Loading [MathJax]/jax/output/SVG/jax.js
Research article

Representation rings of extensions of Hopf algebra of Kac-Paljutkin type

  • In this paper, we focus on studying two classes of finite dimensional Δ-associative algebras, which are extensions of a family of 2n2-dimensional Kac-Paljutkin type semi-simple Hopf algebras H2n2. All their indecomposable modules are classified. Furthermore, their representation rings are described by generators with suitable relations.

    Citation: Dong Su, Shilin Yang. Representation rings of extensions of Hopf algebra of Kac-Paljutkin type[J]. Electronic Research Archive, 2024, 32(9): 5201-5230. doi: 10.3934/era.2024240

    Related Papers:

    [1] Shanshan Liu, Abdenacer Makhlouf, Lina Song . The full cohomology, abelian extensions and formal deformations of Hom-pre-Lie algebras. Electronic Research Archive, 2022, 30(8): 2748-2773. doi: 10.3934/era.2022141
    [2] Hongliang Chang, Yin Chen, Runxuan Zhang . A generalization on derivations of Lie algebras. Electronic Research Archive, 2021, 29(3): 2457-2473. doi: 10.3934/era.2020124
    [3] Yaguo Guo, Shilin Yang . Projective class rings of the category of Yetter-Drinfeld modules over the $ 2 $-rank Taft algebra. Electronic Research Archive, 2023, 31(8): 5006-5024. doi: 10.3934/era.2023256
    [4] Wen Teng, Xiansheng Dai . Nonabelian embedding tensors on 3-Lie algebras and 3-Leibniz-Lie algebras. Electronic Research Archive, 2025, 33(3): 1367-1383. doi: 10.3934/era.2025063
    [5] Xiuhai Fei, Haifang Zhang . Additivity of nonlinear higher anti-derivable mappings on generalized matrix algebras. Electronic Research Archive, 2023, 31(11): 6898-6912. doi: 10.3934/era.2023349
    [6] Ming Ding, Zhiqi Chen, Jifu Li . The properties on F-manifold color algebras and pre-F-manifold color algebras. Electronic Research Archive, 2025, 33(1): 87-101. doi: 10.3934/era.2025005
    [7] Francisco Javier García-Pacheco, María de los Ángeles Moreno-Frías, Marina Murillo-Arcila . On absolutely invertibles. Electronic Research Archive, 2024, 32(12): 6578-6592. doi: 10.3934/era.2024307
    [8] Quanguo Chen, Yong Deng . Hopf algebra structures on generalized quaternion algebras. Electronic Research Archive, 2024, 32(5): 3334-3362. doi: 10.3934/era.2024154
    [9] Agustín Moreno Cañadas, Robinson-Julian Serna, Isaías David Marín Gaviria . Zavadskij modules over cluster-tilted algebras of type $ \mathbb{A} $. Electronic Research Archive, 2022, 30(9): 3435-3451. doi: 10.3934/era.2022175
    [10] Xueru Wu, Yao Ma, Liangyun Chen . Abelian extensions of Lie triple systems with derivations. Electronic Research Archive, 2022, 30(3): 1087-1103. doi: 10.3934/era.2022058
  • In this paper, we focus on studying two classes of finite dimensional Δ-associative algebras, which are extensions of a family of 2n2-dimensional Kac-Paljutkin type semi-simple Hopf algebras H2n2. All their indecomposable modules are classified. Furthermore, their representation rings are described by generators with suitable relations.



    As an invariant of a finite tensor category, the representation ring plays a very important role, having gained great attention in recent years. For example, Chen et al. [1] constructed the representation rings of Taft algebras Hn(q). Wang et al. studied the Green rings of finite-dimensional pointed rank-one Hopf algebras of nilpotent and non-nilpotent types in [2,3]. It should be noted that the small quantum group [4,5] and the Drinfeld double of a Taft algebra [6,7] are not of finite representation type. Their representation rings are not finitely generated, as described by [8] and [9,10], respectively. In [11,12], the authors considered the Grothendieck rings of the quotient algebras of Wu-Liu-Ding algebras (see [13,14] for definitions), and provided their Casimir numbers, which are a class of non-pointed semi-simple Hopf algebras. The representations and Grothendieck rings of the Hopf algebra H2n2 are described in [15]. Furthermore, Guo and Yang [16] explicitly described the Grothendieck rings of the category of Yetter-Drinfeld modules over H2n2 by generators and relations. Sun et al. [17] described the structure of the representation ring of the small quasi-quantum group. For more results on representations and representation rings, one can refer to [18,19,20,21,22,23,24,25].

    In order to understand and extend the concept of representation rings, we can weaken the definition of weak Hopf algebras to more general cases. The most interesting one is the so-called Δ-associative algebra, which was introduced in [26], where the representation ring of Δ-associative algebra was defined and described for Kac-Paljutkin Hopf algebra K8(i.e., H8). One consideration is that the representation rings of Δ-associative algebras may not be commutative and may not even contain the identity (see [26]). In this paper, we provide two classes of Δ-associative algebras to understand them. Roughly speaking, based on Kac-Paljutkin-type Hopf algebras H2n2 (see [27] or [15]), which are generalizations of the 8-dimensional Kac-Paljutkin Hopf algebra, we weaken the definition of Hopf algebra H2n2 to obtain two classes of Δ-associative algebras, ¯H2n2 and ˆH2n2. One can see that ¯H2n2 is still a weak Hopf algebra but ˆH2n2 is a non-trivial Δ-associative algebra. Following that, we describe their representations as well as their representation rings. We show that the representation ring r(¯H2n2) is a commutative ring, but the representation ring r(ˆH2n2) is noncommutative with no identity. However, we can embed the ring r(ˆH2n2) into a ring r(ˆH2n2) with an identity in the natural way. The rings r(¯H2n2) and r(ˆH2n2) are described by generators with suitable relations.

    The paper is organized as follows. In Section 2, we recall the definition of Δ-associative algebra, and review the definition and representations of the semi-simple Hopf algebra H2n2. In Section 3, by weakening the definition of Hopf algebra H2n2, we obtain two classes of Δ-associative algebras ¯H2n2 and ˆH2n2, where ¯H2n2 is a weak Hopf algebra, and ˆH2n2 is just a non-trivial Δ-associative algebra. Some properties of ¯H2n2 and ˆH2n2 are discussed. In Section 4, all irreducible modules of ¯H2n2 are listed, and there are (n2+7n+2)/2 non-isomorphic finite dimensional irreducible modules for ¯H2n2. The decomposition formulas of the tensor product of two arbitrary irreducible ¯H2n2-modules are also established. The representation ring r(¯H2n2), described by generators and relations, shows that r(¯H2n2) is a commutative ring. In Section 5, the representations and representation ring of ˆH2n2 are studied. We found that the representation ring r(ˆH2n2) is a noncommutative ring with no identity, which can be embedded into a ring r(ˆH2n2) with an identity.

    Throughout this research, we work over a fixed algebraically closed field k of characteristic 0, unless otherwise stated. All algebras, Hopf algebras, and modules are defined over k; all modules are left modules and finite dimensional; all maps are k-linearity; dim,, and hom stand for dimk, k, and homk, respectively.

    The definition of weak Hopf algebra was introduced by Li (see [28]). We recall that a k-bialgebra (H,μ,η,Δ,ϵ) is called a weak Hopf algebra if there exists a map Thom(H,H) such that TidT=T and idTid=id, where is the convolution map in hom(H,H). By weakening the definition of weak Hopf algebras, the following definition is established.

    Definition 2.1. An associative k-algebra A with an identity is called a Δ-associative algebra if there exists an algebra homomorphism Δ:AAA such that (Δid)Δ=(idΔ)Δ.

    All Hopf algebras, bialgebras, and weak Hopf algebras are Δ-associative algebras. If the Δ-associative algebra A is not a coalgebra, A is said to be non-trivial.

    In the sequel, we always assume that A is a Δ-associative algebra, and the Sweedler's notations [29] are used. For example, for aA, we denote

    Δ(a)=(a)a(1)a(2).

    Now, suppose that M and N are two A-modules, then MN is also an A-module defined as follows:

    a(mn)=(a)a(1)ma(2)n,

    for all mM,nN.

    We denote by [M] the isomorphism class of an A-module M and

    P={[M]|MmodA}.

    Let R(A) be a free abelian group spanned by P. For all [M],[N]P, we define [M][N]=[MN], and it is easy to see that R(A) is a ring. Let

    r(A)=R(A)[MN][M][N],

    where [MN][M][N] is the ideal of R(A) generated by [MN][M][N] for all [M],[N]P, then r(A) is also a ring and is called the representation ring or the Green ring of A. If A is a non-trivial Δ-associative algebra, then r(A) may have no identity [kε], the trivial 1-dimensional A-module. In this case, let

    r(A)={(k,α)|kk,αr(A)}.

    For (k,α), (k,α)r(A), we define the addition and multiplication in r(A) as the following

    (k,α)+(k,α)=(k+k,α+α),
    (k,α)(k,α)=(kk,kα+kα+αα).

    Then, r(A) is a ring with the identity (1,0), and r(A) is embedded into r(A) naturally.

    In the sequel, we always assume that n>1 and q is a primitive n-th root of unity.

    The Hopf algebra H2n2 can be found in [15,27], which is a generalization of Kac-Paljutkin Hopf algebra.

    Definition 2.2. The Hopf algebra H2n2 is an associative algebra generated by x,y and z, with the following relations

    xn=1, yn=1, xy=yx, zx=yz, zy=xz,z2=1nn1i,j=0qijxiyj.

    The comultiplication, counit, and antipode are as follows

    Δ(x)=xx,ϵ(x)=1,S(x)=x1,Δ(y)=yy,ϵ(y)=1,S(y)=y1,Δ(z)=1nn1i,j=0qijxizyjz,ϵ(z)=1,S(z)=z.

    The Hopf algebra H2n2 is a 2n2-dimensional semi-simple Hopf algebra. It is of a basis {xiyj, xiyjz|0i,jn1}. Indeed,

    =(n1i=0xi)(n1j=0yj)(1+z)

    is the left and right integral of H2n2 and ϵ()=2n20. It is also a quasi-triangular Hopf algebra with a universal R-matrix

    R=1nn1i,j=0qijxjyni.

    Therefore, the representation ring of H2n2 is semi-simple and commutative. The representations and representation ring of the Hopf algebra H2n2 are described in [15]. We list them as follows, where q is any primitive n-th root of unity, but q=e2pπi/n with an even number pZ if n is odd for convenience.

    Set

    σ(m)={1,0mn1,1,nm2n1.

    (a) 1-dimensional irreducible H2n2-module Sm,mZ2n: it is of basis vm, and the actions of H2n2 on Sm are

    x.vm=qmvm,y.vm=qmvm,z.vm=σ(m)qm22vm;

    (b) 2-dimensional irreducible H2n2-module Si,j,0i<jn1: it is of basis vij1 and vij2, and the actions of H2n2 on Si,j are

    x.vij1=qivij1,y.vij1=qjvij1,z.vij1=vij2,x.vij2=qjvij2,y.vij2=qivij2,z.vij2=qijvij1.

    The set

    S={Sm,Si,jmZ2n,0i<jn1}

    forms a complete list of non-isomorphic irreducible H2n2-modules.

    Lemma 2.3. [15,Theorem 1] The decomposition formulas of tensor product of two H2n2-modules in S are as follows.

    (1) (a) For all 0m,mn1, we have

    SmSmSm+m(modn);

    (b) for all 0mn1,nm2n1 and nm+m2n1, or 0mn1,nm2n1 and nm+m2n1, we have

    SmSmSm+m;

    (c) for all 0mn1,nm2n1 and 2nm+m3n1, or 0mn1,nm2n1 and 2nm+m3n1, we have

    SmSmSm+mn;

    (d) for all nm,m2n1 and 2nm+m3n1, we have

    SmSmSm+m(mod2n);

    (e) for all nm,m2n1 and 3nm+m4n1, we have

    SmSmSm+m(mod3n).

    (2) For all mZ2n,0i<jn1, we have

    SmSi,jSm+i(modn),j+m(modn)Si,jSm.

    (3) Set I1={0i<j,k<l<ni+kj+l(modn)} and I2={0i<j,k<l<ni+lj+k(modn)}, then

    (a) if i,j,k,lI1I2, we have

    Si,jSk,lSi+k(modn),j+l(modn)Si+l(modn),j+k(modn);

    (b) if i,j,k,lI1I2, we have

    Si,jSk,lSi+kSj+lSi+l(modn),j+k(modn);

    (c) if i,j,k,lI2I1, we have

    Si,jSk,lSi+k(modn),j+l(modn)Si+l(modn)Sj+k(modn)+n;

    (d) if n is even, and i,j,k,lI1I2, we have

    Si,jSk,lSi+kSj+lSi+l(modn)Sj+k(modn)+n.

    Let

    Al(y,z,u)=[zuy0001zy00000zy0001z]l×l

    and

    Bl(y,z)=det(Al(y,z,1+yn)),Dl(y,z,x)=det(Al(y,z,1+y2m1x)).

    Then, we have

    B0(y,z)=1+yn,B1(y,z)=z,B2(y,z)=z2yyn+1;
    D0(y,z,x)=1+y2m1x,D1(y,z,x)=z,D2(y,z,x)=z2yy2mx.

    [15,Theorem 2,Theorem 3] can be stated as follows.

    Lemma 2.4. Assume that n2 is an integer, we have

    (1) if n3 is odd and m=n12, then

    r(H2n2)Z[y,z]/y2n1,zynz,Bm+1(y,z)ym+1Bm(y,z);

    (2) if n=2, then

    r(H2n2)Z[x,y,z]/y21,x2y2,zxzy,zzy,z2xyxy1;

    (3) if n>2 is even with m=n2, then

    r(H2n2)Z[x,y,z]/xn1,Dm+1(y,z,x)ym+1Dm1(y,z,x),x2y2,zxzy,Dm(y,z,x)ymDm(y,z,x).

    In this section, we establish two classes of extensions of Hopf algebra H2n2 of Kac-Paljutkin type, which are denoted by ¯H2n2 and ˆH2n2, where ¯H2n2 is a weak Hopf algebra in the sense of Li in [28], but ˆH2n2 is a non-trivial Δ-associative algebra.

    Definition 3.1. Let ¯H2n2 be an associative algebra generated by x,y, and z, with the following relations

    xn+1=x, yn+1=y, xy=yx,z=yzxn1, z=xzyn1,z2=1nni,j=1qijxiyj.

    For convenience, we set

    ei=1nnl=1qilxl,fi=1nnk=1qikyk,

    for i=1,2,,n, then

    xsei=qisei,ytei=eiyt,xsfi=fixs,ytfi=qitfi,

    for s,t=1,2,,n, and

    z2=nj=1ejyj=ni=1xifi.

    We also have

    Δ(z)=(nj=1ejyj)(zz)=(ni=1xifi)(zz).

    Let J1=xn,J2=yn and J=J1J2.

    The following lemma helps us to understand the deep relations between the algebra ¯H2n2 and the algebra H2n2.

    Lemma 3.2. z=yzxn1, z=xzyn1 if and only if zJ=Jz=z, zx=yz, zy=xz.

    Proof. Indeed, assume that z=yzxn1, z=xzyn1, then

    z=ynzxn(n1)=J2zJ1andz=xnzyn(n1)=J1zJ2.

    It follows that

    z=J1zJ2=J1J2zJ1J2=JzJ

    and zJ=Jz=z. Also,

    zx=yzxn=yzJ1=y(J2zJ1)=yz, zy=xzyn=xzJ2=x(J1zJ2)=xz.

    Conversely, if zJ=Jz=z and xz=zy, yz=zx, then

    zJi=zJJi=zJ=z=Jz=JiJz=Jiz

    for i=1,2. So we have

    z=zJ2=zyn=xzyn1, z=zJ1=zxn=yzxn1.

    The result follows.

    It is easy to see that

    {J,J1J,J2J,1+JJ1J2}

    is a set of orthogonal central idempotents of ¯H2n2.

    Now, we define three maps Δ:¯H2n2¯H2n2¯H2n2, ε:¯H2n2k, and T:¯H2n2¯H2n2 as follows

    Δ(x)=xx,ϵ(x)=1,T(x)=xn1,Δ(y)=yy,ϵ(y)=1,T(y)=yn1,Δ(z)=1nni,j=1qijxizyjz,ϵ(z)=1,T(z)=z.

    We get the first main result as follows.

    Theorem 3.3. ¯H2n2 is a weak Hopf algebra with the weak antipode T.

    Proof. We will prove this theorem in three steps.

    a) The k-map Δ keeps the defining relations, hence it can be extended into the whole algebra ¯H2n2 such that Δ is an algebra homomorphism. Indeed, we note that

    ei=1nnl=1qilxl,fi=1nnk=1qjkyk,fori=1,2,,n,z2=nj=1ejyj=ni=1xifi,Δ(z)=1nni,j=1qijxizyjz=(nj=1ejyj)(zz)=(ni=1xifi)(zz).

    So it is easy to see that

    (Δ(x))n+1=Δ(x),(Δ(y))n+1=Δ(y),Δ(x)Δ(y)=Δ(y)Δ(x),Δ(y)Δ(z)Δ(x)n1=(yy)(1nni,j=1qijxizyjz)(xn1xn1)=1nni,j=1qijxiyzxn1yjyzxn1=1nni,j=1qijxizyjz=Δ(z),Δ(x)Δ(z)Δ(y)n1=(xx)(1nni,j=1qijxizyjz)(yn1yn1)=1nni,j=1qijxixzyn1yjxzyn1=1nni,j=1qijxizyjz=Δ(z),(Δ(z))2=(nj=1ejyj)(zz)(ni=1xifi)(zz)=ni,j=1ejyi+jeiyi+j=1n2ni,j,l,k=1qiljkxlyi+jxkyi+j=1n2ni,j,l,k=1qjl+i(lk)xlyjxkyj=1nni,j=1qijxiyjxiyj=1nni,j=1qijΔ(x)iΔ(y)j.

    Hence, Δ can define a homomorphism of algebra. Similarly, ε can also define a homomorphism of algebra.

    b) (¯H2n2,Δ,ε) is a coalgebra. Indeed, the k-map Δ is coassociative. To see this fact, we have

    (Δid)Δ(x)=xxx=(idΔ)Δ(x),
    (Δid)Δ(y)=yyy=(idΔ)Δ(y),

    and

    (Δid)Δ(z)=(Δid)(ni=1xizfiz)definitionand(a)__ni,j=1xiejzxiyjzfiz=ni,j=1qijejzxiyjzfiz,=(idΔ)(nj=1ejzyjz)=(idΔ)Δ(z).

    Hence,

    (Δid)Δ(a)=(idΔ)Δ(a)

    for a{x,y,z}. Then, the map Δ is coassociative in ¯H2n2 by the statement a). Furthermore, ε also satisfies the counit axiom.

    By the statements a) and b), we see that ¯H2n2 is a bialgebra.

    c) The map T can define a weak antipode of ¯H2n2 in a natural way. To see this fact, we note that

    (T(x))n+1=((x)n1)n+1=xn1=T(x);(T(y))n+1=((y)n1)n+1=yn1=T(y);T(y)T(x)=(y)n1(x)n1=(x)n1(y)n1=T(x)T(y);(T(x))n1T(z)T(y)=((x)n1)n1z(y)n1=J1z=z=T(z);(T(y))n1T(z)T(x)=((y)n1)n1z(x)n1=J2z=z=T(z);1nni,j=1qij(T(y))j(T(x))i=1nni,j=1qijy(n1)jx(n1)i=1nni,j=1qijxiyj=z2=(T(z))2.

    Thus, the map T can be extended into an anti-algebra homomorphism T:¯H2n2¯H2n2.

    Furthermore, it is easy to see that

    (idTid)(x)=xT(x)x=xn+1=x=id(x),(TidT)(x)=T(x)xT(x)=xn1=T(x),(idTid)(y)=yT(y)y=yn+1=y=id(y),(TidT)(y)=T(y)yT(y)=yn1=T(y),(idTid)(z)=ni,j=1qijejzT(xiyjz)fiz=ni,j=1qijejz2y(n1)jx(n1)ifiz=ni,j=1qijej(nk=1ekyk)fiz=ni,j=1ejfiz=Jz=z=id(z),(TidT)(z)=ni,j=1qijT(ejz)xiyjzT(fiz)=ni,j=1qijze(n1)jz2f(n1)i=ni,j=1qijze(n1)j(nk=1ekyk)f(n1)i=zni,j=1ejfi=Jz=z=T(z).

    In addition, we have

    (x)T(x(1))x(2)=J1=(x)x(1)T(x(2)),(y)T(y(1))y(2)=J2=(y)y(1)T(y(2)),(z)T(z(1))z(2)=ni=1T(xiz)fiz=ni=1ynieiz2=ni=1ynei=J1J2=J,(z)z(1)T(z(2))=nj=1ejzT(yjz)=nj=1ejz2ynj=ni=1ejyn=J1J2=J.

    The elements J1,J2 and J are all the center of ¯H2n2 by Lemma 3.2. Hence,

    (a)a(1)T(a(2)),(a)T(a(1))a(2)C(¯H2n2),

    the center of ¯H2n2, for all a¯H2n2.

    The above facts show that

    (a)a(1)T(a(2))a(3)=a=id(a),(a)T(a(1))a(2)T(a(3))=T(a)

    for all a¯H2n2 by induction.

    This means that T can indeed define a weak antipode of ¯H2n2, and hence ¯H2n2 is a weak Hopf algebra. Here, we point that ¯H2n2 is a noncommutative and noncocommutative weak Hopf algebra.

    Now, we assume that

    J¯H2n2=¯A0,(J1J)¯H2n2=¯A1,(J2J)¯H2n2=¯A2,(1+JJ1J2)¯H2n2=¯A3.

    One can get the following result.

    Proposition 3.4. As algebras, we have

    (1) ¯H2n2=¯A0¯A1¯A2¯A3;

    (2) ¯A0H2n2;

    (3) ¯A1¯A2k[g]/(gn1);

    (4) ¯A3k.

    Proof. (1) The first statement follows from the fact that {J,J1J,J2J,1+JJ1J2} is a set of orthogonal central idempotents.

    (2) For the statement, we note that ¯A0 can be viewed as a subalgebra of ¯H2n2 with the identity element J, generated by Jx, Jy, and Jz with the relations

    (Jx)n=J,(Jy)n=J,(Jx)(Jy)=(Jy)(Jx),(Jz)(Jx)=(Jy)(Jz),(Jz)(Jy)=(Jx)(Jz),(Jz)2=1nn1i,j=0qij(Jx)i(Jy)j.

    Let ρ0:H2n2¯A0 be the map defined by

    ρ0(1)=J,ρ0(x)=Jx,ρ0(y)=Jyρ0(z)=Jz,

    it is straightforward to see that ρ0 is a well-defined surjective algebra homomorphism. Let ϕ:¯H2n2H2n2 be the map given by

    ϕ(1)=1,ϕ(x)=x,ϕ(y)=y,ϕ(z)=z.

    The map ϕ is also a well-defined algebra epimorphism. Considering the restriction ϕ|¯A0 of ϕ on ¯A0, we have ϕ|¯A0ρ0=idH2n2. Hence, ρ0 is injective and ¯A0H2n2 as algebras.

    (3) We note that ¯A1 and ¯A2 are the subalgebras of ¯H2n2 with the unit J1J and J2J, respectively. In ¯A1, we have

    (J1J)y=J1yJ1y=0,(J1J)z=zz=0.

    It follows that ¯A1 is generated by (J1J)x, with the relation

    [(J1J)x]n=J1J.

    Similarly, the subalgebra ¯A2 is generated by (J2J)y, with the relations

    [(J2J)y]n=J2J,

    and (J2J)x=J2xJ2x=0,(J2J)z=zz=0. Finally, it is easy to see that

    ¯A1¯A2k[g]/(gn1).

    (4) It is obvious that ¯A3 is a one-dimensional subalgebra of ¯H2n2 with the unit 1+JJ1J2. Indeed,

    (1+JJ1J2)x=0,(1+JJ1J2)y=0,(1+JJ1J2)z=0.

    So, ¯A3=k(1+JJ1J2).

    The proof is completed.

    By Proposition 3.4, ¯H2n2 is a semi-simple algebra of dimension 2n2+2n+1 with a k-basis:

    {(Jx)i(Jy)j,(Jx)i(Jy)jz,[(J1J)x]i,[(J2J)y]j,1+JJ1J2|1i,jn}.

    Now, we consider the second class of extensions of H2n2. The definition is given as follows.

    Definition 3.5. The associative algebra ˆH2n2 is generated by x,y, and z, with the relations

    xn+1=x, yn+1=y, xy=yx,zx=yz, zy=xz,z2=1nni,j=1qijxiyj.

    We define Δ:ˆH2n2ˆH2n2ˆH2n2, as follows

    Δ(x)=xx,Δ(y)=yy,Δ(z)=1nni,j=1qijxizyjz

    and extend it to ˆH2n2 in the natural way. As previously, let J1=xn,J2=yn and J=J1J2.

    Theorem 3.6. ˆH2n2 is a Δ-associative algebra.

    Proof. We should show that the Δ is an algebra homomorphism of ˆH2n2. In fact,

    (Δ(x))n+1=Δ(x),(Δ(y))n+1=Δ(y),Δ(x)Δ(y)=Δ(y)Δ(x),Δ(y)Δ(z)=(yy)(1nni,j=1qijxizyjz)=(1nni,j=1qijxizyjz)(xx)=Δ(z)Δ(x),Δ(x)Δ(z)=(xx)(1nni,j=1qijxizyjz)=(1nni,j=1qijxizyjz)(yy)=Δ(z)Δ(y).

    Furthermore, we have that

    (Δ(z))2=1nni,j=1qijΔ(x)iΔ(y)j

    by the proof of Theorem 3.3 a). Hence, the map Δ can be extended to a homomorphism of algebra in ˆH2n2. Also, Δ satisfies the coassociative axiom by the proof of Theorem 3.3 b). Hence, ˆH2n2 is a Δ-associative algebra.

    Remark 3.7. The Δ-associative algebra ˆH2n2 may be non-trivial. Indeed, if there exists a k-map ε:ˆH2n2k:

    ϵ(x)=a,ϵ(y)=b,ϵ(z)=c,a,b,ck,

    such that ε is a counit of ˆH2n2. Then, we have

    xε(x)=ε(x)x=x,yε(y)=ε(y)y=y,

    we get that a=b=1 and

    m(ϵid)Δ(z)=c(1nni,j=1qijyj)z=cJ2z=cynz,
    m(idϵ)Δ(z)=c(1nni,j=1qijxi)z=cJ1z=cxnz,

    where m is the multiplication of ˆH2n2.

    If there exists a ck such that cJ1z=cJ2z=z, then ˆH2n2 is trivial. Otherwise, ˆH2n2 is non-trivial.

    Here, we always assume that x, y, and z are freely without any additional conditions in the definition of ˆH2n2. Consequently, in this case, ˆH2n2 is a non-trivial Δ-associative algebra.

    Recall that in ˆH2n2, the set

    {J, J1J+J2J,1+JJ1J2}

    is a set of orthogonal central idempotents. Set

    ˆA0:=JˆH2n2,ˆA1:=(J1J+J2J)ˆH2n2,ˆA2:=(1+JJ1J2)ˆH2n2.

    Proposition 3.8. As an algebra, we have

    (1) ˆH2n2=ˆA0ˆA1ˆA2.

    (2) ˆA0H2n2.

    (3) ˆA1H2H2H2ncopies, where H2 is the Sweedler's algebra.

    (4) ˆA2k[h]/(h2).

    Proof. (1) It is obvious.

    (2) The proof is similar to Proposition 3.4 (2).

    (3) Let

    x1=(J1+J22J)x,y1=(J1+J22J)y,z1=(J1+J22J)z.

    Note that ˆA1 is a subalgebra of ˆH2n2 with the identity J1+J22J, which is generated by x1, y1, and z1 with the relations

    xn1=J1J,yn1=J2J,x1y1=y1x1=0,z21=0,z1x1=y1z1,x1z1=z1y1.

    Let

    λs=1nnt=1qstxt1,μs=1nnt=1qstyt1,

    for s=1,2,,n. One can see that

    {λ1,λ2,,λn,μ1,μ2,,μn}

    is a complete set of primitive orthogonal idempotents of ˆA1, and

    x1λs=λsx1=qsλs,y1λs=λsy1=0,z1λs=μsz1,x1μs=μsx1=0,y1μs=μsy1=qsμs,z1μs=λsz1.

    Thus, the bounded quiver of ˆA1 is as follows

    with the admissible ideal

    αiβi=βiαi=0,(i=1,2,,n).

    Therefore, the set

    {λ1+μ1,λ2+μ2,,λn+μn}

    forms a complete set of orthogonal central idempotents of ˆA1 and

    ˆA1=nl=1(λl+μl)ˆA1.

    In addition, each ideal (λl+μl)ˆA1 of ˆA1 is a 4-dimensional subalgebra of ˆA1 with the identity λl+μl. Also, it has the k-basis {λl,μl,λlz,μlz} for l{1,2,,n}.

    Let H2 be the 4-dimensional Sweedler's algebra generated by η,χ subjecting to the relations

    η2=1,χ2=0,ηχ=χη.

    For l=1,2,,n, let ϱl:(λl+μl)ˆA1H2 be the map given by

    ϱl(λl+μl)=1,ϱl(λl)=1+η2,
    ϱl(μl)=1η2,ϱl(λlz)=1+η2χ,ϱl(μlz)=1η2χ.

    Obviously, ϱl is an algebra isomorphism.

    (4) The ideal ˆA2 is a subalgebra of ˆH2n2 with the identity 1+JJ1J2. Since

    (1+JJ1J2)x=0,(1+JJ1J2)y=0,(1+JJ1J2)z0,

    it can be viewed as an algebra generated by (1+JJ1J2)z with the relation

    [(1+JJ1J2)z]2=0.

    This shows that ˆA2 is isomorphic to k[h]/(h2).

    The proof is completed.

    By Proposition 3.8, we see that ˆH2n2 is not a semi-simple algebra with the dimension 2n2+4n+2. Its radical has a k-basis as follows

    {λlz,μlz,(1+JJ1J2)z|l=1,2,,n}.

    In this section, the aim is to classify all indecomposable representations of the weak Hopf algebra ¯H2n2, and then to characterize its representation ring using generators with generating relations.

    The representation theory of semi-simple algebra H2n2 and k[g]/(gn1) was studied by many authors; see, for example, [15,30,31]. Let ind-R denote the set of all indecomposable representations of some ring R. By Proposition 3.4, we have

    ¯H2n2=¯A0¯A1¯A2¯A3

    as algebras, therefore,

    ind¯H2n2=ind¯A0ind¯A1ind¯A2ind¯A3.

    where ind-¯Ai can be viewed as the subset of ind-¯H2n2 in the natural way. We note that ¯A0H2n2, ¯A1¯A2k[g]/(gn1) and ¯A3k,

    Now, we list all irreducible representations of ¯Ai, i=1,2,3.

    (1) Let Mi,iZn. The 1-dimensional irreducible ¯H2n2-module with basis v(i), the actions of ¯H2n2 on Mi are

    x.v(i)=qiv(i),y.v(i)=0,z.v(i)=0.

    In fact,

    ind-¯A1={Mi|iZn}.

    (2) Let Ni,iZn. The 1-dimensional irreducible ¯H2n2-module with basis w(i), the actions of ¯H2n2 on Ni are

    x.w(i)=0,y.w(i)=qiw(i),z.w(i)=0.

    In fact,

    ind-¯A2={Ni|iZn}.

    (3) Let L be a 1-dimensional irreducible ¯H2n2-module with basis u, the actions of ¯H2n2 on L are

    x.u=0,y.u=0,z.u=0.

    In fact,

    ind-¯A3={L}.

    Therefore, we have the following results.

    Proposition 4.1. The set

    R={Sm, Si,j, Ms, Ns, LmZ2n,0i<jn1,sZn}

    forms a complete list of non-isomorphic irreducible representations of ¯H2n2.

    For the decomposition formulas of the tensor product of two irreducible representations of ¯H2n2, we have the following several lemmas.

    Lemma 4.2. For all iZn, mZ2n and 0s<tn1, we have

    (1) MiSmSmMiMm+i(modn);

    (2) MiSs,tSs,tMiMi+s(modn)Mi+t(modn);

    (3) NiSmSmNiNm+i(modn);

    (4) NiSs,tSs,tNiNi+s(modn)Ni+t(modn);

    (5) LSmSmLL;

    (6) LSs,tSs,tLLL.

    Proof. Suppose that Sm and Ss,t are irreducible ¯H2n2-modules with the basis {vm} and {vst1, vst2}, respectively, for all mZ2n and 0s<tn1.

    (1) Considering the decomposition formulas of tensor product MiSm for iZn. We have

    x.(v(i)vm)=qi+mv(i)vm,y.(v(i)vm)=0,z.(v(i)vm)=0.

    Hence, MiSmMi+m(modn). Similarly, one can show that SmMiMi+m(modn).

    (2) Considering the decomposition formulas of tensor product MiSs,t for all iZn. We have

    x.(v(i)vst1)=qi+sv(i)vst1,y.(v(i)vst1)=0,z.(v(i)vst1)=0,x.(v(i)vst2)=qi+tv(i)vst2,y.(v(i)vst2)=0,z.(v(i)vst2)=0.

    So, MiSs,tMi+s(modn)Mi+t(modn). Similarly, Ss,tMiMi+s(modn)Mi+t(modn) can be proven.

    The remaining cases can be proven in a similar way.

    Lemma 4.3. For all i,iZn, we have

    (1) MiMiMi+i(modn);

    (2) MiNiNiMiL;

    (3) MiLLMiL;

    (4) NiNiNi+i(modn);

    (5) NiLLNiL;

    (6) LLL.

    Proof. It is similar to the proof of Lemma 4.2.

    Let a=[S1], b=[Sn+1], c=[S0,1], d=[M0],e=[N0] in r(¯H2n2), we have the following lemma.

    Lemma 4.4. For all i,iZn, the following statements hold in r(¯H2n2).

    (1) If n is odd, then

    (a) [Mi]=bid, and bnd=d;

    (b) [Ni]=bie, and bne=e;

    (c) [L]=ed;

    (d) cd=d+bd, ce=e+be, d2=d and e2=e.

    (2) If n is even, then

    (a) [Mi]=aid, and and=d;

    (b) [Ni]=aie, and ane=e;

    (c) [L]=ed;

    (d) bd=ad, be=ae, cd=d+ad, ce=e+ae, d2=d and e2=e.

    Proof. (1) If n is odd, it can be concluded that bi=Si+(1)i+1+12n, for iZn and bn=Sn by [15, Lemma 3]. Hence, we get that

    M0Si+(1)i+1+12nSi+(1)i+1+12nM0Mi,M0SnSnM0M0,N0Si+(1)i+1+12nSi+(1)i+1+12nN0Ni,N0SnSnN0N0,M0S0,1S0,1M0M0M1,N0S0,1S0,1N0N0N1,

    by Lemma 4.2. One sees that [Mi]=bid, bnd=d, [Ni]=bie and bne=e. The statements (a) and (b) of (1) are followed.

    Moreover, based on statements (a) and (b), one can get that cd=d+bd, ce=e+be. Also,

    M0N0L,M0M0M0,N0N0N0

    by Lemma 4.3, then [L]=ed, d2=d and e2=e.

    The first claim follows.

    (2) If n is even, by [15, Lemma 4], it can be concluded that

    ai=Si, bi=Si+(1)i+1+12n, for iZn and bn=an=S0.

    The remaining statements can be proven in a similar way of claim (1).

    Corollary 4.5. For the representation ring r(¯H2n2), we have

    (1) if n is odd, the set

    {bk|0k2n1}{cibj|1in12,0jn1}{bld,ble,de|0ln1}

    forms a Z-basis of r(¯H2n2);

    (2) if n is even, the set

    {aibj|0in1,j=0,1}{cibj|1i<n2,0jn1}{cn2bj|0j<n2}{ald,ale,de|0ln1}

    forms a Z-basis of r(¯H2n2).

    Proof. (1) If n is odd, the set

    {bk|0k2n1}{cibj|1in12,0jn1}

    corresponds one-to-one to the set of irreducible ¯H2n2-modules

    {[Si]|0i2n1}{[Si,j]|0i<jn1}

    by [15, Corollary 2 (1)]. We have

    [Mi]=bid,bnd=d,[Ni]=bie,bne=e,[L]=ed

    by Lemma 4.4 (1).

    Similarly, the set {bld,ble,de|0ln1} corresponds one-to-one to the set of irreducible ¯H2n2-modules {[Ml],[Nl],[L]|0ln1}.

    (2) If n is even, the set

    {aibj|0in1,j=0,1}{cibj|1i<n2,0jn1}{cn2bj|0j<n2}

    corresponds one-to-one to the set of irreducible ¯H2n2-modules

    {[Si],[Si,j|0i2n1,0i<jn1}

    by [15, Corollary 2 (2)].

    On the other hand, we have

    [Mi]=aid,and=d,[Ni]=aie,ane=e,[L]=ed,

    by Lemma 4.4 (2). Hence, the set {ald,ale,de|0ln1} corresponds one-to-one to the set of irreducible ¯H2n2-modules {[Ml],[Nl],[L]|0ln1}.

    Accordingly, the proof is finished.

    Now, we can prove the main result of this section.

    Theorem 4.6. Assume that nN and n2, then the representation ring r(¯H2n2) is a commutative ring, which can be characterized by generators and relations as follows:

    (1) if n is odd, then r(¯H2n2)Z[y,z,α,β]/I1, where I1 is the ideal generated by the set

    {y2n1,zynz,Bm+1(y,z)ym+1Bm(y,z),ynαα,ynββ,zααyα,zββyβ,α2α,β2β.}m:=n12;

    (2) if n=2, then r(¯H2n2)Z[y,z,α,β]/I2, where I2 is the ideal generated by the set

    {y21,x2y2,zxzy,zzy,z2xyxy1,x2αα,x2ββ,yαxα,yβxβ,zααxα,zββxβ,,α2αβ2β};

    (3) if n>2 is even, then r(¯H2n2)Z[y,z,α,β]/I3, where I3 is the ideal generated by the set

    {xn1,x2y2,zxzy,Dm+1(y,z,x)ym+1Dm1(y,z,x),Dm(y,z,x)ymDm(y,z,x),xnαα,xnββ,yαxα,yβxβ,zααxα,zββxβ,α2α,β2β}m:=n2.

    Proof. It is easy to see that r(¯H2n2) is commutative.

    (1) By Corollary 4.5, if n is odd, then r(¯H2n2) is generated by b,c,d and e. Hence, there is a unique ring epimorphism

    Φ:Z[y,z,α,β]r(¯H2n2)

    such that

    Φ(y)=b=[Sn+1],Φ(z)=c=[S0,1],Φ(α)=d=[M0],Φ(β)=e=[N0].

    We note that

    b2n=1,cbn=c,Bm+1(b,c)=bm+1Bm(b,c),
    bnd=d,bne=e,cd=d+bd,ce=e+be,d2=d,e2=e,

    by Lemma 2.4, and Lemma 4.4, we have

    Φ(y2n1)=0,Φ(zynz)=0,Φ(Bm+1(y,z)ym+1Bm(y,z))=0,Φ(ynαα)=0,Φ(ynββ)=0,Φ(zααyα)=0,Φ(zββyβ)=0,Φ(α2α)=0,Φ(β2β)=0,

    where m:=n12. It follows that Φ(I1)=0, and Φ induces a ring epimorphism

    ¯Φ:Z[y,z,α,β]/I1r(¯H2n2)

    such that

    ¯Φ(¯ν)=Φ(ν),forallνZ[y,z,α,β],

    where ¯ν=π(ν), and π is a natural epimorphism Z[y,z,α,β]Z[y,z,α,β]/I1.

    We note that the ring r(¯H2n2) is the free Z-module of rank 4n+1+n(n1)2, with the Z-basis

    {bk|0k2n1}{cibj|1in12,0jn1}{bld,ble,de|0ln1}.

    So we can define a Z-module homomorphism:

    Ψ:r(¯H2n2)Z[y,z,α,β]/I1,
    Ψ(cibj)=¯zi¯yj,Ψ(bk)=¯yk,Ψ(bld)=¯yl¯α,
    Ψ(ble)=¯yl¯β,Ψ(de)=¯α¯β,

    where 0k2n1,1in12,0jn1 and 0ln1.

    On the other hand, as the Z-module, Z[y,z,α,β]/I1 is generated by elements

    {¯zi¯yj,¯yk,¯yl¯α,¯yl¯β,¯α¯β|0k2n1,1in12,0jn1,0ln1}.

    Let

    ¯a{¯zi¯yj,¯yk,¯yl¯α,¯yl¯β,¯α¯β|0k2n1,1in12,0jn1,0ln1},

    it is straightforward to check that Ψ¯Φ(¯a)=¯a. Hence Ψ¯Φ=id, which implies that ¯Φ is a monomorphism, and hence ¯Φ is an isomorphism.

    The proofs of the remaining statements of the theorem are similar to the above.

    In this section, the representations and representation ring of the Δ-associative algebra ˆH2n2 are given. Due to the similar research methods used as in Section 4, we directly provide relevant conclusions and omit their proofs.

    By Proposition 3.8, it follows that ˆH2n2=ˆA0ˆA1ˆA2 as algebras, then

    indˆH2n2=indˆA0indˆA1indˆA2.

    where ind-ˆAi can be viewed as the subset of ind-ˆH2n2 in the natural way. As an algebra, we note that

    ˆA0H2n2,ˆA1H2H2H2ncopies,ˆA2K[h]/(h2),

    where H2 is the 4-dimensional Sweedler's algebra.

    It is noted that the algebra H2 and k[y]/(yd) are Nakayama algebras, and their representations are also easy to construct; for examples, see [21,30,31]. Therefore, we can give a complete list of finite dimensional indecomposable representations in ind-ˆA1 and ind-ˆA2 as follows.

    (1) Let ˆMi,iZn. The 1-dimensional irreducible ˆH2n2-module with basis ˆv(i), the actions of ˆH2n2 on ˆMi are

    x.ˆv(i)=0,y.ˆv(i)=qiˆv(i),z.ˆv(i)=0.

    (2) Let ˆNi,iZn. The 1-dimensional irreducible ˆH2n2-module with basis ˆw(i), the actions of ˆH2n2 on ˆNi are

    x.ˆw(i)=qiˆw(i),y.ˆw(i)=0,z.ˆw(i)=0.

    (3) Let Li,iZn. The 2-dimensional indecomposable ˆH2n2-module with basis ˆv(i)1,ˆv(i)2, the actions of ˆH2n2 on Li are

    x.ˆv(i)1=qiv(i)1,y.ˆv(i)1=0,z.ˆv(i)1=ˆv(i)2,x.ˆv(i)2=0,y.ˆv(i)2=qiˆv(i)2,z.ˆv(i)2=0.

    (4) Let Pi,iZn. The 2-dimensional indecomposable ˆH2n2-module with basis ˆw(i)1,ˆw(i)2, the actions of ˆH2n2 on Pi are

    x.ˆw(i)1=0,y.ˆw(i)1=qiˆw(i)1,z.ˆw(i)1=ˆw(i)2,x.ˆw(i)2=qiˆw(i)2,y.ˆw(i)2=0,z.ˆw(i)2=0.

    In fact,

    ind-ˆA1={ˆMi,ˆNi,Li,Pi|iZn}.

    (5) Let Q be a 1-dimensional irreducible ˆH2n2-module with basis ˆu, the actions of ˆH2n2 on Q are

    x.ˆu=0,y.ˆu=0,z.ˆu=0.

    (6) Let R be the 2-dimensional indecomposable ˆH2n2-module with basis ˆu1,ˆu2, the actions of ˆH2n2 on R are

    x.ˆu1=0,y.ˆu1=0,z.ˆu1=ˆu2,x.ˆu2=0,y.ˆu2=0,z.ˆu2=0.

    In fact,

    ind-ˆA2={Q,R}.

    Therefore, we have

    Proposition 5.1. The set

    G={Sm,Si,j,ˆMs,ˆNs,Ls,Ps,Q,RmZ2n,0i<jn1,sZn}

    forms a complete list of non-isomorphic indecomposable ˆH2n2-modules.

    Also, we have several lemmas as follows. As an example, we only give the proof of the first one.

    Lemma 5.2. For all i,iZn, mZ2n and 0s<tn1, we have

    (1) LiSmˆMm+i(modn)ˆNm+i(modn);

    (2) SmLiLm+i(modn);

    (3) LiSs,tˆMi+s(modn)ˆMi+t(modn)ˆNi+s(modn)ˆNi+t(modn);

    (4) Ss,tLiLi+s(modn)Li+t(modn);

    (5) LiLiˆMi+i(modn)ˆNi+i(modn)QQ;

    (6) LiPiˆMi+i(modn)ˆNi+i(modn)QQ;

    (7) PiLiRˆMi+i(modn)ˆNi+i(modn);

    (8) LiQQLiQQ;

    (9) LiRRLiQQQQ.

    Proof. (1) In G, considering the tensor product LiSm for iZn, mZ2n, we have

    x.(ˆv(i)1vm)=qi+mˆv(i)1vm,y.(ˆv(i)1vm)=0,z.(ˆv(i)1vm)=0,x.(ˆv(i)2vm)=0,y.(ˆv(i)2vm)=qi+mˆv(i)2vm,z.(ˆv(i)2vm)=0.

    Hence, LiSmˆNi+m(modn)ˆMi+m(modn).

    (2) Considering the tensor product SmLi for iZn, mZ2n, we have

    x.(vmˆv(i)1)=qi+mvmˆv(i)1,x.(vmˆv(i)2)=0,y.(vmˆv(i)1)=0,y.(vmˆv(i)2)=qi+mvmˆv(i)2,z.(vmˆv(i)1)=1nσ(m)qm22ni,j=1qij+mi+ijvmˆv(i)2,z.(vmˆv(i)2)=0.

    Let π1=vmˆv(i)1,π2=1nσ(m)qm22ni,j=1qij+mi+ijvmˆv(i)2, then

    x.π1=qi+mπ1,y.π1=0,z.π1=π2,x.π2=0,y.π2=qi+mπ2,z.π2=0.

    Thus, SmLiLi+m(modn).

    (3) Considering the tensor product LiSs,t for iZn and 0s<tn1, we have

    x.(ˆv(i)1vst1)=qi+sˆv(i)1vst1,y.(ˆv(i)1vst1)=0,z.(ˆv(i)1vst1)=0,x.(ˆv(i)1vst2)=qi+tˆv(i)1vst2,y.(ˆv(i)1vst2)=0,z.(ˆv(i)1vst2)=0,x.(ˆv(i)2vst1)=0,y.(ˆv(i)2vst1)=qi+sˆv(i)2vst1,z.(ˆv(i)2vst1)=0,x.(ˆv(i)2vst2)=0,y.(ˆv(i)2vst2)=qi+tˆv(i)2vst2,z.(ˆv(i)2vst2)=0.

    Hence, LiSs,tˆMi+s(modn)ˆMi+t(modn)ˆNi+s(modn)ˆNi+t(modn).

    (4) Considering the tensor product Ss,tLi for all iZn and 0s<tn1, we have

    x.(vst1ˆv(i)1)=qi+svst1ˆv(i)1,y.(vst1ˆv(i)1)=0,z.(vst1ˆv(i)1)=1nni,j=1qij+it+tivst2ˆv(i)2,x.(vst2ˆv(i)1)=qi+tvst2ˆv(i)1,y.(vst2ˆv(i)1)=0,z.(vst1ˆv(i)1)=1nni,j=1qij+st+is+tivst1ˆv(i)2,x.(vst1ˆv(i)2)=0,y.(vst1ˆv(i)2)=qi+tvst1ˆv(i)2,z.(vst1ˆv(i)2)=0,x.(vst2ˆv(i)2)=0,y.(vst2ˆv(i)2)=qi+svst2ˆv(i)2,z.(vst2ˆv(i)2)=0.

    Let

    π3=vst1ˆv(i)1,π4=1nni,j=1qij+it+tivst2ˆv(i)2,π5=vst2ˆv(i)1,π6=1nni,j=1qij+st+is+tivst1ˆv(i)2,

    then

    x.π3=qi+sπ3,y.π3=0,z.π3=π4,x.π4=0,y.π4=qi+sπ4,z.π4=0,x.π5=qi+tπ5,y.π5=0,z.π5=π6,x.π6=0,y.π6=qi+tπ6,z.π6=0,

    it follows that Ss,tLiLi+s(modn)Li+t(modn).

    The remaining statements can be proven similarly.

    Lemma 5.3. For i,iZn, mZ2n and 0s<tn1, we have

    (1) ˆMiSmSmˆMiˆMm+i(modn);

    (2) ˆMiSs,tSs,tˆMiˆMi+s(modn)ˆMi+t(modn);

    (3) ˆMiˆMiˆMi+i(modn);

    (4) ˆMiNiNiˆMiQ;

    (5) ˆMiLiLiˆMiQˆMi+i(modn);

    (6) ˆMiPiPiˆMiQˆMi+i(modn);

    (7) ˆMiQQˆMiQ;

    (8) ˆMiRRˆMiQQ.

    Lemma 5.4. For all i,iZn, mZ2n and 0s<tn1, we have

    (1) ˆNiSmSmˆNiˆNm+i(modn);

    (2) ˆNiSs,tSs,tˆNiˆNi+s(modn)ˆNi+t(modn);

    (3) ˆNiˆNiˆNi+i(modn);

    (4) ˆNiLiLiˆNiˆNi+i(modn)Q;

    (5) ˆNiPiPiˆNiQˆNi+i(modn);

    (6) ˆNiQQˆNiQ;

    (7) ˆNiRRˆNiQQ.

    Lemma 5.5. For all i,iZn, mZ2n and 0s<tn1, we have

    (1) PiSmPm+i(modn);

    (2) SmPiˆMm+i(modn)ˆNm+i(modn);

    (3) PiSs,tˆPi+s(modn)ˆPi+t(modn);

    (4) Ss,tPiMi+s(modn)Mi+t(modn)Ni+s(modn)Ni+t(modn);

    (5) PiPiˆMi+i(modn)ˆNi+i(modn)QQ;

    (6) PiQQPiQQ;

    (7) PiRRPiQQQQ.

    Lemma 5.6. For all iZn, mZ2n and 0s<tn1, we have

    (1) QSmSmQQ;

    (2) QSs,tSs,tQQQ;

    (3) QQQ;

    (4) QRRQQQ;

    (5) RSmSmRQQ;

    (6) RSijSijRQQQQ;

    (7) RRQQQQ.

    Let

    a=[s1],b=[sn+1],c=[s0,1],d=[ˆM0],e=[ˆN0],f=[L0],g=[P0].

    Then, we have the following result.

    Lemma 5.7. For all i,i{1,2,,n1}, the following statements hold in r(ˆH2n2).

    (1) If n is odd, then

    (a) [ˆMi]=bid=dbi and [ˆM0]=bnd=dbn=d;

    (b) [ˆNi]=bie=ebi and [ˆN0]=bne=ebn=e;

    (c) [Li]=bif, and [L0]=bnf=f;

    (d) [Pi]=gbi, and [P0]=gbn=g;

    (e) [Q]=ed=de;

    (f) [R]=gfde;

    (g) fb=bg=b(d+e);

    (h) cd=dc=d+bd, ce=ec=e+be, cf=f+bf, fc=cg=d+e+bd+be and gc=g+gb;

    (i) d2=d, e2=e, fg=g2=f2=d+e+2ed, df=fd=dg=gd=d+ed, and ef=fe=eg=ge=e+ed.

    (2) If n is even, then

    (a) [ˆMi]=aid=dai and [ˆM0]=and=dan=d;

    (b) [ˆNi]=aie=eai and [ˆN0]=ane=ean=e;

    (c) [Li]=aif and [L0]=anf=f;

    (d) [Pi]=gai and [P0]=gan=g;

    (e) [Q]=ed=de;

    (f) [R]=gfde;

    (g) fa=ag=a(d+e);

    (h) bd=db=ad, be=eb=ae, bf=af, bg=fb=fa and gb=ga;

    (i) cd=dc=d+ad, ce=ec=e+ae, cf=f+af, fc=cg=d+e+ad+ae and gc=g+ga;

    (j) d2=d, e2=e, fg=g2=f2=d+e+2ed, df=fd=dg=gd=d+ed and ef=fe=eg=ge=e+ed.

    Corollary 5.8. (1) If n is odd, the set

    {bk|1k2n}{cibj|1in12,1jn}{bsd,bse,bsf,gbs,de,gfde|1sn}

    forms a Z-basis of r(ˆH2n2).

    (2) If n is even, the set

    {aibj|1in,j=1,n}{cibj|1i<n2,1jn}{cn2bj|1j<n2,j=n}{asd,ase,asf,gas,de,gfde|1sn}

    forms a Z-basis of r(ˆH2n2).

    Proof. By Lemmas 5.3–5.7, we see that [S0] is no longer the identity of r(ˆH2n2). By Lemma 2.3, [S0]=[S2n], and if n is odd, the set

    {bk|1k2n}{cibj|1in12,1jn}

    corresponds one-to-one to the set of irreducible ˆH2n2-modules

    {[Si]|1i2n}{[Si,j]|0i<jn1};

    if n is even, the set

    {aibj|1in,j=1,n}{cibj|1i<n2,1jn}{cn2bj|1j<n2,j=n}

    corresponds one-to-one to the set of irreducible ˆH2n2-modules

    {[Si],[Si,j]|1i2n,0i<jn1}.

    On the other hand, if n is odd, we have

    [ˆMs]=bsd=dbs,and[ˆM0]=bnd=dbn=d,[ˆNs]=bse=ebs,and[ˆN0]=bne=ebn=e,[Ls]=bsf,and[L0]=bnf=f,[Ps]=gbs,and[P0]=gbn=g,[Q]=ed=de,[R]=gfde

    by Lemma 5.7, where 1sn1. Hence, the set

    {bsd,bse,bsf,gbs,de,gfde|1sn}

    corresponds one-to-one to the set of indecomposable ˆH2n2-modules

    {[ˆMs],[ˆNs],[Ls],[Ps],[Q],[R]|0sn1}.

    The remaining can be proven in a similar ways.

    By Lemma 5.3–5.7, we see that r(ˆH2n2) is a noncommutative ring without an identity. Let r(ˆH2n2) be the ring with the identity extended from the ring r(ˆH2n2) in the natural way. For the general definition one can refer to Section 2. Therefore, r(ˆH2n2) is a ring with the identity (1,0) and

    r(ˆH2n2){(0,α)|αr(ˆH2n2)}r(ˆH2n2).

    Now, we can explicitly describe r(ˆH2n2) by generators with relations.

    Theorem 5.9. Assume that nN and n2, we have the following statements:

    (1) if n is odd, then r(ˆH2n2)Zy,z,α,β,γ,δ/J1, where I1 is the ideal generated by the following set

    T1={y2n+1y,yzzy,zynz,Bm+1(y,z)ym+1Bm(y,z),yααy;ynαα,yββy,ynββ,ynγγ,δynδ,αββα,γyyδ,yδy(α+β),zααz,αzαyα,zββz,βzβyβ,zγγyγ,γzzδ,zδαβyαyβ,δzδδy,α2α,β2β,γδδ2,γ2δ2,δ2αβ2αβ,αγγα,γααδ,αδδα,δαααβ,βγγβ,γββδ,βδδβ,δββαβ.}m:=n12;

    (2) if n=2, then r(ˆH2n2)Zx,y,z,α,β,γ,δ/J2, where I2 is the ideal generated by the following set

    T2={y3y,x3x,xyyx,xzzx,yzzy,x2y2,zxzy,zzy,z2xyxyy2,xααx,x2αα,xββx,x2ββ,x2γγ,δx2δ,αββα,γxx(α+β),xδγx,yααy,yαxα,yββy,yβxβ,yγxγ,yδγy,γyγx,δyδx,zααz,zααxα,zββz,zββxβ,zγγxγ,γzzδ,zδαβxαxβ,δzδδx,α2α,β2β,γδδ2,γ2δ2,δ2αβ2αβ,αγγα,γααδ,αδδα,δαααβ,βγγβ,γββδ,βδδβ,δββαβ.};

    (3) if n>2 is even, then r(ˆH2n2)Zx,y,z,α,β,γ,δ/J3, where I3 is the ideal generated by the following set

    T3={xn+1x,yn+1y,xyyx,xzzx,yzzy,x2y2,zxzy,Dm+1(y,z,x)ym+1Dm1(y,z,x),Dm+1(y,z,x)ymDm(y,z,x),xααx,xnαα,xββx,xnββ,xnγγ,δxnδ,αββα,γxx(α+β),xδγx,yααy,yαxα,yββy,yβxβ,yγxγ,yδγy,γyγx,δyδx,zααz,zααxα,zββz,zββxβ,zγγxγ,γzzδ,zδαβxαxβ,δzδδx,α2α,β2β,γδδ2,γ2δ2,δ2αβ2αβ,αγγα,γααδ,αδδα,δαααβ,βγγβ,γββδ,βδδβ,δββαβ.}m:=n2.

    Proof. It is easy to see that r(ˆH2n2) is a noncommutative ring.

    (1) Let π:Zy,z,α,β,γ,δZy,z,α,β,γ,δ/J1 be the natural epimorphisms and ¯a=π(a) for any aZy,z,α,β,γ,δ.

    By Lemma 5.7(1) and Corollary 5.8(1), if n is odd, r(ˆH2n2) is generated as a ring by b,c,d,e,f,g, and any element of r(ˆH2n2) can be written as

    k1+k2++k6>0ϱk1k2k6bk1ck2dk3ek4fk5gk6

    where ϱk1k2k6,k1,k2,,k6Z. Therefore,

    r(ˆH2n2)={(k,k1+k2++k6>0ϱk1k2k6bk1ck2dk3ek4fk5gk6)|ϱk1k2k6,k1,k2,,k6,kZ}.

    We defined a Z-map

    Φ:Zy,z,α,β,γ,δr(ˆH2n2)

    as

    Φ(1)=(1,0),Φ(y)=(0,b),Φ(z)=(0,c),Φ(α)=(0,d),
    Φ(β)=(0,e),Φ(γ)=(0,f),Φ(δ)=(0,g).

    It is easy to get that Φ can be extended to a ring epimorphism in the natural way. In fact, any υZy,z,α,β,γ,δ can be written as

    υ=k1Zy,z,α,β,γ,δ+k1+k2++k6>0ϱk1k2k6yk1zk2αk3βk4γk5δk6,

    where ϱk1k2k6,k1,k2,,k6,kZ. Thus,

    Φ(υ)=(k,k1+k2++k6>0ϱk1k2k6bk1ck2dk3ek4fk5gk6).

    By Lemma 5.7(1), one sees that Φ(ω)=(0,0) for all ωT1. Hence, Φ(J1)=(0,0) and Φ induces a unique ring epimorphism

    ¯Φ:Zy,z,α,β,γ,δ/J1r(ˆH2n2)

    with ¯Φ(¯ν)=Φ(ν) for all νZy,z,α,β,γ,δ.

    We note that the ring r(ˆH2n2) is the free Z-module of rank 6n+3+n(n1)2, with the Z-basis

    {(1,0)}{(0,bk)|1k2n}{(0,cibj)|1in12,1jn}{(0,bsd),(0,bse),(0,bsf),(0,gbs),(0,de),(0,gfde)|1sn}.

    So we can define a Z-module homomorphism:

    Ψ:r(ˆH2n2)Zy,z,α,β,γ,δ/J1

    by

    Ψ((1,0))=¯1,Ψ((0,bk))=¯yk,Ψ((0,cibj))=¯zi¯yj,Ψ((0,bsd)=¯ys¯α,
    Ψ((0,bse))=¯ys¯β,Ψ((0,bsf))=¯ys¯γ,Ψ((0,gbs))=¯δ¯ys,
    Ψ((0,de))=¯α¯β,Ψ((0,gfde))=¯γ¯δ¯α¯β,

    where 1k2n,1in12,1jn and 1sn. Obviously, as the Z-module, Zy,z,α,β,γ,δ/J1 is generated by the set

    S={¯1}{¯yk|1k2n}{¯zi¯yj|1in12,1jn}{¯ys¯α,¯ys¯β,¯ys¯γ,¯δ¯ys,¯α¯β,¯γ¯δ¯α¯β,|1sn}.

    For any ¯aS, we see that Ψ¯Φ(¯a)=¯a. Hence Ψ¯Φ=id, which implies that ¯Φ is a monomorphism, and hence ¯Φ is an isomorphism.

    The proofs of the remaining statements are similar.

    Remark 5.10. By Theorems 4.6 and 5.9, as the rings, it is easy to see that

    (1) if n is odd, then

    r(H2n2)r(¯H2n2)/α,βr(ˆH2n2)/α,β,γ,δ,y2n1;

    (2) if n is even, then

    r(H2n2)r(¯H2n2)/α,βr(ˆH2n2)/α,β,γ,δ,yn1,xn1.

    We have described the representation rings of two classes of Δ-associative algebras, ¯H2n2 and ˆH2n2, extended from Hopf algebra H2n2 of Kac-Paljutkin type. It may be interesting to consider the category of representations of these representation rings, as this could be helpful in understanding the invariants of the more general tensor categories. It is challenging to consider non-Hermitian linear systems over these rings, similar to non-Hermitian quaternion linear systems over the quaternion algebra. Readers are referred to related works, such as [32,33,34].

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

    This work is supported by National Natural Science Foundation of China (Grant No. 12201187, 12471038) and Natural Science Foundation of Henan Province (Grant No. 222300420156, 242300421385)

    The authors declare there is no conflicts of interest.



    [1] H. Chen, F. V. Oystaeyen, Y. Zhang, The Green rings of Taft algebras, Proc. Amer. Math. Soc., 142 (2014), 765–775. https://doi.org/10.1090/S0002-9939-2013-11823-X doi: 10.1090/S0002-9939-2013-11823-X
    [2] Z. Wang, L. Li, Y. Zhang, Green rings of pointed rank one Hopf algebras of nilpotent type, Algebr. Represent. Theory, 17 (2014), 1901–1924. https://doi.org/10.1007/s10468-014-9484-9 doi: 10.1007/s10468-014-9484-9
    [3] Z. Wang, L. Li, Y. Zhang, Green rings of pointed rank one Hopf algebras of non-nilpotent type, J. Algebra, 449 (2016), 108–137. https://doi.org/10.1016/j.jalgebra.2015.11.002 doi: 10.1016/j.jalgebra.2015.11.002
    [4] R. Suter, Modules over Uq(sl2), Commun. Math. Phys., 163 (1994), 359–393. https://doi.org/10.1007/bf02102012 doi: 10.1007/bf02102012
    [5] H. Kondo, Y. Saito, Indecomposable decomposition of tensor products of modules over the restricted quantum universal enveloping algebra associated to sl2, J. Algebra, 330 (2011), 103–129. https://doi.org/10.1016/j.jalgebra.2011.01.010 doi: 10.1016/j.jalgebra.2011.01.010
    [6] H. Chen, Finite-dimensional representations of a quantum double, J. Algebra, 251 (2002), 751–789. https://doi.org/10.1006/jabr.2002.9144 doi: 10.1006/jabr.2002.9144
    [7] H. Chen, H. S. E. Mohammed, H. Sun, Indecomposable decomposition of tensor products of modules over Drinfeld doubles of Taft algebras, J. Pure Appl. Algebra, 221 (2017), 2752–2790. https://doi.org/10.1016/j.jpaa.2017.01.007 doi: 10.1016/j.jpaa.2017.01.007
    [8] D. Su, S. Yang, Representation rings of small quantum groups ¯Uq(sl2), J. Math. Phys., 58 (2017), 091704. https://doi.org/10.1080/00927872.2019.1612412 doi: 10.1080/00927872.2019.1612412
    [9] H. Chen, The Green ring of Drinfeld double D(H4), Proc. Am. Math. Soc., 142 (2014), 765–775. https://doi.org/10.1007/s10468-013-9456-5 doi: 10.1007/s10468-013-9456-5
    [10] H. Sun, H. S. E. Mohammed, W. Lin, H. Chen, Green rings of Drinfeld doubles of Taft algebras, Commun. Algebra, 48 (2020), 3933–3947. https://doi.org/10.1080/00927872.2020.1752225 doi: 10.1080/00927872.2020.1752225
    [11] R. Yang, S. Yang, The Grothendieck rings of Wu-Liu-Ding algebras and their Casimir numbers (Ⅰ), J. Algebra Appl., 21 (2022), 2250178. https://doi.org/10.1142/S021949882250178X doi: 10.1142/S021949882250178X
    [12] R. Yang, S. Yang, The Grothendieck rings of Wu-Liu-Ding algebras and their Casimir numbers (Ⅱ), Commun. Algebra, 49 (2021), 2041–2073. https://doi.org/10.1080/00927872.2020.1862139 doi: 10.1080/00927872.2020.1862139
    [13] J. Wu, Note on the coradical filtration of D(m,d,ε), Commun. Algebra, 44 (2016), 4844–4850. https://doi.org/10.1080/00927872.2015.1113295 doi: 10.1080/00927872.2015.1113295
    [14] J. Wu, G. Liu, N. Ding, Classification of affine prime regular Hopf algebras of GK-dimension one, Adv. Math., 296 (2016), 1–54. https://doi.org/10.1016/j.aim.2016.03.037 doi: 10.1016/j.aim.2016.03.037
    [15] J. Chen, S. Yang, D. Wang, Grothendieck rings of a class of Hopf algebras of Kac-Paljutkin type, Front. Math. China, 16 (2021), 29–47. https://doi.org/10.1007/s11464-021-0893-x doi: 10.1007/s11464-021-0893-x
    [16] Y. Guo, S. Yang, The Grothendieck ring of Yetter-Drinfeld modules over a class of 2n2-dimensional Kac-Paljutkin Hopf algebras, Commun. Algebra, 51 (2023), 4517–4566. https://doi.org/10.1080/00927872.2023.2213340 doi: 10.1080/00927872.2023.2213340
    [17] H. Sun, H. Chen, Y. Zhang, Representations of the small quasi-quantum group, preprint, arXiv: 2305.06083. https://doi.org/10.48550/arXiv.2305.06083
    [18] C. Cibils, The projective class ring of basic and split Hopf algebras, K-Theory, 17 (1999), 385–393. https://doi.org/10.1023/A:1007703709638 doi: 10.1023/A:1007703709638
    [19] M. Herschend, On the representation ring of a quiver, Algebra Represent. Theory, 12 (2009), 513–541. https://doi.org/10.1007/s10468-008-9118-1 doi: 10.1007/s10468-008-9118-1
    [20] Y. Li, N. Hu, The Green rings of the 2-rank Taft algebras and its two relatives twisted, J. Algebra, 410 (2014), 1–35. https://doi.org/10.1016/j.jalgebra.2014.04.006 doi: 10.1016/j.jalgebra.2014.04.006
    [21] L. Li, Y. Zhang, The Green rings of the Generalized Taft algebras, Contemp. Math, 585 (2013), 275–288. https://doi.org/10.1090/conm/585/11618 doi: 10.1090/conm/585/11618
    [22] D. Su, S. Yang, Green rings of weak Hopf algebras based on generalized Taft algebra, Period. Math. Hung., 76 (2018), 229–242. https://doi.org/10.1007/s10998-017-0221-0 doi: 10.1007/s10998-017-0221-0
    [23] Z. Wang, L. You, H. Chen, Representations of Hopf-Ore extensions of group algebras and pointed Hopf algebras of rank one, Algebra Represent. Theory, 18 (2015), 801–830. https://doi.org/10.1007/s10468-015-9517-z doi: 10.1007/s10468-015-9517-z
    [24] Q. Xie, J. Sun, Non-weight modules over N=1 Lie superalgebras of Block type, Forum Math., 35 (2023), 1279–1300. https://doi.org/10.1515/forum-2022-0267 doi: 10.1515/forum-2022-0267
    [25] R. Yang, S. Yang, Representations of a non-pointed Hopf algebra, AIMS Math., 6 (2021), 10523–10539. https://doi.org/10.3934/math.2021611 doi: 10.3934/math.2021611
    [26] D. Su, S. Yang, Green rings of Δ-associative algebras, J. Beijing Univ. Technol., 43 (2017), 1275–1282. https://doi.org/10.11936/bjutxb2016070014 doi: 10.11936/bjutxb2016070014
    [27] D. Pansera, A class of semisimple Hopf algebras acting on quantum polynomial algebras, in Rings, Modules and Codes, 727 (2019), 303–316. https://doi.org/10.1090/conm/727/14643
    [28] F. Li, Weak Hopf algebras and some new solutions of the quantum Yang-Baxter equation, J. Algebra, 208 (1998), 72–100. https://doi.org/10.1006/jabr.1998.7491 doi: 10.1006/jabr.1998.7491
    [29] M. E. Sweedler, Hopf Algebras, W. A. Benjamin, New York, 1969.
    [30] M. Auslander, I. Reiten, S. O. Smalϕ, Representation Theory of Artin Algebras, University Press, Cambridge, 1995. https://doi.org/10.1017/CBO9780511623608
    [31] I. Assem, D. Simson, A. Skowroński, Elements of the Representation Theory of Associative Algebras: Volume 1: Techniques of Representation Theory, Cambridge University Press, Cambridge, 2006. https://doi.org/10.1017/CBO97805116192 12
    [32] T. Li, Q. Wang, Structure preserving quaternion full orthogonalization method with applications, Numer. Linear Algebra Appl., 30 (2023), e2495. https://doi.org/10.1002/nla.2495 doi: 10.1002/nla.2495
    [33] T. Li, Q. Wang, Structure preserving quaternion biconjugate gradient method, SIAM J. Matrix Anal. Appl., 45 (2024), 306–326. https://doi.org/10.1137/23M1547299 doi: 10.1137/23M1547299
    [34] X. Zhang, W. Ding, T. Li. Tensor form of GPBiCG algorithm for solving the generalized Sylvester quaternion tensor equations, J. Franklin Inst., 360 (2023), 5929–5946. https://doi.org/10.1016/j.jfranklin.2023.04.009 doi: 10.1016/j.jfranklin.2023.04.009
  • This article has been cited by:

    1. Huaqing Gong, Shilin Yang, The representation ring of a non-pointed bialgebra, 2025, 10, 2473-6988, 5110, 10.3934/math.2025234
  • Reader Comments
  • © 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(714) PDF downloads(48) Cited by(1)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog