Assume that you have two functions $f^-,f^+ :(\mathbb R^+)^n \rightarrow \mathbb R$. Also assume that every integral I use converges. Assume that we have the following inequality for every $(a_1,...,a_n) \in (\mathbb R^+)^n$: \begin{align*} \int\limits_{a_1}^{\infty}...\int\limits_{a_n}^{\infty} f^-(x_1,...,x_n)dx_1...dx_n \leq \int\limits_{a_1}^{\infty}...\int\limits_{a_n}^{\infty} f^+(x_1,...,x_n)dx_1...dx_n \end{align*}

Now suppose that we have a function $h:(\mathbb R^+)^n \rightarrow \mathbb R^+$ which is monotonously increasing in every component. Does then the following hold:

\begin{align*} \int\limits_{a_1}^{\infty}...\int\limits_{a_n}^{\infty} h(x_1,...,x_n) \cdot f^-(x_1,...,x_n)dx_1...dx_n \leq \\ \int\limits_{a_1}^{\infty}...\int\limits_{a_n}^{\infty} h(x_1,...,x_n) \cdot f^+(x_1,...,x_n)dx_1...dx_n \end{align*}

The intuition why I believe that this is true is the following:

$f^-$ may not be always smaller than $f^+$ but independently of how late we start the integral, the integral over $f^+$ remains bigger. So independently of how late we start, $f^+$ is "most of the times bigger than $f^-$ or it is drastically bigger than $f^-$ sometimes". So what is the worst that multiplying with $h$ can do? That is, that when for a while $f^-$ happens to be bigger than $f^+$ for these values $h$ should be huge. So if we hadn't the monotonicity of $h$ then we could make $h$ huge on the described sets and small otherwise. But here every increase in $h$ ultimately increases the value of the integral with $f^+$ because $h$ can't become smaller again, and in the long run, $f^+$ is bigger more often / drastically. Thus I think the claim should be true.

Thanks in advance!


Your Answer

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

Browse other questions tagged or ask your own question.