Is it true that $\displaystyle \int_0^{\infty} 2^{-x}\,dx=\sum_{x=0}^{\infty}\lim_{k\to\infty}\dfrac{2^{-x}}{k}$?

Here is my try:

\begin{align} & \sum_{x=0}^{\infty}\lim_{k\to\infty}\dfrac{2^{-x}}{k}\\ &= \lim_{k\to\infty}\sum_{x=0}^{\infty}\dfrac{2^{-x}}{k} \\ &= \lim_{h\to0+}\sum_{x=0}^{\infty}2^{-x}h \\ &= \int_0^{\infty} 2^{-x}\,dx \\ &= \frac{1}{\log{2}} \end{align}

Is it correct? I have doubts because doing

\begin{align} & \sum_{x=0}^{\infty}\lim_{k\to\infty}\dfrac{2^{-x}}{k}\\ &= \lim_{k\to\infty}\dfrac 1k\sum_{x=0}^{\infty}2^{-x} \\ &= \lim_{k\to\infty}\dfrac 2k \\ &= 0 \neq \frac 1{\log 2} \end{align}

I get another answer from the same equation.

  • 1
    $\begingroup$ no,RHS is equal to zero $\endgroup$ – haqnatural Nov 28 '17 at 19:50
  • $\begingroup$ What is $\sum_{x=0}^\infty $ $\endgroup$ – hamam_Abdallah Nov 28 '17 at 19:56

The fallacy is in assuming $\int_0^\infty f(x) dx=\lim_{h\to 0^+}\sum_{n=0}^\infty hf(n)$; the right-hand side should be $\lim_{h\to 0^+}\sum_{n=0}^\infty \frac{h}{2}(f(hn)+f(hn+h))$. You can convince yourself of this by drawing thin trapezia under a curve.


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.