Maths Olympiad Prep

Library / /479 of 520

Number theory Difficulty 7.2 National olympiad, round 2 Prove it

Theorem 1 Let m2,(a,m)=1m \geqslant 2, (a, m)=1, and gg be a primitive root modulo mm. Then the congruence equation (1) has a solution, i.e., aa is an nn-th power residue modulo mm if and only if
(n,φ(m))γ(a),(n, \varphi(m)) \mid \gamma(a),

where γ(a)=γm,g(a)\gamma(a)=\gamma_{m, g}(a) is the index of aa modulo mm with respect to the base gg. Moreover, when there is a solution, (1) has exactly (n,φ(m))(n, \varphi(m)) solutions.

Solution

Prove that if xx1(modm)x \equiv x_{1}(\bmod m) is a solution to (1), then by (a,m)=1(a, m)=1 we know (x1,m)=1\left(x_{1}, m\right)=1. Therefore, by Theorem 5 in §2, there must be a y1y_{1} such that
x1gy1(modm)x_{1} \equiv g^{y_{1}}(\bmod m)

Thus, we have
gny1a(modm)g^{n y_{1}} \equiv a(\bmod m)

Furthermore, by Property I in §3, we have
ny1γ(a)(modφ(m))n y_{1} \equiv \gamma(a)(\bmod \varphi(m))

This shows that yy1(modφ(m))y \equiv y_{1}(\bmod \varphi(m)) is a solution to the linear congruence
nyγ(a)(modφ(m))n y \equiv \gamma(a)(\bmod \varphi(m))

Conversely, if yy1(modφ(m))y \equiv y_{1}(\bmod \varphi(m)) is a solution to (5), then by Property I in §3, we can also deduce that equation (4) holds. Therefore, when x1x_{1} is given by equation (3), xx1(modm)x \equiv x_{1}(\bmod m) must be a solution to (1). This proves that the congruence equation (1) ((a,m)=1)((a, m)=1) and the congruence equation (5) have solutions simultaneously or not. Moreover, for any
x2gy2(modm)x_{2} \equiv g^{y_{2}}(\bmod m)

By Property IV in §1 (taking a=ga=g), we know that x1x2(modm)x_{1} \equiv x_{2}(\bmod m) if and only if
y1y2(modφ(m))y_{1} \equiv y_{2}(\bmod \varphi(m))

Therefore, when the congruence equation (1) ((a,m)=1)((a, m)=1) has solutions, it has the same number of solutions as the congruence equation (5). By Theorem 2 in Chapter 4, §2, the necessary and sufficient condition for (5) to have solutions is that equation (2) holds. When it has solutions, there are exactly (n,φ(m))(n, \varphi(m)) solutions. This, together with the previous discussion, proves the theorem.

Theorem 1 provides a method for solving equation (1) ((a,m)=1)((a, m)=1) when the modulus mm has a primitive root: (i) use the index table to find the index γ(a)\gamma(a) of aa; (ii) solve the congruence equation (5); (iii) if (5) has solutions, then for each solution y1(modφ(m))y_{1}(\bmod \varphi(m)), use the index table to find x1x_{1} satisfying equation (3). The resulting x1(modm)x_{1}(\bmod m) are all the solutions to (1).

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