Maths Olympiad Prep

Library / /161 of 169

Number theory Difficulty 8.0 Shortlist Prove it United States

For integral mm, let p(m)p(m) be the greatest prime divisor of mm. By convention, we set p(±1)=1p(\pm 1) = 1 and p(0)=p(0) = \infty. Find all polynomials ff with integer coefficients such that the sequence {p(f(n2))2n}n0\{p(f(n^2)) - 2n\}_{n \ge 0} is bounded above. (In particular, this requires f(n2)0f(n^2) \ne 0 for n0n \ge 0.)

Solution

The polynomial ff has the required properties if and only if
f(x)=c(4xa12)(4xa22)(4xak2),() f(x) = c(4x - a_1^2)(4x - a_2^2)\cdots(4x - a_k^2), \quad (*)
where a1,a2,,aka_1, a_2, \dots, a_k are odd positive integers and cc is a nonzero integer. It is straightforward to verify that polynomials given by ()(*) have the required property. If pp is a prime divisor of f(n2)f(n^2) but not of cc, then p(2naj)p|(2n-a_j) or p(2n+aj)p|(2n+a_j) for some jkj \le k. Hence p2nmax{a1,a2,,ak}p-2n \le \max\{a_1, a_2, \dots, a_k\}. The prime divisors of cc form a finite set and do affect whether or not the given sequence is bounded above. The rest of the proof is devoted to showing that any ff for which {p(f(n2))2n}n0\{p(f(n^2)) - 2n\}_{n \ge 0} is bounded above is given by ()(*).

Let Z[x]\mathbb{Z}[x] denote the set of all polynomials with integral coefficients. Given fZ[x]f \in \mathbb{Z}[x], let P(f)\mathcal{P}(f) denote the set of those primes that divide at least one of the numbers in the sequence {f(n)}n0\{f(n)\}_{n \ge 0}. The solution is based on the following lemma.

Lemma If fZ[x]f \in \mathbb{Z}[x] is a nonconstant polynomial then P(f)\mathcal{P}(f) is infinite.

Proof: Repeated use will be made of the following basic fact: if aa and bb are distinct integers and fZ[x]f \in \mathbb{Z}[x], then aba-b divides f(a)f(b)f(a)-f(b). If f(0)=0f(0)=0, then pp divides f(p)f(p) for every prime pp, so P(f)\mathcal{P}(f) is infinite. If f(0)=1f(0)=1, then every prime divisor pp of f(n!)f(n!) satisfies p>np>n. Otherwise pp divides n!n!, which in turn divides f(n!)f(0)=f(n!)1f(n!)-f(0)=f(n!)-1. This yields p1p|1, which is false. Hence f(0)=1f(0)=1 implies that P(f)\mathcal{P}(f) is infinite. To complete the proof, set g(x)=f(f(0)x)/f(0)g(x) = f(f(0)x)/f(0) and observe that gZ[x]g \in \mathbb{Z}[x] and g(0)=1g(0)=1. The preceding argument shows that P(g)\mathcal{P}(g) is infinite, and it follows that P(f)\mathcal{P}(f) is infinite. \blacksquare

Suppose fZ[x]f \in \mathbb{Z}[x] is nonconstant and there exists a number MM such that p(f(n2))2nMp(f(n^2)) - 2n \le M for all n0n \ge 0. Application of the lemma to f(x2)f(x^2) shows that there is an infinite sequence of distinct primes {pj}\{p_j\} and a corresponding infinite sequence of nonnegative integers {kj}\{k_j\} such that pjf(kj2)p_j|f(k_j^2) for all j1j \ge 1. Consider the sequence {rj}\{r_j\} where rj=min{kj(modpj),pjkj(modpj)}r_j = \min\{k_j \pmod{p_j}, p_j - k_j \pmod{p_j}\}. Then 0rj(pj1)/20 \le r_j \le (p_j-1)/2 and pjf(rj2)p_j|f(r_j^2). Hence 2rj+1pjp(f(rj2))M+2rj2r_j+1 \le p_j \le p(f(r_j^2)) \le M+2r_j, so 1pj2rjM1 \le p_j - 2r_j \le M for all j1j \ge 1. It follows that there is an integer a1a_1 such that 1a1M1 \le a_1 \le M and a1=pj2rja_1 = p_j - 2r_j for infinitely many jj. Let m=degfm = \deg f. Then pj4mf((pja1)/2)2)p_j|4^m f((p_j-a_1)/2)^2) and 4mf((xa1)/2)2)Z[x]4^m f((x-a_1)/2)^2) \in \mathbb{Z}[x]. Consequently, pjf((a1/2)2)p_j|f((a_1/2)^2) for infinitely many jj, which shows that (a1/2)2(a_1/2)^2 is a zero of ff. Since f(n2)0f(n^2) \ne 0 for n0n \ge 0, a1a_1 must be odd. Then f(x)=(4xa12)g(x)f(x) = (4x-a_1^2)g(x) where gZ[x]g \in \mathbb{Z}[x]. (See the note below.) Observe that {p(g(n2))2n}n0\{p(g(n^2)) - 2n\}_{n \ge 0} must be bounded above. If gg is constant, we are done. If gg is nonconstant, the argument can be repeated to show that ff is given by ()(*).

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.