Let $b_{1},b_{2},\ldots,b_{k}$ be nonnegative numbers and $p_{1} + p_{2} + \ldots + p_{k} = 1$ where each $p_{i}$ is positive. Then

\begin{align*} \sum_{i=1}^{k}p_{i}b_{i}\geq\prod_{i=1}^{k}b^{p_{i}}_{i} \end{align*}


Since the logarithm function is strictly increasing, the proposed inequality is equivalent to \begin{align*} \ln\left(p_{1}b_{1} + p_{2}b_{2} + \ldots + p_{k}b_{k}\right) \geq p_{1}\ln(b_{1}) + p_{2}\ln(b_{2}) + \ldots + p_{k}\ln(b_{k}) \end{align*}

Once $f''(x) < 0$, where $f(x) = \ln(x)$, we conclude that $f$ is concave and the proposed inequality holds.

My question is: am I proving this result correctly? If this is the case, is there another way to prove it?

Any contribution is appreciated.

  • 2
    $\begingroup$ You have done it absolutely correctly. Another method is by using weighted AM-GM as shown in my answer that I'm typing. $\endgroup$ Aug 6, 2020 at 22:17

1 Answer 1


Using $b_1,b_2,\cdots,b_k\ge 0$ with $p_1,p_2,\cdots,p_k$ as the respective weights, weighted AM-GM inequality gives $$\frac{\sum_{i=1}^k p_ib_i}{\sum_{i=1}^k p_i} \ge \left(\prod_{i=1}^k b_i^{p_i}\right)^{\dfrac{1}{\sum_{i=1}^k p_i}}$$ which gives the required inequality.

Note: This is not much different from your approach because the special case of Jensen's inequality for the function $f(x)=\ln(x)$ can be proved by weighted AM-GM or vice-versa, the weighted AM-GM inequality can be proved by the Jensen's inequality for $f(x)=\ln(x)$ as you have done, since the inequality you want to prove is directly the weighted AM-GM inequality in disguise. Also, these two can be proven independently without using each other.

  • 1
    $\begingroup$ Thanks for the contribution! $\endgroup$
    – user0102
    Aug 6, 2020 at 22:25

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.