a)

First we'll calculate cumulative distribution function of `Y=sqrt(B)` and that can be done this way:

`F_Y(x)=P(sqrt(B) leq x)=P(B leq x^2) = 1-e^(-ax^2)`

The second equality is valid because we are interested only in `x>0.`

Now to get density function of `Y` you simply derivate distribution function.

`f_Y(x)=(dF_Y)/(dx) (x) = 2ae^(-ax^2)x`

b)

This case is very similar, but here we have `Y=1/a ln(B)` so:

`F_Y(x)=P(1/a ln(B) leq x)=P(B leq e^(ax)) = 1-e^(-ae^(ax))`

`f_Y(x)=(dF_Y)/(dx) (x) = a^2 e^(-ae^(ax)+ax)`