Find the sum of all positive integers n such that for some positive integer m ≥ n , the following inequality is satisfied for all real x : { 1 5 x } − { n x } − { m x } ≤ 0 .
Details and assumptions
The function { a } denotes the fractional part of a . It can be calculated as { a } = a − ⌊ a ⌋ . As an explicit example, { 5 } = 0 , { 3 . 7 5 } = 0 . 7 5 , { − 4 . 2 } = 0 . 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.
Very nicely done! This was a very difficult problem, it does not have a short solution.
This can most probably be generalized to all other positive integers in place of 1 5 , but the ideas required for that are substantially deeper.
Very nice. I got a bit lazy after reducing to the following assumptions:
(a) m ≥ n > 1 5 , since for n ≤ 1 5 we can always pick m = 1 5 ;
(b) m 1 + n 1 ≥ 1 5 1 . [ This follows by letting x be a small positive number, e.g. x = 2 1 . ]
Upon these, there're only finitely many cases left for ( m , n ) . Indeed, we have
1 5 1 ≤ m 1 + n 1 ≤ n 1 = n 2
so n ≤ 3 0 . Each case of n then gives only finitely many possibilities of m . So in theory, we can just brute force all ( m , n ) . Furthermore, we can restrict ourselves to the case where x = k is an integer, since if the condition holds for x = k , then it holds for all x ∈ [ k , k + 1 ) thanks to condition (b).
This reduces to the case of finitely many possibilities for ( m , n , k ) , since we only need to consider k modulo lcm ( m , n , 1 5 ) .
[ Curious question: are any solutions of ( m , n ) for m < 1 5 and n = 1 5 ? ]
Log in to reply
Suppose there was a solution m , n with n < 1 5 and m = 1 5 .
Putting x = n shows that we cannot have m > 1 5 .
Suppose then that m < 1 5 . We still must also have that 1 5 divides { m , n } . Moreover, since 1 5 does not divide n , we have ( m , n ) > 1 .
One of the numbers must be a multiple of 5 , but cannot be 5 (since that would force the other number to be 1 5 ). Hence one of the numbers is 1 0 . The other number must be an even number of 3 . Thus there are only two possibilities:
n = 6 , m = 1 0 . This does not work ( x = − 6 ).
n = 1 0 , m = 1 2 . Thus does not work ( x = 1 2 ).
Nicely done, Mark H.
My solution had several small differences (that I won't get into) and one more significant difference: namely, I proved that if n isn't a multiple of 15, then 15n must be a multiple of m. (Consider x = n ( 1 5 λ ± 1 ) for integer λ .) I found that handy in a number of places.
Because the given inequality is satisfied for all real x, we consider two cases:
Case 1: m=15, then the given condition is obviously satisfied, we get 15 possible values of n, from 1 to 15.
Case 2: m ≥ n > 1 5 .
We have: { 1 5 x } -{ m x }-{ n x } ≤ { 1 5 x } -{ m x + n x }
≤ { 1 5 x − m x − n x }.
The given condition is satisfied for all real x when 1 5 1 = m 1 + n 1 .
We get (m,n)=(16,240);(18,90);(20,60);(24,40);(30,30).
The answer is: 2 1 5 ∗ 1 6 + 1 6 + 1 8 + 2 0 + 2 4 + 3 0 = 2 2 8
Why are these the only possible values of n and m ?
I didn't understood your Case 2(inequility)
Log in to reply
Since ⌊ x ⌋ + ⌊ y ⌋ ≤ x + y , it follows that ⌊ x ⌋ + ⌊ y ⌋ ≤ ⌊ x + y ⌋ for all real x , y , and from this it follows that { x + y } ≤ { x } + { y } for all x , y . The inequality uses this result twice.
Problem Loading...
Note Loading...
Set Loading...
It is clear that any 1 ≤ n ≤ 1 5 will do, since we can choose m = 1 5 .
Suppose then that m ≥ n ≥ 1 6 . I first claim that the result holds for any m ≥ n ≥ 1 6 with 1 5 1 = m 1 + n 1 . If this identity holds, then ⌊ 1 5 x ⌋ − ⌊ n x ⌋ − ⌊ m x ⌋ + { 1 5 x } − { n x } = { m x } If { 1 5 x } − { n x } ≥ 0 , it is clear that it is equal to { m x } . If { 1 5 x } − { n x } < 0 , it is clear that it is equal to { m x } − 1 . Thus we deduce that { 1 5 x } − { n x } − { m x } = 0 , − 1 ≤ 0 for all real x .
Solving the equation n − 1 + m − 1 = 1 5 − 1 for m ≥ n ≥ 1 6 leads to the equation ( m − 1 5 ) ( n − 1 5 ) = 2 2 5 , and so we can have n = 1 6 , 1 8 , 2 0 , 2 4 , 3 0 , with respective values of m being 2 4 0 , 9 0 , 6 0 , 4 0 , 3 0 . I claim there are no other values of n that will do.
Suppose now that m ≥ n ≥ 1 6 and that m , n satisfy the stated condition. Then (putting x = 1 ) we deduce that 1 5 − 1 ≤ n − 1 + m − 1 . Hence 2 2 5 ≥ ( m − 1 5 ) ( n − 1 5 ) ≥ ( n − 1 5 ) 2 so that 1 6 ≤ n ≤ 3 0 . We also note, if we put x = { m , n } to be the lowest common multiple of m and n , that 1 5 must divide { m , n } .
If m and n are coprime, find integers a , b such that n a − m b = 1 . Putting x = n a we deduce that { 1 5 a n } ≤ m 1 . Thus 1 5 must divide a n . Hence 1 5 divides a n 2 − b m n = n .
Consider the HCF ( m , n ) of m and n .
If ( 1 5 , n ) = 1 5 , we must have n = 3 0 .
If ( 1 5 , n ) = 5 , we must have n = 2 0 , 2 5 . If n = 2 5 , then m would have to be a multiple of 3 between 2 5 and 3 7 . 5 . But 1 5 does not divide n , so n and m cannot be coprime. Thus m must be a multiple of 5 as well, and hence must be 3 0 . But putting x = 1 2 5 shows that this cannot work. Thus we deduce that n = 2 0 is the only option.
If ( 1 5 , n ) = 3 , then n = 1 8 , 2 1 , 2 4 , 2 7 . If n = 2 7 then m would be a multiple of 5 between 1 6 and 3 3 . 7 5 . Since ( n , m ) = 1 , m would be a multiple of 3 as well, and so would have to be 3 0 . Putting x = − 2 7 shows that does not work. If n = 2 1 then m would be a multiple of 5 between 1 6 and 5 2 . 5 . Since ( n , m ) = 1 , m would have to be a multiple of either 3 or 7 as well, and hence must be one of 3 0 , 3 5 , 4 5 . But m = 3 0 does not work (put x = − 2 1 ), and m = 3 5 does not work ( x = 4 2 ), and m = 4 5 does not work ( x = − 8 4 ). Thus only n = 1 8 , 2 4 are possible.
If ( 1 5 , n ) = 1 , then m would have to be multiple of 1 5 between 1 6 and 2 4 0 . Thus m = 1 5 k where 2 ≤ k ≤ 1 6 , and 2 ≤ ( n , m ) = ( n , k ) ≤ 1 6 . Note that we cannot have k = 1 5 . Find integers a , b with a n + b m = ( n , m ) . If ( n , m ) < 1 6 , putting x = a n gives 1 5 ( n , m ) = { 1 5 a n } ≤ { m a n } = m ( n , m ) , which is impossible. Thus ( n , m ) = 1 6 , and hence n = 1 6 is the only option.
Thus the only possible values of n are 1 , 2 , … , 1 5 , 1 6 , 1 8 , 2 0 , 2 4 , 3 0 , and the sum of these numbers is 2 2 8 .