0

I need some help for notation. I need to present fractions in four different format and I'd like to get it right. I just take $1/999$ for example, but of course it could be any fraction with positive integer denominator. So far my notation looks like this, leaving few parts undecided and unclear:

a)

$$1/999 = 0.001001001... = 0.\overline{001}$$

b)

$$.001 + .000001 + .000000001 + ...$$

How do you denote sequence like this up to infinity?

c)

$$1*10^{-(3*(n+0))} + 1*10^{-(3*(n+1))} + 1*10^{-(3*(n+2))} + ...$$

Again I don't know how to denote continuity and increment of the power $3n$.

d)

$$\sum_{k=1}^{\infty} \frac{(10^{\lfloor\strut\log_{10}(p)\rfloor+1} - p)^{k-1}}{(10^{\lfloor\strut\log_{10}(p)\rfloor+1})^k}$$

where $p = 999$

Summation part is pretty much ok, except i don't know how to incorporate p = 999 part on it and ow to tie everything together.

After all, my purpose is to show that $a = b = c = d$ which context is why I prefer to ask them on one question rather than all on different questions.

MarkokraM
  • 671

1 Answers1

1

a) $1/999 = 0.001001001\ldots = 0.\overline{001}$

Either of these is fairly standard notation. The overline format $0.\overline{001}$ is a little more explicit, so I think it would be preferred.

b) $.001 + .000001 + .000000001 + \cdots$

This denotes an infinite sequence in the way that $1, 2, 3, \ldots$ indicates an infinite sequence: a little informally, but I think it is recognizable as a mathematical notation.

The main problem with this one is it is hard to count the zeros. One way to help see how big each number is is to consider a power of $0.001$, like this:

$$0.001^1 + 0.001^2 + 0.001^3 + \cdots.$$

Using the $\sum$ summation notation, this is

$$\sum_{k=1}^\infty 0.001^k.$$

Alternatively, you could write this as

$$10^{-3} + 10^{-6} + 10^{-9} + \cdots$$

but now it is starting to look like

c) $1 \cdot 10^{-(3 \cdot (n+0))} + 1 \cdot 10^{-(3 \cdot (n+1))} + 1 \cdot 10^{-(3 \cdot (n+2))} + \cdots$

But there's a problem here: what is $n$? If you set $n=1$ then this evaluates to $10^{-3 \cdot 1} + 10^{-3 \cdot 2} + 10^{-3 \cdot 3} + \cdots$ which is OK. If you set $n$ to any other value then the sum will not add up to $1/999$.

In $\sum$ notation, this series is

$$\sum_{k=0}^{\infty} 10^{-3 \cdot (n + k)}$$

and we can rewrite it as follows:

$$\sum_{k=0}^{\infty} 10^{-3 \cdot (n + k)} = 10^{-3n} \sum_{k=0}^{\infty} \left( 10^{-3} \right)^k = 10^{-3n} \left( 1 + 0.001 + 0.000001 + \cdots \right). $$

d) $\displaystyle\sum_{k=1}^{\infty} \frac{(10^{\lfloor\strut\log_{10}(p)\rfloor+1} - p)^{k-1}}{(10^{\lfloor\strut\log_{10}(p)\rfloor+1})^k}$

This looks really complicated, but let's try setting $N = 10^{\lfloor\strut\log_{10}(p)\rfloor+1}$. Then

\begin{align} \sum_{k=1}^{\infty} \frac{(10^{\lfloor\strut\log_{10}(p)\rfloor+1} - p)^{k-1}}{(10^{\lfloor\strut\log_{10}(p)\rfloor+1})^k} & = \sum_{k=1}^{\infty} \frac{(N - p)^{k-1}}{N^k} \\ & = \sum_{k=0}^{\infty} \frac{(N - p)^k}{N^{k+1}} \\ & = \frac 1N \sum_{k=0}^{\infty} \left(1 - \frac pN\right)^k \\ \end{align}

The sum in this last equation is a geometric series $\sum r^k$ with ratio $r = \left(1 - \frac pN\right)$. The limit of such a series starting from the $k=0$ term is $\sum_{k=0}^{\infty} r^k = \frac{1}{1-r}$. Applying this result, we have

$$\sum_{k=0}^{\infty} \left(1 - \frac pN\right)^k = \frac{1}{1 - \left(1 - \frac pN\right)}$$

and therefore

\begin{align} \sum_{k=1}^{\infty} \frac{(10^{\lfloor\strut\log_{10}(p)\rfloor+1} - p)^{k-1}}{(10^{\lfloor\strut\log_{10}(p)\rfloor+1})^k} & = \frac 1N \left(\frac{1}{1 - \left(1 - \frac pN\right)}\right) \\ & = \frac 1N \left(\frac Np\right) \\ & = \frac 1p. \end{align}

So it turns out it doesn't make much difference what $N$ is.

Edit: the rest of this answer has been revised to reflect upon the related question, Doubling sequences of the cyclic decimal parts of the fraction numbers.

Now, it turns out that in the case where $p = 999$, the series in part d) is term-for-term the same as the series in part b), because $10^{\lfloor\strut\log_{10}(999)\rfloor+1} - 999 = 1$. The sum then comes down to

$$\sum_{k=1}^{\infty} \frac{1^{k-1}}{(10^{\lfloor\strut\log_{10}(999)\rfloor+1})^k} = \sum_{k=1}^{\infty} \frac{1}{1000^k} = \sum_{k=1}^{\infty} 0.001^k$$

which is a more explicit way of writing the series $0.001^1 + 0.001^2 + 0.001^3 + \cdots$.

But as the calculations above showed, the formula for part d) works even when $p$ is not all $9$s (that is, even when $p \neq 16^n - 1$ for any integer $n$). For example, take $p = 98$. The formula then says

$$\frac{1}{98} = \sum_{k=1}^{\infty} \frac{(100 - 98)^{k - 1}}{100^k} = \sum_{k=1}^{\infty} \frac{2^{k-1}}{100^k} = 0.01 + 0.0002 + 0.000004 + 0.00000008 + \cdots$$

which well illustrates the "powers of $2$" pattern within the repeating $42$-digit sequence in

$$\frac{1}{98} = 0.0\overline{102040816326530612244897959183673469387755}.$$

More could be said along these lines, but the question Doubling sequences of the cyclic decimal parts of the fraction numbers is a better context for that.

David K
  • 98,388
  • Do you have time for a little chat about your answer? I remember on some SE site it was possible to have a conversation screen, but I can't see it here... – MarkokraM Jul 24 '15 at 09:31
  • hope you have time to read and comment at least the last part of this: https://stackedit.io/viewer#!provider=gist&gistId=2eec41d340c0c0970c86&filename=Decimal+expansion+and+notation – MarkokraM Jul 24 '15 at 11:19
  • @MarkokraM There's a "chat" button at the bottom of this Web page that enables creating a new room. I did this for http://chat.stackexchange.com/rooms/26191/decimal-notations-for-fractions. (The stackedit link did not seem to work for me.) – David K Jul 24 '15 at 11:39
  • I will add it. Is it possible to show by which principle you get sigma sign out of the equation on part d? – MarkokraM Jul 25 '15 at 11:31
  • I have changed how this answer ends, now that I know what the reason for part d) is. I think this is about as much as we need to say about the notational question. Of course there is a lot more to say about the "doubling sequences" question that inspired this one, and I will try to post some kind of answer to that question too. (It may just take a little while longer to do that.) – David K Jul 25 '15 at 14:56
  • Brilliant, that is more than enough for OP. I'm looking forward answer for other wider "doubling sequences" topic. At first I didn't want to mix it with the current one, but true, that is what I'm after. – MarkokraM Jul 25 '15 at 17:19