Answer: Here we have 100 independent random variables which have an exponential distribution. As per the formula,
Hence in this case,
\[ f_z = (1/2) \lambda e^{-\lambda|z|} \]
Answer: Now for the convolution W = X + Y, we know the probability density would be: \[ f_W(w) = \int_{-\infty}^{\infty} f_X(x) f_Y(w-x) dx \]
Here Z = X1 - X2 = X1 + (-X2)
Hence \[ f_Z(z) = \int_{-\infty}^{\infty} f_X(x) f_{-Y}(z-x) dx \]
Now, we can safely say: \[ f_{-Y}(z-x) = f_Y(x-z) \]
Hence, \[ f_Z(z) = \int_{-\infty}^{\infty} f_X(x) f_Y(x-z) dx \]
Now for z < 0, \[ f_Z(z) = \int_{0}^{\infty} \lambda e^{-\lambda x} \lambda e^{-\lambda (x-z)} dx \]
We are taking 0 to \(\infty\) because \[ f_X(x) = 0 \] for x < 0 \[ f_Z(z) = \lambda e^{\lambda z} \int_{0}^{\infty} \lambda e^{-2\lambda x} dx \] \[ = (\lambda /2) e^{\lambda z} \] when z < 0
Z and -Z have same distribution.
Hence with that, for z \(\ge\) 0, we will have to take the negative the exponent \[ f_Z(z) = (\lambda /2) e^{-\lambda z} \] when z \(\ge\) 0
Combining both the ranges z < 0 and z \(\ge\) 0, we can say: \[ f_z = (1/2) \lambda e^{-\lambda|z|} \]