While reviewing multivariate calculus with a student, I encountered this exercise:
Evaluate the integral $\displaystyle\iint_We^{xy}\,\mathrm dx\,\mathrm dy$, where $W$ is the region bounded by $1\le xy\le4$ and $1\le\dfrac xy\le2$.
I made the change of coordinates
$$\begin{cases}u(x,y)=xy\\[1ex]v(x,y)=\dfrac xy\end{cases}$$
for which the Jacobian and its determinant are
$$\mathbf J=\begin{bmatrix}y&x\\[1ex]\dfrac1y&-\dfrac x{y^2}\end{bmatrix},\quad\det\mathbf J=-\dfrac{2x}y=-2v$$
So the integral is
$$\iint_We^{xy}\,\mathrm dx\,\mathrm dy=2\int_{v=1}^{v=2}\int_{u=1}^{u=4}ve^u\,\mathrm du\,\mathrm dv=3e^4-3e$$
I check my answer with Mathematica to find out the answer instead appears to be
Integrate[
E^(x y) Boole[1 <= x y <= 4 && 1 <= x/y <= 2],
{x, 0, Infinity}, {y, 0, Infinity}
]
which returns a different answer of $\dfrac{(e^4-e)\log2}2$.
Is there a mistake in my work? Or is Mathematica correct? Or is the integral I'm computing with
Integratenot equivalent to the one I'm supposed to compute?
Yes indeed, as you noted, we have
$$du\,dv=|\det\mathbf J|dx\,dy=-\dfrac{2x}y=-2v\implies dx\,dy =-\frac1{2v}du\,dv$$
and it should be
$$\color{red}{\frac12}\int_{v=1}^{v=2}\int_{u=1}^{u=4}\frac{e^u}v\,\mathrm du\,\mathrm dv$$