Maths Olympiad Prep

Library / /2 of 7

, 2020

Number theory Difficulty 8.4 Shortlist Prove it Taiwan

Let N>25000N > 2^{5000} be a positive integer. Prove that if 1a1<<ak<1001 \le a_1 < \dots < a_k < 100 are distinct positive integers then the number
i=1k(Nai+ai) \prod_{i=1}^{k} (N^{a_i} + a_i)
has at least kk distinct prime factors.

Solution

First, we prove the following lemma.

Lemma. For any positive integers X,AX, A and BB with ABA \neq B we have
gcd(XA+A,XB+B)<AB+BA. \text{gcd}(X^A + A, X^B + B) < A^B + B^A.

Proof of lemma. Note that XB+BX^B + B divides XAB(B)AX^{AB} - (-B)^A, and that
gcd(XA+A,(XA)B(B)A)=gcd(XA+A,(A)B(B)A). \text{gcd}(X^A + A, (X^A)^B - (-B)^A) = \text{gcd}(X^A + A, (-A)^B - (-B)^A).
We now have two cases. If ABBAA^B \neq B^A, then (A)B(B)A0(-A)^B - (-B)^A \neq 0, and
gcd(XA+A,XB+B)(A)B(B)AAB+BA. \text{gcd}(X^A + A, X^B + B) \le |(-A)^B - (-B)^A| \le A^B + B^A.
The only other case is when (A,B)=(2,4)(A, B) = (2, 4) or (A,B)=(4,2)(A, B) = (4, 2). But here we can check it manually, since gcd(X2+2,X4+4)=gcd(X2+2,8)8\text{gcd}(X^2 + 2, X^4 + 4) = \text{gcd}(X^2 + 2, 8) \le 8. ■

Now, we claim that for every 1m<1001 \le m < 100 there exists a prime power PmP_m dividing Nm+mN^m+m which does not divide any other Nc+cN^c+c (for 1c<1001 \le c < 100). This will immediately solve the problem, because no two of the PiP_i can have a common factor.

Assume for contradiction this is not the case. Then for some mm, it must be the case that
Nm+m1k<100kmgcd(Nm+k,Nk+m)1k<100km(mk+km). N^m + m \le \prod_{1 \le k < 100 \atop k \ne m} \text{gcd}(N^m + k, N^k + m) \le \prod_{1 \le k < 100 \atop k \ne m} (m^k + k^m).

Let α\alpha be a real number such that mαmm \le \alpha^m for all integers mm. We may then write
N<1k<100km(mk+km)1/m<1k<100((m1/m)k+k)<1k<100(αk+k)=α49501k<100(1+kαk)<α4950(1+1α+1α2++99α9999)99 \begin{align*} N &< \prod_{\substack{1 \le k < 100 \\ k \ne m}} (m^k + k^m)^{1/m} \\ &< \prod_{1 \le k < 100} ((m^{1/m})^k + k) \\ &< \prod_{1 \le k < 100} (\alpha^k + k) \\ &= \alpha^{4950} \prod_{1 \le k < 100} \left(1 + \frac{k}{\alpha^k}\right) \\ &< \alpha^{4950} \cdot \left(1 + \frac{\frac{1}{\alpha} + \frac{1}{\alpha^2} + \dots + \frac{99}{\alpha^{99}}}{99}\right)^{99} \end{align*}
with the last line being AM-GM. If we take even the crude bound α=2\alpha = 2 now, then we easily get an upper bound of 24950(1+299)100<250002^{4950} (1 + \frac{2}{99})^{100} < 2^{5000}, which is a contradiction. This concludes the proof.

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 translated into English from zh; metadata (topic, difficulty) added by this project.