0
$\begingroup$

I am studying Fubini theorem for riemann integrals and I was wondering if there are examples of a bounded function $f : A\times B \to \mathbb{R}^n$ such that $A \subset \mathbb{R}^n, B \subset \mathbb{R}^m$ are rectangles and such that one of the iterated integrals $\int_A f, \int_B$ f exist but the other does not, and even worst, if the both iterated integral exist but not the integral $\int_{A\times B} f$.

I do appreciate any suggestions, examples, etc!

Thanks a lot!

$\endgroup$
1
$\begingroup$

Hint: There exists $E\subset [0,1]^2,$ with $E$ dense in $[0,1]^2,$ such that each horizontal and vertical line intersects $E$ in at most finitely many points. Think about $f(x,y) = \chi_E(x,y).$

$\endgroup$
  • $\begingroup$ Thank you so much for being so present answering so well all questions I ve been posting $\endgroup$ – L.F. Cavenaghi May 15 '16 at 0:34
1
$\begingroup$

If we have $A=\mathbb{R}$, $B=\mathbb{R}$, take $f(x, y)=e^{-x^2}x$. We can show that $\int_{A}f(x, y)dx=0$. But $\int_Bf(x,y)dy=\lim_{n,m\to \infty}(n+m)f(x, y)\to \infty$.

For bounded $A$, $B$, we can also let $g(x)$ denote a nonintegrable function, then $g(x, y)=g(y)$ is constant in $X$, so $\int_{A}g(x, y)dx$ exists, but $\int_{B}g(x,y)dy=\int g(y)dy$ doesn't exist by assumption.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.