What is the distribution of $X+Y$ where $X \sim U(0,\frac{L}{2})$ and $Y \sim U(\frac{L}{2},L)$?

84 Views Asked by At

I started along these lines:

Let $Z = X + Y$ where $\frac{L}{2}< z < \frac{3L}{2}$, then,

$$f_{X+Y}(z)=f_{Z}(z) = \int f_{X}(x)f_{Y}(z-x)dx$$

However, I am not sure how to fill in the bounds of the integral to properly account for the domains of $X$ and $Y$. Also, I think $f_{Z}(z)$ needs to computed via 2 cases:

1) $\frac{L}{2}< z < L$

2) $L< z < \frac{3L}{2}$

Since in case (1) $X$ can take any value in its domain, whereas in case (2) $X$ must take the value $\frac{L}{2}$.

Regardless, I am having a hard time properly setting up the integrals. Any insight would be appreciated.

Update: After some more thinking, for case (1), I think this is the integral:

$$f_{Z}(z) = \int_{0}^{z} f_{X}(y-z)f_{Y}(y)dy$$

And for case (2):

$$f_{Z}(z) = \int_{z}^{\frac{3L}{2}} f_{X}(X=\frac{L}{2})f_{Y}(z-\frac{L}{2})dy$$

However, I am unsure of the above.

1

There are 1 best solutions below

0
On

$P(X+Y \le z)=P(X\le z-Y)$

Therefore,

$F(z) = \int_{L/2}^L P(X \le z-y) \frac{2}{L} dy$

Now, it is obvious that:

$$P(X \le z-y) = 0 \text{ when } z-y<0$$

$$P(X \le z-y) = \frac{2(z-y)}{L} \text{ when } 0\le z-y \le \frac{L}{2}$$

$$P(X \le z-y) = 1 \text { when } z-y \ge \frac{L}{2}$$

Therefore, it follows that:

$$F(z) = 0 \text{ when } z \le \frac{L}{2}$$

$$F(z) = \int_{L/2}^L \frac{2(z-y)}{L} \frac{2}{L} dy \text{ when } \frac{L}{2} \le z \le L$$

$$F(z) = \int_{L/2}^L \frac{2}{L} dy \text{ when } L \le z \le \frac{3L}{2}$$