Some time ago, I ran across the following inequality (if I remember rightly): $\bigg(\sum_{i=1}^{n}v_{i}^{k_{1}}\bigg)\geq n\bigg(\frac{\sum_{i=1}^{n}v_{i}^{k_{2}}}{n}\bigg)^{\frac{k_{1}}{k_{2}}}$, for $v_{1},...,v_{n}$ positive real numbers and $k_{1},k_{2}$ real numbers, with $k_{1}\geq k_{2}$. Where did it originate? Does it have a name?

It is a direct consequence of Hölder's inequality for a discrete probability measure. In Hardy-Littlewood-Pólya it is listed as Theorem 16, and doesn't have a special name. They credit the inequality to Schlömilch (1858), Reynaud and Duhamel (1823), and Chrystal (1900).

A better way to formulate the inequality is to define the $\ell_p$ weighted mean of a finite list of positive numbers $a = (a_1, \ldots, a_n)$ by $$ \mathfrak{M}_p(a) = \left( \frac{1}{n} \sum_{i = 1}^n a_i^p \right)^{1/p}$$ and the inequality states $$ \mathfrak{M}_p \leq \mathfrak{M}_q $$ whenever $q \geq p$. It fits in the general hierarchy of inequality of means, which also includes the inequality of arithmetic and geometric means.

Your Answer


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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