. May 12, 2010 #3
    i came up with an answer on my own, is the process correct? i looked at the two arguments of f(x, z-x) and try to deduce which bounds are significant.

    1) argument "x"

    0 < x < z - x.

    so x/2 < x < z/2. the z/2 is significant across all z in (0,2) since for z < 2, z/2 < 1. x < z/2 < 1, so the value of z/2 takes precedent over the numerical upper bound of 1.

    the lower bound of 0 is significant

    2) argument "z - x"

    x < z - x < 1

    so z - 1 < x < z - x.

    the upper bound of z - x is not significant as x < z/2 is equivalent to x < z - x.

    0 < z < 2 => -1 < z - 1 < 1, so the lower bound z - 1 < x is significant only for some z.

    for 0 < z < 1, -1 < z -1 < 0 is totally not relevant compared to 0 as z-1 < 0 < x anyway. but for 1 < z < 2 we have 0 < z - 1 < 1 which is relevant as 0 < z -1 < x, so z - 1 is the greatest lower bound.

    3) Intervals

    Hence the intervals are 0 < z < 1 where 0 < x < z/2, and for 1 < z < 2 where z - 1 < x < z/2.

    4 ) Algorithm

    Basically I followed this self made algo which i hope is correct!

    1. figure out the bounds for the two arguments of f(x, z - x)
    2. compare it to the bounds for x, e.g. a < x < b
    3. find the least upper bound and greatest lower bound. partition z if need be.


    Thanks!
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook