Déjà vu?

Find the smallest positive integer that cannot be written as

2 a 2 b 2 c 2 d \dfrac{2^a-2^b}{2^c-2^d}

for positive integers a a , b b , c c , and d d .


The answer is 11.

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.

3 solutions

Patrick Corn
May 27, 2014

WLOG we can assume a > b a > b and c > d c > d (we can switch them if necessary, and the quotient has to be positive). Now note that b d b \ge d because otherwise the quotient is not an integer. Divide through by 2 d 2^d to get something of the form 2 x 2 y 1 2 z 1 2^x \frac{2^y-1}{2^z-1} , where x x is nonnegative and y , z y,z are positive.

It's clear that this is an integer if and only if z y z|y . In that case, we can rewrite as 2 x ( 1 + 2 z + 2 2 z + + 2 y z ) 2^x(1+2^z+2^{2z}+\cdots+2^{y-z}) . So the integers that can be written as a quotient like this are precisely the integers which, when we write their binary expansion, have their 1 1 's spaced evenly apart.

To get the smallest integer that fails this criterion, we clearly need at least three 1 1 's in the expansion, and since 111 111 works (with z = 1 z = 1 and y = 3 y = 3 ), we need at least four binary digits. This leads to 1101 1101 and 1011 1011 ; the smaller of these, 1011 1011 , is the decimal number 11 \fbox{11} .

Very nice interpretation. This gives a description of all such numbers.

"It's clear that this is an integer if and only if z y z \mid y " could use a bit more explanation.

Calvin Lin Staff - 7 years ago

Log in to reply

Ok. First, 2 x 2 y 1 2 z 1 2^x \frac{2^y-1}{2^z-1} is an integer if and only if ( 2 z 1 ) 2 x ( 2 y 1 ) (2^z-1) | 2^x(2^y-1) , but 2 z 1 2^z-1 and 2 x 2^x are relatively prime, so this happens if and only if ( 2 z 1 ) ( 2 y 1 ) (2^z-1)|(2^y-1) . Certainly z y z|y is sufficient. To see that it is necessary, write y = q z + r y=qz+r with 0 r < z 0 \le r < z and note that ( 2 z 1 ) ( 2 y 1 ) (2^z-1) | (2^y-1) if and only if ( 2 z 1 ) ( 2 r 1 ) (2^z-1) | (2^r-1) (e.g. look mod ( 2 z 1 ) (2^z-1) ). But 2 z 1 > 2 r 1 2^z-1 > 2^r-1 so the only possibility is r = 0 r = 0 .

Patrick Corn - 7 years ago

Log in to reply

Wow we both posted two not entirely different arguments at the exact same time. :P

It is easy to verify that the order of 2 ( m o d 2 j 1 ) 2 \pmod{2^j-1} is j . j. If m m is any positive integer smaller than j , j, 2 m 1 < 2 j 1 , 2^m-1 < 2^j-1, so 2 m ≢ 1 ( m o d 2 j 1 ) . 2^m \not \equiv 1 \pmod{2^j-1}. Hence, ord 2 ( 2 j 1 ) = j . \text{ord}_2 (2^j-1) = j. It follows that 2 k 1 ( m o d 2 j 1 ) 2^k \equiv 1 \pmod{2^j-1} if and only if k k is a multiple of j . j.

I agree, really nice solution! I hadn't thought of generalizing my arguments.

tfyuidsgvcjkdysitfkA

Anuj Shikarkhane - 6 years, 10 months ago

hey i don't know what u did there ........... but if we take a=4 b=2 c=3 and d=1 then the equation solves to 2 which is smaller than 11

Abhishek Kalra - 7 years ago

Log in to reply

The problem says "cannot be written," not "can be written."

Patrick Corn - 7 years ago

What a solution . Just perfect. Keep it up

Mayank Chaturvedi - 7 years ago

Call a positive integer n n good if it can be expressed as 2 a 2 b 2 c 2 d \dfrac{2^a-2^b}{2^c-2^d} for some positive integers a , b , c , d . a,b,c,d.


We first make the following observations.

  • All powers of 2 2 are good.

Proof: Note that 2 n = 2 n + 2 2 n + 1 2 2 2 . 2^n = \dfrac{2^{n+2} - 2^{n+1}}{2^2 - 2}. \blacksquare

  • All numbers of the form 2 k 1 2^k-1 are good.

Proof: Note that 2 k 1 = 2 k + 1 2 2 2 2 . 2^k - 1 =\dfrac{2^{k+1} - 2}{2^2-2}. \blacksquare

  • All numbers of the form 2 k + 1 2^k+1 are good.

Proof: Note that 2 k + 1 = 2 2 k + 1 2 2 k + 1 2 . 2^k+1 = \dfrac{2^{2k+1} - 2}{2^{k+1} - 2}. \blacksquare

  • n n is good if and only if its largest odd factor is good.

Proof: Let j j be the largest odd factor of n , n, and let n = 2 m j . n= 2^m j. Let n = 2 a 2 b 2 c 2 d . n= \dfrac{2^a- 2^b}{2^c-2^d}. Then, j = 2 a 2 b 2 c + m 2 d + m . j= \dfrac{2^a-2^b}{2^{c+m} - 2^{d+m}}. Conversely, if j = 2 p 2 q 2 r 2 s j = \dfrac{2^p - 2^q}{2^r-2^s} is good, so is n = 2 m j = 2 p + m 2 q + m 2 r 2 s . n = 2^mj = \dfrac{2^{p+m} - 2^{q+m}}{2^r - 2^s}. \blacksquare


By the above observations, we immediately reject 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10. We shall now show that 11 11 is not a good number. Suppose 11 = 2 a 2 b 2 c 2 d , 11 = \dfrac{2^a-2^b}{2^c-2^d}, where a > b a>b and c > d . c>d. Rearrange it as 2 b ( 2 a b 1 ) = 11 × 2 d ( 2 c d 1 ) . 2^b (2^{a-b} - 1) = 11 \times 2^d (2^{c-d} - 1). Matching the highest powers of 2 2 in both sides, b = d . b=d. Then, the equation becomes 2 b ( 2 a b 1 ) = 11 × 2 b ( 2 c b 1 ) . 2^b (2^{a-b}-1) =11 \times 2^b (2^{c-b} - 1). Again, rewrite the equation as 2 a b + 10 = 11 × 2 c b . 2^{a-b} + 10 = 11 \times 2^{c-b}. Note that c b 1 , c-b \leq 1, since otherwise we would have that 4 11 × 2 c b 4 \mid 11 \times 2^{c-b} and 4 2 a b , 4 \mid 2^{a-b} , which implies 4 10. 4 \mid 10. Again, since c > d = b , c>d=b, c b 0. c-b \neq 0. So it follows that c b = 1. c-b=1. Our equation becomes 2 a b = 12 , 2^{a-b} = 12, not possible.


In conclusion, the answer is 11 . \boxed{11}.

why not 5 ?

Omar El Mokhtar - 7 years ago

Log in to reply

Because 5 = 2 2 + 1. 5= 2^2+1.

Prove that product of 4consecutive integers plus 1 is a perfect square

Anubhav Tyagi - 7 years ago

Log in to reply

Why are you posting this here?

Let the four integer be (n-1), n, (n+1), (n+2) Acc. to ques. (n-1)×n×(n+1)×(n+2)+1 = ( n 2 n + 1 ) 2 (-n^{2}-n+1)^{2}

Lu Chee Ket
Jun 2, 2014

2, 4, 6, 8, 12, 14, 16, 24, 28, 30,32, 48, 50, 56, 60, 62, 64, 92, 96, ...; multiplied with {11, 13, 19, 22, 26, ...} cannot be on them; 11 is hence the smallest answer.

C.K. Lu, 40, Malaysia

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...