Remarkable subsets

Let A Z A \subset \mathbb Z be a non-trivial set of integers ( ( i.e. A A \not= \emptyset and A Z ) . A \not= \mathbb Z). We call such a set A A a " remarkable set of type N N " if it has the following properties:

  • If a a is an element of A , A, then a -a is also an element of A A .

  • If a a is an element of A A , then a + N a+N is also an elements of A A .

  • If a , b a,b are elements of A A (not necessarily different), then a + 2 b a + 2b is also an element of A A .

How many (non-trivial) remarkable sets of type 18 are there?


Bonus: Generalize. If N N = 2 e 2 3 e 3 5 e 5 2^{e_2}\cdot 3^{e_3}\cdot 5^{e_5}\cdots is the prime factorization of N N , how many non-trivial remarkable sets of type N N exist? What do they look like?


The answer is 8.

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.

2 solutions

Arjen Vreugdenhil
Jan 26, 2018

Let b b be the smallest positive element in A A . Then b A -b \in A by rule 1, and by rule 3, all k b A kb \in A for all odd k k .

There are now two possibilities:

  • 0 A 0 \in A , in which case (by rule 2 again) k b A kb \in A for all k k : A { , 4 b , 3 b , 2 b , b , 0 , b , 2 b , 3 b , 4 b , } A \supseteq \{\dots, -4b, -3b, -2b, -b, 0, b, 2b, 3b, 4b, \dots \}

  • 0 ∉ A 0 \not\in A , in which case k b A kb \in A for all odd k k . A { , 5 b , 3 b , b , b , 3 b , 5 b , } A \supseteq \{\dots, -5b, -3b, -b, b, 3b, 5b, \dots \}

In fact, equality holds in either case: set A A contains no other elements. Proof: if a A a \in A is not an odd multiple of k g kg , use rule 3 repeatedly to add or subtract 2 b 2b until you find a number a a' strictly between b -b and + b +b . If a < 0 a' < 0 , swap it out for a > 0 -a' > 0 , which must also be an element by rule 1. Then either a = 0 a' = 0 , showing that a = k b a = kb for even k k ; or 0 < a < b 0 < a < b , which contradicts the assumption that b b is minimal.

Now use rule 2 to conclude that N N should be a multiple of the "step size" in A A : b N b | N (first case) or 2 b N 2b | N (i.e. b 1 2 N b | \tfrac12N , provided N N is even) (second case). In the first case we must rule out b = 1 b = 1 because then A = Z A = \mathbb Z trivially. For N = 18 N = 18 we find

  • five sets of the first kind: b = 2 , 3 , 6 , 9 , 18 b = 2, 3, 6, 9, 18 A 1 = { , 8 , 6 , 4 , 2 , 0 , 2 , 4 , 6 , 8 , } A 2 = { , 12 , 9 , 6 , 3 , 0 , 3 , 6 , 9 , 12 , } A 3 = { , 24 , 18 , 12 , 6 , 0 , 6 , 12 , 18 , 24 , } A 4 = { , 36 , 27 , 18 , 9 , 0 , 9 , 18 , 27 , 36 , } A 5 = { , 72 , 54 , 36 , 18 , 0 , 18 , 36 , 54 , 72 , } A_1 = \{\dots, -8, -6, -4, -2, 0, 2, 4, 6, 8, \dots \} \\ A_2 = \{\dots, -12, -9, -6, -3, 0, 3, 6, 9, 12, \dots \} \\ A_3 = \{\dots, -24, -18, -12, -6, 0, 6, 12, 18, 24, \dots \} \\ A_4 = \{\dots, -36, -27, -18, -9, 0, 9, 18, 27, 36, \dots \} \\ A_5 = \{\dots, -72, -54, -36, -18, 0, 18, 36, 54, 72, \dots \}

  • three sets of the second kind: b = 1 , 3 , 9 b = 1, 3, 9 A 6 = { , 7 , 5 , 3 , 1 , 1 , 3 , 5 , 7 , } A 7 = { , 21 , 15 , 9 , 3 , 3 , 9 , 15 , 21 , } A 8 = { , 63 , 45 , 27 , 9 , 9 , 27 , 45 , 63 , } A_6 = \{\dots, -7, -5, -3, -1, 1, 3, 5, 7, \dots \} \\ A_7 = \{\dots, -21, -15, -9, -3, 3, 9, 15, 21, \dots \} \\ A_8 = \{\dots, -63, -45, -27, -9, 9, 27, 45, 63, \dots \}

The answer to the problem is: there exist eight \boxed{\text{eight}} remarkable sets of type 18 18 .


In general, the number of sets A A of the given description is n ( N ) = ( τ ( N ) 1 ) + τ ( 1 2 N ) , n(N) = (\tau(N) - 1) + \tau(\tfrac12 N), where τ ( x ) \tau(x) is the number of positive divisors of x x . Given the prime factor decomposition N = 2 e 2 3 e 3 5 e 5 p e p , N = 2^{e_2}\cdot 3^{e_3}\cdot 5^{e_5} \cdots p^{e_p}\cdots, this becomes n ( N ) = ( 2 e 2 + 1 ) ( e 3 + 1 ) ( e 5 + 1 ) ( e p + 1 ) 1. n(N) = (2e_2+1)\cdot (e_3 + 1)\cdot (e_5 + 1)\cdots(e_p+1)\cdots - 1. (Check: if N = 18 N = 18 then e 2 = 1 , e 3 = 2 e_2 = 1, e_3 = 2 , and n ( N ) = ( 2 + 1 ) ( 2 + 1 ) 1 = 3 3 1 = 8 n(N) = (2+1)\cdot (2+1) - 1 = 3\cdot 3 - 1 = 8 .)

I'd like to dive right into this with an unverified fibonacci take

Ryan Matthew - 3 years, 4 months ago

Log in to reply

That is how new problems are born: for which N N are τ ( N ) 1 \tau(N) - 1 and τ ( 1 2 N ) \tau(\tfrac12 N) successive Fibonacci numbers?

Arjen Vreugdenhil - 3 years, 4 months ago

What represent N in this problem ? Is it , Natural numbers ?

Pranshul Goyal - 3 years, 4 months ago

Log in to reply

In this problem, we have to search for the number of " remarkable sets of type N N " when N = 18 N = 18 . So in this case, the specific one, N = 18 N = 18 .

Javier Álvarez - 3 years, 4 months ago
Meneghin Mauro
Feb 8, 2018

Arjen has the brilliant solution, but I thought I would post the result of my little script

first column is the generator, following values are elements of the set between 0-18. I added 0 as optional value for odd generators, even generators already produce 0

0 0 18

1 1 3 5 7 9 11 13 15 17

1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

2 0 2 4 6 8 10 12 14 16 18

3 3 9 15

3 0 3 6 9 12 15 18

6 0 6 12 18

9 9

9 0 9 18

options: 9

This includes the trivial set Z, so the solution is 8

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...