Function of two random variables

Click For Summary

Homework Help Overview

The discussion revolves around finding the distribution function of the random variable Z, defined as Z = X/(X+Y), where X and Y are independent, exponentially distributed random variables with parameter a.

Discussion Character

  • Exploratory, Assumption checking, Mathematical reasoning

Approaches and Questions Raised

  • Participants explore the relationship between Z and the random variables X and Y, discussing the need for double integrals and the definition of the integration region. Questions arise regarding the correct setup of inequalities and the implications of the bounds on z.

Discussion Status

The discussion is active, with participants providing insights into the integration region and the importance of considering the signs of factors involved in the inequalities. There is a focus on clarifying the conditions under which certain inequalities hold, but no consensus has been reached on a specific method or solution.

Contextual Notes

Participants note the need to be cautious about the bounds of z and the implications of the positivity of X and Y, which may affect the integration setup. The original poster expresses difficulty in understanding the problem and finding relevant materials.

Gauss M.D.
Messages
151
Reaction score
1

Homework Statement



We have two independent, exponentially distributed random variables X and Y (with parameter a).

Z = X/(X+Y)

What is Z:s distributon function?

Homework Equations


The Attempt at a Solution



I think I need some intuition to what I'm really doing with these, I'm having a really tough time and would be extremely grateful for a good explanation.

P(Z < z) = P(X/(X+Y) < z)... Solving this problem corresponds to doing a double integral of f(x,y) over some interval (?) defined by z. Am I getting that correctly? I don't understand how to find that interval. Help??!

Been at this for days but I'm still struggling. Having a hard time finding good materials that cover this aswell...
 
Physics news on Phys.org
It's going to be an integral over some region in the plane defined by z, not an interval

The region is exactly what you have written down already. It's defined by the equation x/(x+y) < z
 
Ok so we can write the area as: X < Yz/(1-z)

So ∫∫ f(x,y) dx dy, were we integrate from 0 to Yz/(1-z) with respect to x, and from 0 to ∞ with respect to y?
 
You have to be careful about whether 1-z is positive or negative when you divide both sides of the inequality though (so your region of integration is slightly different). Other than that it looks good
 
Office_Shredder said:
You have to be careful about whether 1-z is positive or negative
The need for care starts before that. You can't even get to X < z(X+Y) without worrying about the signs of factors you're multiplying by.
 
Can you elaborate on what "being careful" entails here? Obviously z will have bounds but I guess you're talking about something else?
 
Gauss M.D. said:
Can you elaborate on what "being careful" entails here? Obviously z will have bounds but I guess you're talking about something else?

For z > 0, [tex]\{X/(X+Y) \leq z \} = \{ X \leq z(X+Y) \} = \{ (1-z) X \leq z Y \}.[/tex] What do you get if 0 < z < 1? What do you get if z > 1?
 
Well... if X and Y are positive, X/(X+Y) < z tells us that z < 1, right?
 

Similar threads

Replies
2
Views
2K
Replies
9
Views
3K
Replies
6
Views
1K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K