26th IMO 1985 shortlist

------
 
 
Problem 19

Show that the sequence [n √2] for n = 1, 2, 3, ... contains infinitely many powers of 2.

 

Solution

Solution by Demetres Christofides

Suppose not, so that the highest power of 2 in the sequence is 2m-1. Since √2 < 2, the difference between [n√2] and [(n+1)√2] is at most 2. Since √2 > 1, the difference is at least 1. So it is either 1 or 2. So for some N we must have [N√2] = 2m - 1, [(N+1)√2] = 2m + 1.

We claim that for k = 0, 1, 2, ... we have [2kN√2] = 2m+k - 1, [(2kN+1)√2] = 2m+k + 1. We use induction on k. It is true for k = 0. So suppose it is true for k. Then 2kN√2 < 2m+k, so 2k+1N√2 < 2m+k+1. Also (2kN+1)√2 > 2m+k + 1, so 2kN√2 > 2m+k - (√2 - 1) and hence 2k+1N√2 > 2m+k+1 - 2(√2 - 1) > 2m+k+1 - 1. Hence [2k+1N√2] = 2m+k+1 - 1. So [(2k+1N+1)√2] = 2m+k+1 or 2m+k+1 + 1. By assumption it is not 2m+k+1, so it must be 2m+k+1 + 1. That establishes the result for k+1 and hence for all k.

But √2 is irrational, so N√2 = 2m - h for some positive real h. Take k such that h > 1/2k, then 2kN√2 = 2m+k - h2k < 2m+k - 1, so [2kN√2] < 2m+k - 1. Contradiction. So the assumption must be wrong and there must be infinitiely many powers of 2.

 


 

26th IMO shortlist 1985

© John Scholes
jscholes@kalva.demon.co.uk
11 Sep 2002