betaFeedback

WorksMathematics

On complemented skew braces

Abstract

We investigate complemented skew left braces, that is, skew braces in which every skew subbrace admits a complement. We establish that if every minimal subbrace of a finite skew brace is complemented, then the brace is supersoluble and all its skew subbraces of prime power order are elementary abelian. We also show that the converse does not hold in general, providing a finite counterexample. Furthermore, we prove that every subbrace of a finite direct product of square-free skew braces is complemented. Finally, we extend our investigation to the infinite setting, proving that every multipermutational complemented skew brace is hypercyclic and that all its skew subbraces whose additive groups are pp-groups are elementary abelian.

Introduction

The study of complemented subgroups plays an important role in the structural theory of finite groups. Recall that a subgroup HH of a group GG is said to be complemented in GG if there exists a subgroup KK of GG such that G=HKG = HK and H∩K=1H \cap K = 1. Groups in which every subgroup is complemented are called complemented groups and form a particularly interesting class, since the existence of complements imposes strong restrictions on the subgroup structure. and provides a natural framework for understanding group decompositions and factorisations.

One of the classical results in this direction is due to Hall [8] and the fourth author and Guo [3], establishing the following characterisation of finite complemented groups.

Theorem 1.1. Let GG be a finite group. Then the following statements are equivalent:

  1. GG is complemented.

  1. Every subgroup of prime order of GG is complemented in GG.

  1. GG is supersoluble and its Sylow subgroups are elementary abelian.

  1. GG is isomorphic to a subgroup of a direct product of square-free finite groups.

The aim of this paper is to investigate the extent to which this characterisation can be generalised to the setting of skew left braces.

A skew (left) brace is a set BB endowed with two group operations (B,+)(B,+), the additive group, and (B,⋅)(B,\cdot), the multiplicative group, that satisfy the following distributivity-like property:

a(b+c)=ab−a+ac for all a,b,c∈B.a(b+c)=ab-a+ac\text{ for all }a,b,c\in B.

Skew braces have become an important algebraic tool in the study of set-theoretic solutions of the Yang–Baxter equation. The algebraic structure of a skew brace determines a solution, while properties such as ideals, skew subbraces, quotients and decompositions play an important role in understanding the associated solutions. Consequently, identifying classes of skew braces with particularly strong decomposition properties may provide new approaches to constructing and studying solutions of the Yang–Baxter equation.

One of the key facts about skew braces is that the multiplicative group acts via automorphisms on the additive group by means of the lambda map

λ:(B,⋅)⟶Aut⁡(B,+),λa(b)=−a+ab for all a,b∈B.\lambda:(B,\cdot)\longrightarrow\operatorname{Aut}(B,+),\qquad\lambda_a(b)=-a+ab\text{ for all }a,b\in B.

In case this action is trivial, we say that BB is a trivial skew brace. In this case a+b=aba+b=ab for all a,b∈Ba,b\in B. Furthermore, if BB is trivial and (B,+)(B,+) is abelian (respectively elementary abelian) we say that BB is abelian (respectively elementary abelian).

A subset AA of BB is a skew subbrace of BB if AA is subgroup of (B,+)(B,+) and (B,⋅)(B,\cdot) simultaneously. A skew subbrace of a brace BB is called simply subbrace throughout.

This leads us to consider skew braces in which every subbrace admits a complement that is itself a subbrace. According to [2 Definition 8], a subbrace AA of BB is complemented in BB if there exists a subbrace CC such that

B=A+C=AC,A∩C=0.B = A + C = AC,\qquad A \cap C = 0.

In this case, CC is called a complement of AA in BB.

We shall refer to skew braces where every subbrace is complemented as complemented skew braces.

Recall that a minimal subbrace of a brace BB is a subbrace of prime order.

Our first main result establishes a structural consequence of the complementability of minimal subbraces.

Theorem A. Let BB be a finite skew brace in which every minimal subbrace is complemented in BB. Then BB is supersoluble and every subbrace of prime power order is elementary abelian.

In particular, this result provides an analogue of the implication from condition 2 to condition 3 in the classical characterisation of complemented groups. Moreover, the complementability of a finite skew brace imposes strong restrictions on both of its underlying group structures.

Corollary 1.2. Let BB be a finite complemented skew brace. Then (B,+)(B,+) and (B,⋅)(B,\cdot) are complemented groups.

However, the converse of Theorem A does not hold in general. More precisely, we show that the conditions of supersolubility and having elementary abelian pp-subbraces are not sufficient to guarantee that every minimal subbrace of a finite skew brace is complemented.

Theorem B. There exists a finite supersoluble skew brace BB in which every pp-subbrace is elementary abelian, but BB has minimal subbraces that are not complemented.

Thus, unlike the group case, the structural conditions appearing in the classical characterisation do not provide a complete characterisation of complemented skew braces. This illustrates the additional difficulties arising from the compatibility between the additive and multiplicative operations.

We also investigate the role of square-free skew braces, which provide a natural analogue of the groups appearing in condition 4 of the classical result.

Theorem C. Every subbrace of a finite direct product of square-free skew braces is complemented.

Finally, we consider infinite complemented braces. Although the finite characterisation of complemented groups relies on finiteness in an essential way, an infinite analogous structural result can be obtained for multipermutational skew braces.

Theorem D. Let BB be a multipermutational complemented skew brace. Then BB is hypercyclic and every subbrace whose additive group is a pp-group is elementary abelian.

Preliminaries

Let BB be skew brace. If XX is a subset of BB we denote by ⟨X⟩\langle X\rangle (respectively ⟨X⟩+\langle X\rangle_{+} and ⟨X⟩∙\langle X\rangle_{\bullet}) the smallest subbrace (additive subgroup and multiplicative subgroup respectively) of BB that contains XX.

The star product of two elements of BB is defined by

a∗b=−a+ab−b=λa(b)−ba * b = -a + ab - b = \lambda_{a}(b) - b

for all a,b∈Ba,b \in B.

The star product of two subsets X,Y⊆BX,Y \subseteq B is defined by

X∗Y=⟨x∗y∣x∈X, y∈Y⟩+.X * Y = \langle x * y \mid x \in X,\, y \in Y\rangle_{+}.

A subbrace AA of BB is a left ideal of BB if it is λ\lambda-invariant, or equivalently, B∗A⊆AB * A \subseteq A. If AA is also normal in the additive and multiplicative group we say that AA is an ideal of BB. The relevance of ideals lies in the fact that we can make quotients with them. Two important subbraces of BB that play a key role here are the kernel of lambda and the fixed points set.

Ker⁡λ={a∈B∣λa(b)−b=a∗b=0 for all b∈B},\operatorname{Ker}\lambda= \{a \in B \mid\lambda_{a}(b) - b = a * b = 0 \text{ for all } b \in B\},
Fix⁡(B)={b∈B∣λa(b)−b=a∗b=0 for all a∈B}.\operatorname{Fix}(B) = \{b \in B \mid\lambda_{a}(b) - b = a * b = 0 \text{ for all } a \in B\}.

The former is a subbrace while the latter is a left ideal of BB. An important ideal of BB is the socle of BB

Soc⁡(B)=Ker⁡λ∩Z(B,+).\operatorname{Soc}(B) = \operatorname{Ker}\lambda\cap Z(B,+).

We can construct a socle series as follows

Soc⁡1(B)=Soc⁡(B),Soc⁡n+1(B)/Soc⁡n(B)=Soc⁡(B/Soc⁡n(B)),n≥1.\operatorname{Soc}_{1}(B) = \operatorname{Soc}(B), \quad\operatorname{Soc}_{n+1}(B)/\operatorname{Soc}_{n}(B) = \operatorname{Soc}(B/\operatorname{Soc}_{n}(B)), \quad n \geq1.

We will say that BB is multipermutational if there exists n≥1n \ge1 such that Soc⁡n(B)=B\operatorname{Soc}_{n}(B) = B. Moreover, the smallest n≥1n \ge1 for which this happens is called the multipermutational level of BB.

According to [2 Definition 18], a skew brace BB is said to be soluble if there exists an ideal series

0=I0⊆I1⊆⋯⊆In−1⊆In=B0 = I_0 \subseteq I_1 \subseteq\cdots\subseteq I_{n-1} \subseteq I_n = B

of BB such that Ii+1/IiI_{i+1}/I_i is abelian.

We will say that an ideal II of a skew brace BB is BB-hypercyclic if there exists an ascending series of ideals of BB

0=I0⊆I1⊆⋯⊆Iα⊆Iα+1⊆⋯⊆Iγ=I0 = I_0 \subseteq I_1 \subseteq\cdots\subseteq I_\alpha\subseteq I_{\alpha+1} \subseteq\cdots\subseteq I_\gamma= I

(here α\alpha and γ\gamma are ordinal numbers) such that Iα+1/IαI_{\alpha+1}/I_\alpha has prime order. If the series is finite, we will say that II is BB-supersoluble (see [1 Definition 3.1]). If B=IB = I, we will simply say that BB is hypercyclic.

Let pp be a prime. A pp-subbrace of BB is a subbrace AA of BB such that (A,+)(A,+) and (A,⋅)(A,\cdot) are pp-groups, that is, groups in which every element has order a power of pp. In the finite case that means that ∣A∣|A| is a power of pp.

If BB is a finite skew brace and ∣B∣=ptm|B| = p^t m, where pp a prime and (p,m)=1(p,m) = 1, BB has subbraces of order ptp^t (see [5]). Such pp-subbraces are called Sylow pp-subbraces.

Proof of Theorem A

The following result is elementary, therefore, we omit its proof.

Proposition 3.1. Let BB be a complemented skew brace, AA a subbrace of BB and II an ideal of BB. Then AA and B/IB/I are complemented skew braces.

Proposition 3.2. Let BB be a finite pp-skew brace. If every minimal subbrace of BB is complemented in BB, then BB is elementary abelian.

Proof. Suppose that the result is false, and let BB be a minimal counterexample. By Proposition 3.1 and the minimality of BB, all subbraces are elementary abelian. We can also suppose that BB is a non-trivial brace, because in other case, the result is true by Theorem 1.1.

By [7 Theorem 4.8 and Corollary 2.27], Fix⁡(B)≠0\operatorname{Fix}(B) \ne0. Since Fix⁡(B)\operatorname{Fix}(B) is a trivial brace, every subgroup of Fix⁡(B)\operatorname{Fix}(B) is a subbrace of BB. Hence, we can consider a subbrace FF of BB of order pp contained in Fix⁡(B)\operatorname{Fix}(B). Then there exists a proper subbrace CC of BB such that B=F+C=FCB = F + C = FC and F∩C=0F \cap C = 0. Let c0∈Cc_0 \in C and b∈Bb \in B. Then there exists f∈Ff \in F and c∈Cc \in C such that b=f+cb = f + c. Since f∈Fix⁡(B)f \in\operatorname{Fix}(B) and CC is elementary abelian, we have that

λc0(b)=λc0(f+c)=λc0(f)+λc0(c)=f+c=b.\lambda_{c_0}(b) = \lambda_{c_0}(f + c) = \lambda_{c_0}(f) + \lambda_{c_0}(c) = f + c = b.

Hence, c0∈Ker⁡λc_0 \in\operatorname{Ker} \lambda, and C⊆Ker⁡λC \subseteq\operatorname{Ker} \lambda. Therefore ∣B:Ker⁡λ∣≤p|B : \operatorname{Ker} \lambda| \leq p. Since BB is not trivial, we have that ∣B:Ker⁡λ∣=p|B : \operatorname{Ker} \lambda| = p and C=Ker⁡λC = \operatorname{Ker} \lambda. Furthermore, this forces Ker⁡λ∩Fix⁡(B)=0\operatorname{Ker} \lambda\cap\operatorname{Fix}(B) = 0, because otherwise we could suppose that F⊆Ker⁡λ∩Fix⁡(B)F \subseteq\operatorname{Ker} \lambda\cap\operatorname{Fix}(B) which is a contradiction. Therefore, Fix⁡(B)\operatorname{Fix}(B) is a complement of Ker⁡λ\operatorname{Ker} \lambda in BB and ∣Fix⁡(B)∣=p|\operatorname{Fix}(B)| = p.

On the other hand, Ker⁡λ≠0\operatorname{Ker} \lambda\neq0, since in other case, B=Fix⁡(B)B = \operatorname{Fix}(B) and BB would be trivial. Since Ker⁡λ\operatorname{Ker} \lambda is a trivial brace, Ker⁡λ\operatorname{Ker} \lambda contains a subbrace KK of order pp. Let DD be a complement of KK in BB. Let d0∈Dd_0 \in D and b∈Bb \in B. Then there exists k∈Kk \in K and d∈Dd \in D such that b=kdb = kd. Since k∈Ker⁡λk \in\operatorname{Ker} \lambda and DD is elementary abelian, we have that

λb(d0)=λkd(d0)=λk(λd(d0))=λk(d0)=d0.\lambda_b(d_0) = \lambda_{kd}(d_0) = \lambda_k(\lambda_d(d_0)) = \lambda_k(d_0) = d_0.

Therefore, d0∈Fix⁡(B)d_0 \in\operatorname{Fix}(B) and so D⊆Fix⁡(B)D \subseteq\operatorname{Fix}(B). Analogously to the previous case, KK is complemented by Fix⁡(B)\operatorname{Fix}(B) in BB. Hence K=Ker⁡λK = \operatorname{Ker} \lambda and ∣Ker⁡λ∣=p|\operatorname{Ker} \lambda| = p.

Finally, ∣B∣=∣Ker⁡λ∣∣Fix⁡(B)∣=p2|B| = |\operatorname{Ker} \lambda||\operatorname{Fix}(B)| = p^2. In this case (B,+)(B,+) and (B,⋅)(B,\cdot) are abelian groups, therefore, Ker⁡λ⊆Fix⁡(B)\operatorname{Ker} \lambda\subseteq\operatorname{Fix}(B), a contradiction. □

Proposition 3.3. Let BB be a finite skew brace with every minimal subbrace complemented in BB, and let pp be the largest prime dividing ∣B∣|B|. Then BB has a Sylow pp-subbrace that is an ideal of BB.

Proof. We argue by induction on ∣B∣|B|. If BB is a pp-skew brace, then BB is its Sylow pp-subbrace and the result holds. Suppose that ∣B∣|B| is divisible by at least two primes. By Cauchy’s Theorem for braces (see [5]), BB has a subbrace QQ of order qq, where qq is the smallest prime dividing ∣B∣|B|. Let CC be a complement of QQ in BB.

By induction, CC has a Sylow pp-subbrace PP that is an ideal of CC. In particular, (P,+)(P,+) and (P,⋅)(P,\cdot) are characteristic in (C,+)(C,+) and (C,⋅)(C,\cdot) respectively. Since ∣B:C∣=q|B : C| = q and qq is the smallest prime dividing ∣B∣|B|, we have that (C,+)(C,+) and (C,⋅)(C,\cdot) are normal in (B,+)(B,+) and (B,⋅)(B,\cdot) respectively. Hence, (P,+)(P,+) and (P,⋅)(P,\cdot) are normal in (B,+)(B,+) and (B,⋅)(B,\cdot) respectively. Since q≠pq \neq p, we have that PP is a Sylow subbrace of BB. Since PP is characteristic in (B,+)(B,+) and normal in (B,⋅)(B,\cdot), it follows that PP is an ideal of BB. □

Proof of Theorem A. We argue by induction on ∣B∣|B|. Let pp be the largest prime dividing ∣B∣|B|, by Proposition 3.3, BB has a Sylow pp-subbrace PP that is an ideal of BB. By Proposition 3.2, PP is elementary abelian as PP is a complemented skew brace by Proposition 3.1. Let T/PT/P be a minimal subbrace of B/PB/P. Then ∣T/P∣=q|T/P| = q for some q≠pq \ne p. By the Schur-Zassenhaus theorem for skew braces (see [4]), TT has a minimal subbrace SS such that T=S+P=SPT = S + P = SP and S∩P=0S \cap P = 0. Let UU be a complement of SS in BB. Then B/P=T/P+(U+P)/P=(T/P)((U+P)/P)B/P = T/P + (U + P)/P = (T/P)((U + P)/P) and T∩(U+P)=(T∩U)+P=PT \cap(U + P) = (T \cap U) + P = P (note that T∩UT \cap U is contained in PP). Therefore, every minimal subbrace of B/PB/P is complemented in B/PB/P. By induction B/PB/P is supersoluble, and PP is elementary abelian. Therefore, BB is soluble.

Let NN be a minimal ideal of BB. By [2], NN is abelian. Consider AA a subbrace of NN of order pp, and let CC be a complement of AA in BB. Since NN is an ideal of BB, N∩CN \cap C is an ideal of CC. Moreover, NN is abelian; therefore, AA normalises N∩CN \cap C additively and multiplicatively. Hence, N∩CN \cap C is normal in C+A=CA=BC + A = CA = B additively and multiplicatively. Let c0∈N∩Cc_0 \in N \cap C and b∈Bb \in B. Then there exists a∈Aa \in A and c∈Cc \in C such that b=acb = ac. Since NN is abelian,

λb(c0)=λca(c0)=λc(λa(c0))=λc(c0)∈N∩C.\lambda_b(c_0) = \lambda_{ca}(c_0) = \lambda_c(\lambda_a(c_0)) = \lambda_c(c_0) \in N \cap C.

Therefore, N∩CN \cap C is an ideal of BB. The minimality of NN yields N∩C=0N \cap C = 0. It follows that N=A+(N∩C)=AN = A + (N \cap C) = A. Hence, NN is of order pp. By induction, B/NB/N is supersoluble, and so BB is supersoluble.

Let qq be a prime. By Proposition 3.1, every minimal subbrace of a qq-subbrace SS of BB is complemented. Thus, by Proposition 3.2, SS is elementary abelian.

Proof of Corollary 1.2. By Theorem A, BB is supersoluble and every subbrace of prime power order is elementary abelian. In particular, the Sylow subbraces of BB are elementary abelian. By [1 Lemma 3.5], (B,+)(B,+) and (B,⋅)(B,\cdot) are supersoluble. Consequently, both (B,+)(B,+) and (B,⋅)(B,\cdot) are supersoluble groups with elementary abelian Sylow subgroups. By Theorem 1.1, (B,+)(B,+) and (B,⋅)(B,\cdot) are complemented groups.

The example of Theorem B

According to [2], a subbrace SS of a skew brace BB is called a maximal subbrace of BB if SS is the only proper subbrace of BB containing SS. The Frattini subbrace Φ(B)\Phi(B) of BB is defined as the intersection of all maximal subbraces of BB, if such exist, and BB otherwise ([2 Definition 24]). If BB is finite, every proper subbrace of BB is contained in a maximal subbrace of BB and so Φ(B)\Phi(B) is a proper subbrace of BB.

Lemma 4.1. Let BB be a finite skew brace with every minimal subbrace complemented in BB. Then Φ(B)=1\Phi(B) = 1.

Proof. Assume that Φ(B)≠1\Phi(B) \ne1, and let pp be a prime dividing ∣Φ(B)∣|\Phi(B)|. Then, by Cauchy’s theorem for skew braces (see [5]), Φ(B)\Phi(B) has a subbrace AA of order pp. Let CC be a complement of AA in BB. Then B=A+CB = A + C. Since CC is a proper subbrace of BB, there exists a maximal subbrace DD of BB such that C⊆DC \subseteq D. Since A⊆DA \subseteq D, it follows that B=DB = D, a contradiction. Hence Φ(B)=1\Phi(B) = 1. □

Example 4.2. Let BB be the skew brace defined in [6 Example 3.2]. This skew brace corresponds to S(18, 14) of the YangBaxter library [10] of GAP [9].

Notice that Ker⁡λ=⟨b,x⟩+\operatorname{Ker} \lambda= \langle b, x\rangle_{+} and Fix⁡(B)=⟨a,b⟩+\operatorname{Fix}(B) = \langle a, b\rangle_{+}.

⟨b⟩+=⟨b⟩∙\langle b\rangle_{+} = \langle b\rangle_{\bullet} is normal in (B,+)(B,+) and (B,⋅)(B,\cdot) and lambda invariant. Therefore, it is a cyclic ideal. Moreover, B/⟨b⟩+B/\langle b\rangle_{+} is abelian; thus, BB is supersoluble.

Every 2-subbrace has order 2, so it is elementary abelian. Also, every 3-subbrace is elementary abelian since they are contained in Fix⁡(B)\operatorname{Fix}(B), which it is the only Sylow 3-subgroup of (B,+)(B,+).

Furthermore, Ker⁡λ\operatorname{Ker} \lambda and Fix⁡(B)\operatorname{Fix}(B) are maximal subbraces since their indices is prime. They also contain every 2-subgroup and 3-subgroup of (B,+)(B,+) respectively. We show that Ker⁡λ\operatorname{Ker} \lambda is the only subbrace of order 6. The subgroups of (B,+)(B,+) with order 6 are

⟨a,x⟩+,⟨a,b+x⟩+,⟨a,2b+x⟩+,and Ker⁡λ=⟨b,x⟩+.\langle a,x\rangle_{+},\quad\langle a,b+x\rangle_{+},\quad\langle a,2b+x\rangle_{+},\quad\text{and } \operatorname{Ker}\lambda=\langle b,x\rangle_{+}.

Since

a⋅x=a+b+x,a⋅(b+x)=a+2b+x,and a⋅(2b+x)=a+x,a\cdot x=a+b+x,\quad a\cdot(b+x)=a+2b+x,\quad\text{and } a\cdot(2b+x)=a+x,

the first three subgroups are not closed under the product, hence, Ker⁡λ\operatorname{Ker} \lambda is the only subbrace of BB with order 6.

In conclusion, Ker⁡λ\operatorname{Ker} \lambda and Fix⁡(B)\operatorname{Fix}(B) are the only maximal subbraces. Therefore, Φ(B)=Ker⁡λ∩Fix⁡(B)=⟨b⟩+\Phi(B)=\operatorname{Ker}\lambda\cap\operatorname{Fix}(B)=\langle b\rangle_{+}. By Lemma 4.1, BB is not complemented.

Proof of Theorem C

Proof of Theorem C. As in the case of groups, the finite direct product of complemented skew braces is complemented (see [8]). By Proposition 3.1, subbraces of complemented skew braces are complemented. Hence, we only need to show that a square-free skew brace is complemented. Suppose that BB is a square-free skew brace. By [1 Corollary 3.9], BB is supersoluble, in particular it is soluble. Hence, by Hall’s theorem for soluble braces (see [5]), BB has a subbrace for every integer dividing ∣B∣|B| and, therefore, every subbrace has a complement in BB. □

Proof of Theorem D

Lemma 6.1. Let pp be a prime and let BB be a skew brace such that (B,+)(B,+) is pp-elementary abelian. Let f∈Fix⁡(B)f \in\operatorname{Fix}(B) and s∈Soc⁡(B)s \in\operatorname{Soc}(B). Then ⟨f,s⟩\langle f,s\rangle is finite.

Proof. Let n∈Zn \in\mathbb{Z} and write sn=λnf(s)=λfn(s)s_n=\lambda_{nf}(s)=\lambda_{f^n}(s). Consider

A=⟨f,sn∣0≤n≤p−1⟩+={αf+∑n=0p−1βnsn | 0≤α,βn≤p−1}.A=\langle f,s_n\mid0\leq n\leq p-1\rangle_+=\left\{\alpha f+\sum_{n=0}^{p-1}\beta_n s_n\ \middle|\ 0\leq\alpha,\beta_n\leq p-1\right\}.

Obviously, AA is a subgroup of (B,+)(B,+). Let us see that is a subbrace. First, notice that sn=sms_n=s_m if n≡m(modp)n\equiv m\pmod p, hence, sn∈As_n\in A for all n∈Zn\in\mathbb{Z}. Moreover, sn=nf∗s+s∈Soc⁡(B)s_n=n_f*s+s\in\operatorname{Soc}(B) for all n∈Zn\in\mathbb{Z}. Consider a,b∈Aa,b\in A. Then

a=αf+∑n=0p−1βnsn=(∑n=0p−1βnsn)(αf),b=α′f+∑n=0p−1βn′sna=\alpha f+\sum_{n=0}^{p-1}\beta_n s_n=\left(\sum_{n=0}^{p-1}\beta_n s_n\right)(\alpha f),\qquad b=\alpha'f+\sum_{n=0}^{p-1}\beta'_n s_n

for some 0≤α,α′,βn,βn′≤p−10\leq\alpha,\alpha',\beta_n,\beta'_n\leq p-1. Then

ab=a+λa(b)=λαf(α′f+∑n=0p−1βn′sn)=α′f+∑n=0p−1βn′λαf(sn)ab=a+\lambda_a(b)=\lambda_{\alpha f}\left(\alpha'f+\sum_{n=0}^{p-1}\beta'_n s_n\right)=\alpha'f+\sum_{n=0}^{p-1}\beta'_n\lambda_{\alpha f}(s_n)

Notice that

λαf(sn)=λfα(λfn(s))=λf(α+n)(s)=λ(α+n)f(s)=sα+n∈A.\lambda_{\alpha f}(s_n)=\lambda_{f^\alpha}(\lambda_{f^n}(s))=\lambda_{f^{(\alpha+n)}}(s)=\lambda_{(\alpha+n)f}(s)=s_{\alpha+n}\in A.

Hence, ab∈Aab\in A. Moreover,

a−1=(αf)−1(∑n=0p−1βnsn)−1=(−αf)(−∑n=0p−1βnsn)a^{-1}=(\alpha f)^{-1}\left(\sum_{n=0}^{p-1}\beta_n s_n\right)^{-1}=(-\alpha f)\left(-\sum_{n=0}^{p-1}\beta_n s_n\right)
=((−α)f)(∑n=0p−1(−βn)sn).=\left((-\alpha)f\right)\left(\sum_{n=0}^{p-1}(-\beta_n)s_n\right).

Since ff and sns_n have additive orders either 11 or pp, we have that (−α)f(-\alpha)f and ∑n=0p−1(−βn)sn\sum_{n=0}^{p-1}(-\beta_n)s_n are in AA. Since the product of any two elements of AA belongs to AA, it follows that a−1∈Aa^{-1}\in A. Hence AA is a finite subbrace of BB containing ⟨f,s⟩\langle f,s\rangle.

Proposition 6.2. Let BB be a multipermutational complemented skew brace such that (B,+)(B,+) is a pp-group for a prime pp. Then BB is elementary abelian. Proof. First, let us prove that (B,+)(B,+) is elementary abelian. By [7 Theorem 2.20], (B,+)(B,+) is nilpotent. We argue by induction on the nilpotency class of (B,+)(B,+).

Suppose that (B,+)(B,+) is abelian, and consider A={a∈B∣pa=0}A=\{a\in B\mid pa=0\}. Suppose that A≠BA\ne B. Since (B,+)(B,+) is abelian, (A,+)(A,+) is a characteristic subgroup of (B,+)(B,+); therefore, AA is a subbrace of BB. Let CC be a subbrace of BB such that B=A+CB=A+C and A∩C=0A\cap C=0. If C≠0C\ne0, then CC has an element of order pp by Cauchy’s theorem, a contradiction. Hence C=0C=0, B=AB=A and (B,+)(B,+) is elementary abelian.

Now suppose that the nilpotency class of (B,+)(B,+) is nn and that the result is true for multipermutational complemented skew braces whose additive group has nilpotency class n−1n-1. Since Z(B,+)Z(B,+) is a characteristic subgroup of (B,+)(B,+), it is a subbrace of BB, so that there exists a subbrace DD complementing Z(B,+)Z(B,+) in BB. We have that (D,+)≅(B,+)/Z(B,+)(D,+)\cong(B,+)/Z(B,+). By [7 Lemma 2.6 and Theorem 2.20] and Proposition 3.1, DD is a multipermutational complemented skew brace and the nilpotency class of (D,+)(D,+) is less than nn. By induction, (D,+)(D,+) is elementary abelian. Since Z(B,+)Z(B,+) is an abelian subbrace of BB, we have that Z(B,+)Z(B,+) is elementary abelian. Finally, (B,+)=Z(B,+)+(D,+)(B,+)=Z(B,+)+(D,+), so that (B,+)(B,+) is elementary abelian.

Next, we show that BB is a trivial brace by induction on the multipermutational level nn of BB. It is clear that we may assume n>1n>1 and that the result is true for multipermutational complemented skew braces of multipermutational level n−1n-1. Let CC be a complement of Soc⁡(B)\operatorname{Soc}(B) in BB. Then C≅B/Soc⁡(B)C\cong B/\operatorname{Soc}(B). By [7 Lemma 2.6 and Theorem 2.20] and Proposition 3.1, CC is a multipermutational complemented skew brace with multipermutational level n−1n-1. By induction, CC is a trivial brace. Let c0∈Cc_0\in C and b∈Bb\in B. Then there exists s∈Soc⁡(B)s\in\operatorname{Soc}(B) and c∈Cc\in C such that b=scb=sc. Since s∈Soc⁡(B)s\in\operatorname{Soc}(B) and CC is trivial, we have that

λb(c0)=λsc(c0)=λs(λc(c0))=c0.\lambda_b(c_0)=\lambda_{sc}(c_0)=\lambda_s(\lambda_c(c_0))=c_0.

Hence, c0∈Fix⁡(B)c_0\in\operatorname{Fix}(B), and C⊆Fix⁡(B)C\subseteq\operatorname{Fix}(B).

Consider f∈Fix⁡(B)f\in\operatorname{Fix}(B) and s∈Soc⁡(B)s\in\operatorname{Soc}(B). By Lemma 6.1, A=⟨f,s⟩A=\langle f,s\rangle is finite. Therefore, AA is a finite complemented pp-skew brace. Thus, by Theorem A, AA is trivial, in particular, λf(s)=s\lambda_f(s)=s.

Let a,b∈Ba,b\in B. Since B=Fix⁡(B)+Soc⁡(B)=Fix⁡(B)Soc⁡(B)B=\operatorname{Fix}(B)+\operatorname{Soc}(B)=\operatorname{Fix}(B)\operatorname{Soc}(B), there exists f,f′∈Fix⁡(B)f,f'\in\operatorname{Fix}(B) and s,s′∈Soc⁡(B)s,s'\in\operatorname{Soc}(B) such that a=fsa=fs and b=f′+s′b=f'+s'. Therefore,

λa(b)=λfs(f′+s′)=f′+λf(s′)=f′+s′=b.\lambda_a(b)=\lambda_{fs}(f'+s')=f'+\lambda_f(s')=f'+s'=b.

Consequently, BB is a trivial brace.

Recall that if GG is a periodic abelian group and pp is a prime, then P={g∈G:png=0P = \{g \in G : p^{n}g = 0 for some n≥0}n \ge0\} is the unique Sylow pp-subgroup of GG. If GG is a skew brace, then PP is the Sylow pp-subbrace of GG.

Lemma 6.3. Let BB be a skew brace and let II be an ideal of BB contained in Soc⁡(B)\operatorname{Soc}(B). If Soc⁡(B)\operatorname{Soc}(B) is periodic, then the Sylow subgroups of (I,+)(I,+) are ideals of BB.

Proof. Let (P,+)(P,+) be a Sylow pp-subbrace of (I,+)(I,+) for some prime pp. Since II is a trivial brace, PP is a subbrace of BB. Moreover, (P,+)(P,+) and (P,⋅)(P,\cdot) are characteristic in (I,+)(I,+) and (I,⋅)(I,\cdot), respectively. Since (I,+)(I,+) and (I,⋅)(I,\cdot) are normal in (B,+)(B,+) and (B,⋅)(B,\cdot), respectively, it follows that (P,+)(P,+) and (P,⋅)(P,\cdot) are characteristic in (B,+)(B,+) and (B,⋅)(B,\cdot) respectively. Finally, since II is λ\lambda-invariant and (P,+)(P,+) is characteristic in (I,+)(I,+), we have that PP is λ\lambda-invariant. Therefore, PP is an ideal of BB. □\square

Lemma 6.4. If BB is a complemented skew brace, then Soc⁡(B)\operatorname{Soc}(B) is periodic.

Proof. Assume that Soc⁡(B)\operatorname{Soc}(B) has an element ss of infinite additive order. Since Soc⁡(B)\operatorname{Soc}(B) is an abelian subbrace, the additive infinite cyclic group ⟨s⟩\langle s\rangle and its subgroup ⟨2s⟩\langle2s\rangle are subbraces of BB. By Proposition 3.1, ⟨2s⟩\langle2s\rangle has a complement in ⟨s⟩\langle s\rangle, a contradiction. Hence, Soc⁡(B)\operatorname{Soc}(B) is periodic. □\square

Proposition 6.5. Let BB be a complemented skew brace and let II be a non-zero ideal of BB contained in Soc⁡(B)\operatorname{Soc}(B). Then there exists an ideal of BB of prime order contained in II.

Proof. By Lemma 6.4, II is a periodic abelian skew brace. Let (P,+)(P,+) be a non-zero Sylow subgroup of (I,+)(I,+). By Lemma 6.3, PP is an ideal of BB. By Proposition 3.1, PP is complemented. Since PP is a trivial brace, we can consider a subbrace of PP of order pp, and then take a complement UU in PP. Then UU is a subbrace of PP such that ∣P:U∣=p|P:U|=p.

Consider a complement CC of UU in BB, and let V=C∩PV=C\cap P. Then V∩U⊆U∩C=0V\cap U\subseteq U\cap C=0, and

P=P∩B=P∩(UC)=U(P∩C)=UV.P=P\cap B=P\cap(UC)=U(P\cap C)=UV.

Hence, ∣V∣=p|V|=p. By Proposition 3.1, CC is complemented. Let QQ be a complement of VV in CC. Since PP is an ideal of BB, we have that V=C∩PV=C\cap P is an ideal of CC; in particular, VV is lambda invariant by the elements of QQ. Moreover, since V⊆PV\subseteq P, we have that VV is lambda invariant by the elements of PP. Thus, VV is lambda invariant by the elements of PQ=UVQ=UC=BPQ=UVQ=UC=B. Consequently, VV is a left ideal of BB. Finally, note that every left ideal of BB contained in Soc⁡(B)\operatorname{Soc}(B) is an ideal. □\square

Lemma 6.6. Let BB be a skew brace and let I⊆JI \subseteq J be ideals of BB such that JJ is BB-hypercyclic. Then J/IJ/I is B/IB/I-hypercyclic.

Proof. Let

J0=0⊆⋯⊆Jα⊆Jα+1⊆⋯⊆Jγ=JJ_0 = 0 \subseteq\cdots\subseteq J_\alpha\subseteq J_{\alpha+1} \subseteq\cdots\subseteq J_\gamma= J

be an ascending series of ideals of BB such that Jα+1/JαJ_{\alpha+1}/J_\alpha has prime order. Denote Iα=(Jα+I/I)I_\alpha= (J_\alpha+ I/I). Then

I0=0⊆⋯⊆Iα⊆Iα+1⊆⋯⊆Iγ=J/II_0 = 0 \subseteq\cdots\subseteq I_\alpha\subseteq I_{\alpha+1} \subseteq\cdots\subseteq I_\gamma= J/I

is an ascending series of ideals of B/IB/I. Moreover,

Iα+1/Iα≅(Jα+1+I)/(Jα+I)=(Jα+1+(Jα+I))/(Jα+I)≅Jα+1/(Jα+1∩(Jα+I))=Jα+1/(Jα+(Jα+1∩I)).\begin{aligned} I_{\alpha+1}/I_\alpha&\cong(J_{\alpha+1}+I)/(J_\alpha+I) \\ &= (J_{\alpha+1}+(J_\alpha+I))/(J_\alpha+I) \\ &\cong J_{\alpha+1}/(J_{\alpha+1}\cap(J_\alpha+I)) \\ &= J_{\alpha+1}/(J_\alpha+(J_{\alpha+1}\cap I)). \end{aligned}

Since Jα+1/JαJ_{\alpha+1}/J_\alpha has prime order, either Iα+1/IαI_{\alpha+1}/I_\alpha has prime order or Iα+1=IαI_{\alpha+1}=I_\alpha. Therefore, if we remove the repeated terms from the chain we have an ascending chain of ideals of B/IB/I whose successive quotients have prime order. Thus, J/IJ/I is B/IB/I-hypercyclic. □

Proposition 6.7. Let BB be a complemented skew brace. Then Soc⁡(B)\operatorname{Soc}(B) is BB-hypercyclic.

Proof. Let S\mathcal{S} be the set of all BB-hypercyclic ideals of BB contained in Soc⁡(B)\operatorname{Soc}(B). Obviously 0∈S0\in\mathcal{S}, so it is not an empty set. Let

I0=0⊆⋯⊆Iα⊆Iα+1⊆⋯I_0 = 0 \subseteq\cdots\subseteq I_\alpha\subseteq I_{\alpha+1} \subseteq\cdots

be an ascending chain of elements of S\mathcal{S}. Consider I=⋃αIαI=\bigcup_\alpha I_\alpha. Clearly, II is an ideal of BB contained in Soc⁡(B)\operatorname{Soc}(B). Moreover, by Lemma 6.6, we can find an ascending series of ideals of BB from IαI_\alpha to Iα+1I_{\alpha+1} such that the quotients have prime order. Therefore, we can construct an ascending series of ideals of BB from 00 to II such that the quotients have prime order. Therefore, I∈SI\in\mathcal{S}.

By Zorn’s lemma, there exists MM, a maximal element of S\mathcal{S}. If M⊊Soc⁡(B)M\subsetneq\operatorname{Soc}(B), then Soc⁡(B)/M\operatorname{Soc}(B)/M is a non-zero ideal of B/MB/M contained in Soc⁡(B/M)\operatorname{Soc}(B/M). By Proposition 3.1, B/MB/M is a complemented skew brace. Then, by Proposition 6.5, there exists an ideal I/MI/M of B/MB/M of prime order contained in Soc⁡(B)/M\operatorname{Soc}(B)/M. Since MM is BB-hypercyclic and I/MI/M has prime order, II is BB-hypercyclic. This contradicts the maximality of MM. Thus, Soc⁡(B)=M\operatorname{Soc}(B)=M is BB-hypercyclic. □

Proof of Theorem D. We show that BB is hypercyclic by induction on the multipermutational level of BB. By Proposition 6.7, Soc⁡(B)\operatorname{Soc}(B) is BB-hypercyclic and, by Proposition 3.1, B/Soc⁡(B)B/\operatorname{Soc}(B) is a multipermutational complemented skew brace of smaller multipermutational level. By induction, B/Soc⁡(B)B/\operatorname{Soc}(B) is hypercyclic, and hence BB is hypercyclic.

By [7 Lemma 2.6 and Theorem 2.20] and Proposition 3.1, every subbrace of BB is a complemented multipermutational skew brace. Hence, by Proposition 6.2, every subbrace whose additive group is a pp-group is elementary abelian for every prime pp.

Acknowledgements

The first author was supported by Guangdong Basic and Applied Research Foundation (2025A1515012072). The second and fourth authors are supported by the grant PID2024-159495NB-I00, funded byMI-CIU/AEI/10.13039/501100011033 and by ERDF/EU, and by the grant CI-AICO/2023/007 from the Conselleria d’Educació, Cultura, Universitats i Ocupació of the Generalitat Valenciana. The second author is funded by the Spanish Ministry of Science, Innovation and Universities through an FPU predoctoral grant FPU24/01319. The third author is grateful to the Universitat de València for their support to host researchers affected by the war in Ukraine.

References

  1. [1]A. Ballester-Bolinches, R. Esteban-Romero, M. Ferrara, V. Pérez-Calabuig, and M. Trombetti. Finite skew braces of square-free order and supersolubility. Forum Math. Sigma, 12:e39, 2024.
  2. [2]A. Ballester-Bolinches, R. Esteban-Romero, P. Jiménez-Seral, and V. Pérez-Calabuig. Soluble skew left braces and soluble solutions of the Yang–Baxter equation. Adv. Math., 455:109880, 2024.
  3. [3]A. Ballester-Bolinches and X. Guo. On complemented subgroups of finite groups. Arch. Math. (Basel), 72(3):161–166, 1999.
  4. [4]A. Ballester-Bolinches, P. Pérez-Altarriba, and V. Pérez-Calabuig. On finite trifactorised groups and dominance Sylow and Hall theorems for finite skew braces, 2026. Preprint.
  5. [5]A. Ballester-Bolinches, P. Pérez-Altarriba, and V. Pérez-Calabuig. On finite trifactorised groups and Sylow and Hall theorems for skew braces. Ricerche Mat., 2026.
  6. [6]Sergio Camp-Mora and P. Pérez-Altarriba. Carter-like theorems for skew braces, 2026. preprint.
  7. [7]F. Cedó, A. Smoktunowicz, and L. Vendramin. Skew left braces of nilpotent type. Proc. London Math. Soc., 118(6):1367–1392, 2019.DOI
  8. [8]P. Hall. Complemented Groups. J. London Math. Soc., s1-12(3):201–204, 1937.DOI
  9. [9]The GAP Group. GAP – Groups, Algorithms, and Programming, Version 4.15.1. The GAP Group, 2025. http://www.gap-system.org.
  10. [10]L. Vendramin and O. Konovalov. YangBaxter: combinatorial solutions for the Yang–Baxter equation, 2025. version 0.10.7, https://gap-packages.github.io/YangBaxter/.

Paper details

vlnc:W_uz2ghcwh
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents