Consider two positive real numbers $\frac{1}{a}$ and $\frac{1}{b}$. Show that the 'mean' $$\dfrac{1}{\frac{a+b}{2}}=\dfrac{2}{a+b}$$ will always be less than the arithmetic mean $$\dfrac{a+b}{2ab}.$$

Proving this is not difficult, but I was wondering if there was a name for this inequality, or if it demonstrates another known inequality in action, much like how many known inequalities are simple consequences of say the AM/GM Inequality.

  • 1
    $\begingroup$ It is probably more insightful if you replace $\frac 1a$ and $\frac 1b$ by $a$ and $b$. $\endgroup$ – Arnaud Mortier Aug 24 '18 at 6:55
  • $\begingroup$ @ArnaudMortier If we looked at $a$ instead of $1/a$, we would explicitly have to include a non-$0$ criterion. There’s a tradeoff. $\endgroup$ – gen-z ready to perish Aug 24 '18 at 7:18
  • 3
    $\begingroup$ @ChaseRyanTaylor since the numbers have to be positive either way, that's not an issue, at least in English. $\endgroup$ – Especially Lime Aug 24 '18 at 7:29

Yes this is the HM-AM inequality and it is a consequence of AM-GM.

Take a look here

RMS-AM-GM-HM inequalities

  • 1
    $\begingroup$ HM is for Harmonic Mean. $\endgroup$ – mathreadler Aug 24 '18 at 7:07
  • $\begingroup$ @mathreadler Yes exactly! $\endgroup$ – user Aug 24 '18 at 7:20

This is just saying that the inverse of the mean is at most the mean of the inverses:

$$\frac{1}{\frac{a+b}2} \leq \frac{\frac{1}{a} + \frac{1}{b}} 2$$

This follows from the convexity of the function $\frac 1 x$ (on the positive real numbers). Of course, convexity is closely related to the AM-GM inequality, so maybe it's not surprising that this can also been seen as related to AM-GM.


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.