Olympiad Maths Prep

Track / Stage 6 / 78 of 400 #1078 of 2000

Problem 1078

National olympiad, first round
Number theory Difficulty 6.1 Prove it

If the prime factorization of mm is

m=p1α1p2α3prαr m=p_{1}^{\alpha_{1}} p_{2}^{\alpha_{3}} \ldots p_{r}^{\alpha_{r}}

then let us introduce the following notation:

S(m)=(1)a1+a2++ar S(m)=(-1)^{a_{1}+a_{2}+\ldots+a_{r}}

noting that S(1)=1S(1)=1. Prove that the sum extended over all divisors dd of some number nn

S(d)={10 \sum S(d)=\left\{\begin{array}{l} 1 \\ 0 \end{array}\right.

according to whether nn is a perfect square or not. (The unit and nn itself are also counted among the divisors).

This one wants a proof. Work it on paper, read the official solution, then mark yourself honestly — the ladder only means something if the record is true.

Official solution

If we carry out the indicated product in the following expression:

{(1p1+p12p13p13++(1)α1p1α1)×(1p2+p22p23p23++(1)α2p2α2)×(1pr+pr2pr3pr3++(1)αrprαr) \left\{\begin{array}{l} \left(1-p_{1}+p_{1}^{2}-p_{1}^{3}-p_{1}^{3}+\ldots+(-1)^{\alpha_{1}} p_{1}^{\alpha_{1}}\right) \times \\ \left(1-p_{2}+p_{2}^{2}-p_{2}^{3}-p_{2}^{3}+\ldots+(-1)^{\alpha_{2}} p_{2}^{\alpha_{2}}\right) \times \\ \cdots \\ \left(1-p_{r}+p_{r}^{2}-p_{r}^{3}-p_{r}^{3}+\ldots+(-1)^{\alpha_{r}} p_{r}^{\alpha_{r}}\right) \end{array}\right.

we obtain an algebraic sum whose terms will be products formed from the prime divisors (p1,p2,,p3)\left(p_{1}, p_{2}, \ldots, p_{3}\right), where p1p_{1} will appear in the 0,1,2,0,1,2, \ldots or α1\alpha_{1}-th power; p2p_{2} in the 0,1,2,0,1,2, \ldots or α2\alpha_{2}-th power; ... and finally prp_{r} in the 0,1,2,0,1,2, \ldots or αr\alpha_{r}-th power. Each term will thus be a divisor of mm, and since all such products that can be formed under the mentioned conditions are present in the expansion of (1), each divisor of mm will appear exactly once among the terms; the divisors will now appear in this sum with a positive sign for which α1+α2+\alpha_{1}+\alpha_{2}+\ldots is even, so S(d)=1S(d)=1, while with a negative sign those terms will appear in the expansion for which S(d)=1S(d)=-1. Thus, to obtain the desired S(d)\sum S(d), it is sufficient to substitute 1 for all pp's in (1). In this case, in the expansion of (1), 1 or -1 will appear in place of each term, depending on whether S(d)=1S(d)=1 or S(d)=1S(d)=-1. Thus, indeed, the expansion of (1) gives the value of S(d)\sum S(d) when 1 is substituted for all pp's.

The number m=p1α1p2α2prαrm=p_{1}^{\alpha_{1}} p_{2}^{\alpha_{2}} \ldots p_{r}^{\alpha_{r}} is a perfect square if and only if all α\alpha's are even; in this case, the expressions in each row of (1) will be equal to 1, and thus their product: S(d)=1\sum S(d)=1. However, if even one α\alpha, for example, αr\alpha_{r}, is odd, i.e., mm is not a perfect square, then the expression in the rr-th row of (1) will be 0, and thus the entire expression S(d)=0\sum S(d)=0.

(Dénes König, Budapest.)

The problem was also solved by: Bartók I., Dömény I., Haar A., Szücs A.

Source: NuminaMath-1.5, licensed Apache-2.0. Statement and solution reproduced as published; topic, difficulty and ordering added by this site.