Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I want to construct a sequence of rational numbers whose sum converges to an irrational number and whose sum of absolute values converges to 1.

I can find/construct plenty of examples that has one or the other property, but I am having trouble find/construct one that has both these properties.

Any hints(not solutions)?

share|cite|improve this question

Find two irrational numbers $a > 0$ and $b < 0$ such that $a-b = 1$ but $a+b$ is irrational. Create a series with positive terms that sum to $a$ and another series with negative terms that sum to $b$. Combine the two series.

share|cite|improve this answer
would n't combining the terms of the two series, suffer from the issues one encounters rearranging the terms in a series which causes changes to the sum of the series. – Comic Book Guy Oct 27 '12 at 20:29
The series will be absolutely convergent and can be rearranged at will. – mrf Oct 27 '12 at 20:32
that's a good point. Thanks – Comic Book Guy Oct 27 '12 at 20:36
Great, it worked out! Tanks! – susan Oct 27 '12 at 21:22
this is a great example!!! – Wenkui Liu Jan 23 at 21:25

Here is a probabilistic way. Consider an i.i.d. sequence $(X_n)_{n\geqslant1}$ of Bernoulli random variables such that $\mathbb P(X_n=+1)=\mathbb P(X_n=-1)=\frac12$. Then $X=\sum\limits_{n\geqslant1}2^{-n}X_n$ is uniformly distributed on $[-1,1]$ hence $X$ is irrational with probability $1$. And naturally, $\sum\limits_{n\geqslant1}2^{-n}|X_n|=1$.

A deterministic way is as follows. Pick any irrational number $z$ in $(0,1)$, with binary expansion $z=\sum\limits_{n\geqslant1}z_n2^{-n}$, where each $z_n$ is in $\{0,1\}$. For every $n\geqslant1$, consider $x_n=(2z_n-1)2^{-n}$. Then $\sum\limits_{n\geqslant1}x_n=2z-1$ is irrational and $|2z_n-1|=1$ for every $n\geqslant1$ hence $\sum\limits_{n\geqslant1}|x_n|=1$.

share|cite|improve this answer
This seems unnecessarily complicated to me, but I find it very interesting that it's your first response! – Benjamin Dickman Oct 27 '12 at 20:29
Thanks! Both nice solutions! – susan Oct 27 '12 at 21:23

Let $\alpha$ be an irrational number in $(0,1)$. Then for suitable choice of $\epsilon_n\in\{\pm1\}$, you can achieve $$\sum_{n=1}^\infty \frac{\epsilon_n}{2^n}=\alpha\qquad\text{and}\qquad\sum_{n=0}^\infty \frac{1}{2^n}=1.$$ To do so, define $\epsilon_n$ recursively: If $\epsilon_k$ is already known for $k<n$ such that $|s_{n-1}-\alpha|<\frac{1}{2^{n-1}}$ for the partial sum $s_{n-1}=\sum_{k=1}^{n-1}\frac{\epsilon_k}{2^k}$, select $\epsilon_n=+1$ if $\alpha>s_{n-1}$ and $\epsilon_n=-1$ if $\alpha<s_{n-1}$. The case $\alpha=s_{n-1}$ does not occur, because $\alpha$ is irrational. With this choice, $|s_n-\alpha|=|s_{n-1}+\frac{\epsilon_n}{2^n}-\alpha|= |\frac1{2^n}-\frac{\alpha-s_{n-1}}{\epsilon_n}|<\frac 1{2^n}$ because $<\frac{\alpha-s_{n-1}}{\epsilon_n}<2\cdot\frac 1{2^n}$.

Another way to look at this is: The $\frac{1-\epsilon_n}2\in\{0,1\}$ are essentially the binary expansion of $\frac{1-\alpha}2$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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