Maths Olympiad Prep

Library / /32 of 42

Algebra Difficulty 6.7 National olympiad Prove it Romania

Let kNk \in \mathbb{N}^*. We say that the ring (A,+,)(A, +, \cdot) has the property CP(k)CP(k), if for every a,bAa, b \in A there is a cAc \in A, such that ak=bk+cka^k = b^k + c^k.
a) Give an example of a finite ring (A,+,)(A, +, \cdot), which does not have the property CP(k)CP(k) for any positive integer kk, with k2k \ge 2.
b) Let nNn \in \mathbb{N}, n3n \ge 3, and M(n)={mN(Zn,+,) has the property CP(m)}M(n) = \{m \in \mathbb{N}^* \mid (\mathbb{Z}_n, +, \cdot) \text{ has the property } CP(m)\}. Prove that M(n)M(n) is a monoid with respect to multiplication, included in the set 2N+12 \cdot \mathbb{N} + 1 of odd positive integers.

Solution

For any kNk \in \mathbb{N}^* we denote Pk(A)={akaA}P_k(A) = \{a^k \mid a \in A\}. The condition CP(k)CP(k) is then equivalent with
xyPk(A),for any x,yPk(A), x - y \in P_k(A), \quad \text{for any } x, y \in P_k(A),
meaning that Pk(A)P_k(A) is a subgroup of the additive group (A,+)(A, +).

a) For A=Z4A = \mathbb{Z}_4, we have that P2k(A)={0^,1^}P_{2k}(A) = \{\hat{0}, \hat{1}\}, respectively P2k+1(A)={0^,1^,3^}P_{2k+1}(A) = \{\hat{0}, \hat{1}, \hat{3}\}, for any kNk \in \mathbb{N}^*. These are not subgroups of the group (Z4,+)(\mathbb{Z}_4, +). Hence, the ring (Z4,+,)(\mathbb{Z}_4, +, \cdot) does not have the property CP(k)CP(k) for any kNk \in \mathbb{N}, with k2k \ge 2.

b) For nNn \in \mathbb{N}, n3n \ge 3, we consider the ring (Zn,+,)(\mathbb{Z}_n, +, \cdot). Because 1^Pk(Zn)\hat{1} \in P_k(\mathbb{Z}_n) for any kNk \in \mathbb{N}, k2k \ge 2, and (Zn,+)(\mathbb{Z}_n, +) is cyclic, generated by 1^\hat{1}, it follows that CP(k)    Pk(Zn)=ZnCP(k) \iff P_k(\mathbb{Z}_n) = \mathbb{Z}_n. Equivalently, CP(k)    CP(k) \iff the function pk:ZnZnp_k : \mathbb{Z}_n \to \mathbb{Z}_n, defined by pk(x)=xkp_k(x) = x^k for any xZnx \in \mathbb{Z}_n, is bijective.
Then we can rewrite M(n)={mNpmM(n) = \{m \in \mathbb{N}^* \mid p_m is bijective\}.
Because for even kk we have that pk(1^)=pk(1^)p_k(\hat{1}) = p_k(-\hat{1}), and 1^1^\hat{1} \ne -\hat{1}, it follows that any mM(n)m \in M(n) is odd, so that M(n)2N+1M(n) \subseteq 2 \cdot \mathbb{N} + 1.
Because p1=idZnp_1 = \text{id}_{\mathbb{Z}_n} is bijective, we have that 1M(n)1 \in M(n).
Let m1,m2M(n)m_1, m_2 \in M(n) be arbitrary. Since the functions pm1p_{m_1} and pm2p_{m_2} are bijective, the function pm1m2=pm1pm2p_{m_1 m_2} = p_{m_1} \circ p_{m_2} is also bijective, as the composition of two bijective functions. We deduce that m1m2M(n)m_1 \cdot m_2 \in M(n).
It follows that M(n)M(n) is a submonoid of the monoid (N,)(\mathbb{N}^*, \cdot).

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 and solution reproduced as published; topic and difficulty added by this site.