Maths Olympiad Prep

Library / /33 of 61

Number theory Difficulty 6.5 National Olympiad Prove it Canada

Problem:
Let S={1,4,8,9,16,}S=\{1,4,8,9,16, \ldots\} be the set of perfect powers of integers, i.e. numbers of the form nkn^{k} where n,kn, k are positive integers and k2k \geq 2. Write S={a1,a2,a3}S=\left\{a_{1}, a_{2}, a_{3} \ldots\right\} with terms in increasing order, so that a1<a2<a3a_{1}<a_{2}<a_{3} \cdots. Prove that there exist infinitely many integers mm such that 99999999 divides the difference am+1ama_{m+1}-a_{m}.

Solution

Solution:
The idea is that most perfect powers are squares. If an=x2a_{n}=x^{2} and an+1=(x+1)2a_{n+1}=(x+1)^{2}, then an+1an=2x+1a_{n+1}-a_{n}=2x+1. Note that 99992x+19999 \mid 2x+1 is equivalent to x4999(mod9999)x \equiv 4999 \pmod{9999}. Hence we will be done if we can show that there exist infinitely many x4999(mod9999)x \equiv 4999 \pmod{9999} such that there are no perfect powers strictly between x2x^{2} and (x+1)2(x+1)^{2}.

Assume otherwise, so that there exists a positive integer NN such that: for x4999(mod9999)x \equiv 4999 \pmod{9999} and xNx \geq N, there is a perfect power bxexb_{x}^{e_{x}} (ex2e_{x} \geq 2) between x2x^{2} and (x+1)2(x+1)^{2}. Without loss of generality, we can take NN to be 4999(mod9999)\equiv 4999 \pmod{9999}. Note that x2x^{2} and (x+1)2(x+1)^{2} are consecutive squares, hence exe_{x} is odd, and thus ex3e_{x} \geq 3. Let tnt_{n} be the number of odd perfect powers that are at most nn.

By tallying the bxexb_{x}^{e_{x}} up (clearly they are all distinct), for any m1m \geq 1 we have at least mm perfect odd powers between 11 and (N+9999m)2(N+9999m)^{2}, so that
t(N+9999m)2m t_{(N+9999m)^{2}} \geq m
In particular, for large enough nn we have
tnn10000 t_{n} \geq \frac{\sqrt{n}}{10000}
Now, if xfnx^{f} \leq n then xnfx \leq \sqrt[f]{n}. Also, nxf2fn \geq x^{f} \geq 2^{f} so flog2(n)f \leq \log_{2}(n). So we have
tni=3log2(n)nlog2(n)n3 t_{n} \leq \sum_{i=3}^{\log_{2}(n)} \sqrt*{n} \leq \log_{2}(n) \sqrt[3]{n}
Combining with the previous inequality, we have
n610000log2(n) \sqrt[6]{n} \leq 10000 \log_{2}(n)
for all large enough nn. However, this inequality is false for all large nn, contradiction. Therefore the problem statement holds.

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty) added by this project.