Browse · MathNet
PrintIMO 2019 Shortlisted Problems
2019 number theory
Problem
Let , and let be a positive integer. Prove that there exists a constant such that, if satisfies , then there exist such that . (Here is the set of positive integers, and denotes the greatest integer less than or equal to .)
Solution
First, observe that if is a positive integer, then exactly when To see why, observe that if and only if for some . In other words, , which is equivalent to (1).
Now, write , where . Observe that the set of differences is not altered by shifting , so we may assume that with .
From (1), we learn that for each since . Furthermore, we must have whenever ; otherwise, we would have Since , this implies that , contradicting (1).
Now, we have a sequence , with We use the following fact: for any , we have To see why this is the case, let , so . Then since the numerator is a positive integer. Because , inequality (2) follows.
Let , for . Then , and we have Here, the first inequality holds because , the second follows from (2), the third follows from an easy application of the AM-HM inequality (or Cauchy-Schwarz), and the fourth follows from the fact that .
Rearranging this, we obtain which provides the required bound on .
---
Alternative solution.
Let , so . Thus, is the complementary Beatty sequence to (in other words, and are disjoint with ). Write . Suppose that has no differences in , so all its differences are in and we can set for .
For any , we have . Because , we also have for some positive integer . Thus, . The right hand side must equal either or , the latter of which is not a member of as . Therefore, and so we have .
For we now put , and we have that is, . We also have so .
With the above inequalities, an argument similar to (3) (which uses the fact that for positive integers ) proves that , which again rearranges to give
---
Alternative solution.
Again, define , so all differences between elements of are in . We start by making the following observation. Suppose we have a set such that all of the differences between elements of are in . Then .
To see why, observe that any two sums of the form with are different; otherwise, we would have , and so . However, then the left hand side is in whereas the right hand side is in . Thus, is a set of size all of whose elements are no greater than , yielding the claimed inequality.
With this in mind, it suffices to construct a set , all of whose differences are in and whose size is at least for some constant .
To do so, we will use well-known facts about the negative Pell equation ; in particular, that there are infinitely many solutions and the values of are given by the recurrence and . Therefore, we may choose to be a solution with .
Now, we claim that we may choose . Indeed, we have and so from which it follows that . Combined with (1), this shows that all differences between elements of are in .
---
Alternative solution.
As in Solution 3, we will provide a construction of a large set , all of whose differences are in .
Choose to be a solution to the Pell-like equation ; such solutions are given by the recurrence and , and so we can choose such that . Furthermore, it is known that for such a and for , if , and if . Indeed, this is a statement of the fact that is a best rational approximation to , from below in the first case and from above in the second.
Now, consider the sequence . The Erdős-Szekeres theorem tells us that this sequence has a monotone subsequence with at least elements; if that subsequence is decreasing, we may reflect (using (4) or (5)) to ensure that it is increasing. Call the subsequence for .
Now, set . We have for (because the corresponding inequality for the fractional parts holds by the ordering assumption on the ), which means that all differences between elements of are indeed in . Since , this is the required set.
Now, write , where . Observe that the set of differences is not altered by shifting , so we may assume that with .
From (1), we learn that for each since . Furthermore, we must have whenever ; otherwise, we would have Since , this implies that , contradicting (1).
Now, we have a sequence , with We use the following fact: for any , we have To see why this is the case, let , so . Then since the numerator is a positive integer. Because , inequality (2) follows.
Let , for . Then , and we have Here, the first inequality holds because , the second follows from (2), the third follows from an easy application of the AM-HM inequality (or Cauchy-Schwarz), and the fourth follows from the fact that .
Rearranging this, we obtain which provides the required bound on .
---
Alternative solution.
Let , so . Thus, is the complementary Beatty sequence to (in other words, and are disjoint with ). Write . Suppose that has no differences in , so all its differences are in and we can set for .
For any , we have . Because , we also have for some positive integer . Thus, . The right hand side must equal either or , the latter of which is not a member of as . Therefore, and so we have .
For we now put , and we have that is, . We also have so .
With the above inequalities, an argument similar to (3) (which uses the fact that for positive integers ) proves that , which again rearranges to give
---
Alternative solution.
Again, define , so all differences between elements of are in . We start by making the following observation. Suppose we have a set such that all of the differences between elements of are in . Then .
To see why, observe that any two sums of the form with are different; otherwise, we would have , and so . However, then the left hand side is in whereas the right hand side is in . Thus, is a set of size all of whose elements are no greater than , yielding the claimed inequality.
With this in mind, it suffices to construct a set , all of whose differences are in and whose size is at least for some constant .
To do so, we will use well-known facts about the negative Pell equation ; in particular, that there are infinitely many solutions and the values of are given by the recurrence and . Therefore, we may choose to be a solution with .
Now, we claim that we may choose . Indeed, we have and so from which it follows that . Combined with (1), this shows that all differences between elements of are in .
---
Alternative solution.
As in Solution 3, we will provide a construction of a large set , all of whose differences are in .
Choose to be a solution to the Pell-like equation ; such solutions are given by the recurrence and , and so we can choose such that . Furthermore, it is known that for such a and for , if , and if . Indeed, this is a statement of the fact that is a best rational approximation to , from below in the first case and from above in the second.
Now, consider the sequence . The Erdős-Szekeres theorem tells us that this sequence has a monotone subsequence with at least elements; if that subsequence is decreasing, we may reflect (using (4) or (5)) to ensure that it is increasing. Call the subsequence for .
Now, set . We have for (because the corresponding inequality for the fractional parts holds by the ordering assumption on the ), which means that all differences between elements of are indeed in . Since , this is the required set.
Techniques
Pell's equationsCounting two waysCauchy-SchwarzQM-AM-GM-HM / Power MeanFloors and ceilings