Consider the sum of the reciprocals of every positive integer that doesn't contain the digit 2:
1 1 + 3 1 + 4 1 + ⋯ + 1 0 1 + 1 1 1 + 1 3 1 + ⋯ + 1 9 1 + 3 0 1 + 3 1 1 ⋯ .
Does this sum converge?
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Why doesn't this trick work when we are only removing those entries with the digit 0?
Log in to reply
The series with the digit 0 removed will still converge (to ≈ 2 3 . 1 ).
I'm not sure what the "fun fact" is referring to. We could prove it in a similar manner, but be careful with counting the number of terms.
Oh! I'm sorry. Calvin's right; it still does converge with all those containing 0 removed. I somehow didn't remove them all while doing a similar pattern as above. Let me edit that.
Log in to reply
What if we were to remove a specific string of digits? Would it still converge if we were not allowed to use integers which contain (say) 314159?
Log in to reply
@Calvin Lin – It still will. In such case, we are bound to consider that b n < 1 0 k n terms, where k is the number of digits of the substring to be removed in the series.
Log in to reply
@Efren Medallo – That's the right idea, but it think you have a typo in the inequality. Note that for k = 1 (and the digit is non-zero), we have ∣ b n ∣ < 9 n . We need lim b n b n + 1 < 1 0 in order for this simple counting argument to work directly.
Fix a positive integer d , and consider all k satisfying the given condition having exactly d digits. There are 8 ⋅ 9 d − 1 such k (the first digit of k has 8 choices: all except 0 and 2; the other digits have 9 choices: anything except 2). All these k , being having d digits, must be at least 1 0 d − 1 . Thus all these k contribute at most 1 0 d − 1 8 ⋅ 9 d − 1 to the sum; each of these k 's contribute k 1 ≤ 1 0 d − 1 1 , and there are 8 ⋅ 9 d − 1 of them. Summing over all d , we have that the sum S satisfies
S ≤ d = 1 ∑ ∞ 1 0 d − 1 8 ⋅ 9 d − 1 = 8 ⋅ d ′ = 0 ∑ ∞ 1 0 d ′ 9 d ′ = 8 ⋅ 1 0 = 8 0
Thus S has an upper bound. Since all terms of S are increasing, by monotone convergence theorem S converges .
This can be generalized to ruling out any finite sequence of decimal digits. For example, if we take out all terms having the sequence 2 0 1 7 , the series will also converge.
Neat proof!
Problem Loading...
Note Loading...
Set Loading...
Let us divide the sum S into a different sub-series we'll call b n , such that b n is the sum of all elements of S of denominator k for 1 0 n − 1 ≤ k < 1 0 n .
So from this, we can get that
b 1 = 1 + 3 1 + 4 1 + . . . + 9 1 b 2 = 1 0 1 + 1 1 1 + . . . + 9 9 1
and so on.
Now we investigate the number of terms in each of b n .
It will be easy to note that b 1 has less than 9 terms, or that b 1 < 9 .
Similarly, we can also see that b 2 has less than 9 2 terms, and that b 2 < 1 0 9 2 .
In general, b n will have less than 9 n terms, and b n < 1 0 n − 1 9 n .
From here we can deduce that
S < 9 + 1 0 9 2 + 1 0 2 9 3 + . . .
with which the RHS depicts a geometric series with common ratio 1 0 9 .
So, we can say that S < 9 0 . Thus, the series is convergent .
*Notes:
S = 9 0 .
Fun fact : This also works for any series of the same format, but with any other digit removed (say, 6 , or 7 ).