We give the relationship between indecomposable modules over the finite dimensional k-algebra A and the smash product A♯G respectively, where G is a finite abelian group satisfying G⊆Aut(A), and k is an algebraically closed field with the characteristic not dividing the order of G. More precisely, we construct all indecomposable A♯G-modules from indecomposable A-modules and prove that an A♯G-module is indecomposable if and only if it is an indecomposable G-stable module over A. Besides, we give the relationship between simple, projective and injective modules in modA and those in modA♯G.
Citation: Wanwan Jia, Fang Li. Invariant properties of modules under smash products from finite dimensional algebras[J]. AIMS Mathematics, 2023, 8(3): 6737-6748. doi: 10.3934/math.2023342
[1] | Jinlei Dong, Fang Li, Longgang Sun . Derived equivalence, recollements under H-Galois extensions. AIMS Mathematics, 2023, 8(2): 3210-3225. doi: 10.3934/math.2023165 |
[2] | Yang Zhang, Jizhu Nan . A note on the degree bounds of the invariant ring. AIMS Mathematics, 2024, 9(5): 10869-10881. doi: 10.3934/math.2024530 |
[3] | Ruifang Yang, Shilin Yang . Representations of a non-pointed Hopf algebra. AIMS Mathematics, 2021, 6(10): 10523-10539. doi: 10.3934/math.2021611 |
[4] | Yaguo Guo, Shilin Yang . Projective class rings of a kind of category of Yetter-Drinfeld modules. AIMS Mathematics, 2023, 8(5): 10997-11014. doi: 10.3934/math.2023557 |
[5] | Jaruwat Rodbanjong, Athipat Thamrongthanyalak . Characterizations of modules definable in o-minimal structures. AIMS Mathematics, 2023, 8(6): 13088-13095. doi: 10.3934/math.2023660 |
[6] | Yupei Zhang, Yongzhi Luan . Dimension formulas of the highest weight exceptional Lie algebra-modules. AIMS Mathematics, 2024, 9(4): 10010-10030. doi: 10.3934/math.2024490 |
[7] | Wanxia Wang, Shilin Yang . On finite-dimensional irreducible modules for the universal Askey-Wilson algebra. AIMS Mathematics, 2023, 8(8): 18930-18947. doi: 10.3934/math.2023964 |
[8] | Yunpeng Xue . On enhanced general linear groups: nilpotent orbits and support variety for Weyl module. AIMS Mathematics, 2023, 8(7): 14997-15007. doi: 10.3934/math.2023765 |
[9] | Xianguo Hu . Universal enveloping Hom-algebras of regular Hom-Poisson algebras. AIMS Mathematics, 2022, 7(4): 5712-5727. doi: 10.3934/math.2022316 |
[10] | Haicun Wen, Mian-Tao Liu, Yu-Zhe Liu . The counting formula for indecomposable modules over string algebra. AIMS Mathematics, 2024, 9(9): 24977-24988. doi: 10.3934/math.20241217 |
We give the relationship between indecomposable modules over the finite dimensional k-algebra A and the smash product A♯G respectively, where G is a finite abelian group satisfying G⊆Aut(A), and k is an algebraically closed field with the characteristic not dividing the order of G. More precisely, we construct all indecomposable A♯G-modules from indecomposable A-modules and prove that an A♯G-module is indecomposable if and only if it is an indecomposable G-stable module over A. Besides, we give the relationship between simple, projective and injective modules in modA and those in modA♯G.
Let A be a finite dimensional algebra over an algebraically closed field k with characteristic p, and G be an arbitrary finite group each element of which acts as an algebra automorphism on A. Then we have the skew group algebra of A by G, denoted by A♯G. The representation theory of skew group algebras has been widely studied [2,3,11,12]. It is well known that the smash product A♯G retains many properties from A when the order of G is invertible in k. For example, A is of finite representation type (1-Gorenstein, selfinjective, of finite global dimension) if and only if A♯G is.
Since the algebras A and A♯G have a lot of properties in common, it is of interest to study the relationship between modules over A and A♯G and consider whether A-modules can induce A♯G-modules and if an A-module can induce A♯G-modules, how to describe all non-isomorphic classes of such induced A♯G-modules.
In [13], the authors give the relationship under the assumption A=kQ is a path algebra and G is a cyclic group. In [7], the authors discuss the relationship for algebra A=kQ/I and G is a finite abelian subgroup of automorphism group of bound quiver (Q,I). In this paper, we investigate the relationship between indecomposable modules over the finite dimensional k-algebra A and the smash product A♯G respectively, where G is a finite abelian group satisfying G⊆Aut(A), and k is an algebraically closed field with the characteristic not dividing the order of G. We prove that an A♯G-module is indecomposable if and only if it is indecomposable G-stable module and describe all the A♯G-module structures from the same G-stable module.
It should be noted that if algebra A is not basic, it is no longer isomorphic to a quotient of the path algebra. In general, the representations of a finite dimensional (non-basic) algebra is characterized via its corresponding basic algebra. In this paper, we show that the representations of skew group algebra of a finite dimensional non-basic algebra can be induced directly not through the representations of its corresponding basic algebra.
The article is organized as follows. In the next section, we introduce the basic notations in the context of smash products and G-stable modules. We devote Section 3 to induced A♯G-modules from G-stable modules. Section 4 focuses on the construction of all indecomposable A♯HM-modules from an indecomposable A-module M with maximal stable subgroup HM of G and describes the number of non-isomorphic indecomposable induced A♯HM-modules from an indecomposable A-module M. Section 5 states the main theorem which constructs all indecomposable A♯G-modules from an indecomposable A-module and gives the number of non-isomorphic indecomposable induced A♯G-modules from an indecomposable A-module M. Then the relationship between simple, projective and injective modules in modA and those in modA♯G is discussed. In the last section, we give the relation between the stable category of a path algebra and the corresponding smash products to be abelian.
We fix an algebraically closed field k. Let Q=(Q0,Q1) be a finite quiver given by the vertex set Q0 and the arrow set Q1. For an arrow a in a quiver, we write s(a) and t(a) for its source and target respectively. Arrows in quivers are composed as functions, that is if ab is a path then s(a)=t(b). The path algebra kQ is the algebra generated by all paths (including those of length zero) of Q, with multiplication induced by composition of paths. In this paper, algebras are assumed to be associative unital finite dimensional k-algebras and all modules are always unital and finitely generated.
We introduce the definition of skew group algebras as smash products, which was well-known in the theory of Hopf algebras.
Definition 2.1. Let G be a finite group acting on an algebra A over a field by automorphisms. The smash product A♯G of A by G is the algebra defined by:
(i) its underlying vector space is A⊗kkG;
(ii) multiplication is given by
(a⊗g)(b⊗h)=ag(b)⊗gh |
for a,b∈A and g,h∈G, extended by linearity and distributivity.
Usually, we also call A♯G the skew group algebra of A by G.
From now on, algebra A is finite dimensional and G⊆Aut(A) is a finite group with order n. Let k be an algebraically closed field whose characteristic does not divide the order of G. In this paper, We will deal with the smash product A♯G.
For M∈modA and g∈G, we define an A-module gM by taking the same underlying vector space as M with the new module action
a⋅m=g−1(a)m. |
Let ϕ:M→N be a module homomorphism, and set gϕ=ϕ as a linear map. Then gϕ:gM→gN can be viewed as a homomorphism of modules under the new module action. Indeed,
ϕ(a⋅m)=ϕ(g−1(a)m)=g−1(a)ϕ(m)=a⋅ϕ(m). |
Definition 2.2. An A-module M is G-stable module if gM≅M for any g∈G.
We say that an A-module M is an indecomposable G-stable module if it is not isomorphic to the proper direct sum of two G-stable modules. Let HM={g∈G|gM≅M}. Then HM is a subgroup of G. We call HM maximal stable subgroup of G for M. Denote RM={g1,g2,⋯,gs} a complete set of left coset representatives of HM in G. By the Krull-Remak-Schmidt theorem for modules over finite dimensional k-algebras, we have the following lemma
Lemma 2.1. [7] With the above notations, any indecomposable G-stable module M is precisely the representation of the form
M≅⨁g∈RNgN=g1N⊕g2N⊕⋯⊕gsN, |
where N is an indecomposable A-module. Moreover, the Krull-Remak-Schmidt theorem holds for G-stable modules.
Proof. First, define f:h(gM)→hgM such that f(m)=m for all m∈M, g,h∈G. Since
f(a⋅m)=f(h−1(a)⋅m)=f(g−1h−1(a)⋅m)=f((hg)−1(a)m)=a⋅f(m), |
we have f is an A-module automorphism. Therefore, h(gM)≅hgM and ⨁g∈RNgN is an indecomposable G-stable module.
Let M be an indecomposable G-stable module. Then gM≅M for any g∈G. If N is a summand of M as A-module, we have the isomorphism classes of {gN|g∈G} are summands of M as A-module. Therefore M≅⨁g∈RNgN.
Let X be a G-stable module. Then gX≅X as A-module for any g∈G. So we can write
X≅M1⊕M2⊕⋯⊕Mt, |
where each Mi is of the form
Mi≅⨁g∈RNigNi=gi1Ni⊕gi2Ni⊕⋯⊕gisNi |
with Ni an indecomposable A-module, RNi={gi1,gi2,⋯,gis} a complete set of left coset representatives of HNi in G. The lemma follows from the Krull-Remak-Schmidt theorem for A-modules.
In this section, the conclusions can be found in [5,6,7,13] when A is a finite dimensional path algebra or a path algebra with relation. Here we give their proofs by the similar method when A is a finite dimensional algebra.
Lemma 3.1. Every module M of the smash product A♯G is a G-stable A-module.
Proof. Define f:gM→M such that f(m)=g(m) for all m∈M, g∈G. It is well-defined since M is an A♯G module. Then we have that for any a∈A,
f(a⋅m)=g(a⋅m)=g(g−1(a)m)=(a♯g)m=a(g(m))=af(m). |
Then f is an A-module homomorphism. It is easy to check that f is an isomorphism if we define f−1:M→gM such that f−1(m)=g−1(m) for all m∈M. Therefore, M≅gM as A-modules which means M is a G-stable A-module.
We will show that any G-stable A module induces an A♯G-module.
Proposition 3.1. [4] Let M be a G-stable A-module. Then for any g∈G there exists an isomorphism ϕg:gM→M such that ϕng=ϕggϕg⋯gn−1ϕg is the identity.
Proof. Let M be a G-stable A-module and u:gM→M be an isomorphism.
If w=ugu⋯gn−1u is identity, then we get the required isomorphism ϕg=u.
If w=ugu⋯gn−1u is not identity, we can find some isomorphism y:M→M such that ϕg=yu satisfying that ϕng=ϕggϕg⋯gn−1ϕg is the identity.
For this purpose, we note that gw=u−1wu, hence that g(wm)=u−1wmu and gψ=u−1ψu for all m∈N and all ψ∈k[w]⊂End(M).
Suppose y∈k[w] and set ϕg=yu. Then by induction
ϕng=ϕggϕg⋯gn−1ϕg=yugygu⋯gn−1ygn−1u=yuu−1yugu⋯gn−1ygn−1u=y2ugugu−1u−1yugug2u⋯gn−1ygn−1u=y3ugug2u⋯gn−1ygn−1u=⋯=ynugu⋯gn−1u=ynw. |
Since n does not divide the characteristic of k, the equation ynw=1 has a solution y=n√w−1 in k[w]. Therefore ϕg=yu is the required isomorphism.
Theorem 3.1. Let M be a G-stable A-module. Then M has an induced A♯G-module structure.
Proof. We define g(m)=ϕg(m) for any g∈G, m∈M. First, by Proposition 3.1, gi(m)=ϕig(m) and ϕg1ϕg2=ϕg1g2. Hence, g is well-defined. Since for a♯g1,b♯g2∈A♯G, m∈M,
(a♯g1)((b♯g2)(m))=(a♯g1)(bϕg2(m))=aϕg1(bϕg2(m))=aϕg1(g1(b)⋅ϕg2(m)))=ag1(b)(ϕg1ϕg2(m))=ag1(b)(ϕg1g2(m))=(ag1(b)♯g1g2)(m)=((a♯g1)(b♯g2))(m), |
we have that M has an induced A♯G-module structure.
The main purpose of this section is to construct all indecomposable A♯HM-modules from an indecomposable A-module M with maximal stable subgroup HM of G and give the number of non-isomorphic indecomposable induced A♯HM-modules from an indecomposable A-module M.
Here are two lemmas which will be used.
Lemma 4.1. [12] Let M,N be indecomposable A-modules, and G⊆Aut(A) be a finite subgroup of the k-automorphism group of A with order n. Then:
(i) (A♯G)⊗AM≅⨁g∈GgM as A-modules;
(ii) (A♯G)⊗AM≅(A♯G)⊗AN as A♯G-modules if and only if N≅gM for some g∈G;
(iii) The number of summands in the decomposition of (A♯G)⊗AM into a direct sum of indecomposables is at most the order of HM, where HM= {g∈G,gM≅M}.
Lemma 4.2. [8,9] Let H be a finite dimensional semisimple Hopf algebra and A be a finite dimensional H-module algebra. Then, for any A♯H -module M, it holds that M is a direct summand of A♯H⊗AM as an A♯H -module.
For an indecomposable A-module M, denote HM={g∈G,gM≅M} and RM={g1,g2,⋯,gs} a complete set of left coset representatives of HM in G.
In the following discussion, we assume HM is abelian. In particular, we can assume G is an abelian group. Since kHM is semisimple, we have
kHM≅r=n/s⨁i=1Hi | (4.1) |
as kHM-modules, where Hi is one dimensional irreducible HM-representation for i=1,2,⋯,r.
By Theorem 3.1, the HM-stable A-module M induces an A♯HM-module structure. Then Hi⊗kM has an A♯HM-module structure if we define
(a♯g)(1⊗m)=g(1)⊗a♯g(m) |
for any i∈{1,2,⋯r}, a♯g∈A♯HM,1⊗m∈Hi⊗kM.
Lemma 4.3. With the above notations and assumpution, we have
(i) Hi⊗kM≅M as A-modules for i=1,2,⋯,r;
(ii) Hi⊗kM is an indecomposable A♯HM-module for any i=1,2,⋯,r;
(iii) Hi⊗kM≇Hj⊗kM as A♯HM-modules if i≠j.
Proof. (ⅰ) Define f:M→Hi⊗kM such that f(m)=1⊗m for all m∈M. Then f is bijective. Since f(a(m))=1⊗a(m)=a(1⊗m)=af(m) for all a∈A,m∈M, then f is an A-module isomorphism.
(ⅱ) Since for any a♯g1,b♯g2∈A♯G, m∈M,
(1♯1)(1⊗m)=1(1)⊗1♯1(m)=1⊗m |
((a♯g1)(b♯g2))(1⊗m)=(ag1(b)♯g1g2)(1⊗m)=g1g2(1)⊗(ag1(b)♯g1g2)(m)=g1g2(1)⊗((a♯g1)(b♯g2))(m)=(a♯g1)(g2(1)⊗(b♯g2)(m)=(a♯g1)((b♯g2)(1⊗m)), |
we have Hi⊗kM is an A♯HM-module. By (ⅰ), it is an indecomposable A-module. Then Hi⊗kM is an indecomposable A♯HM-module for any i=1,2,⋯,r.
(ⅲ) Before proving (ⅲ), we claim that HomA(M,Hi⊗kM)≅Hi⊗kEndA(M) as A♯HM-modules for any i=1,2,⋯,r.
First, HomA(M,Hi⊗kM) has an A♯HM-module structure with
(a♯l)(f)(m)=(a♯l)f(m),foranyf∈HomA(M,Hi⊗kM),a♯l∈A♯HM,m∈M. |
EndA(M) has an A♯HM-module structure with
(a♯l)(f)(m)=(a♯l)f(m),foranyf∈EndA(M),a♯l∈A♯HM,m∈M. |
Hi⊗kEndA(M) has an A♯HM-module structure with
(a♯l)(1⊗f)=l(1)⊗(a♯l)f,foranyf∈EndA(M),a♯l∈A♯HM. |
Now we define Φ:HomA(M,Hi⊗kM)→Hi⊗kEndA(M) by Φ(f)=1⊗¯f for any f∈HomA(M,Hi⊗kM), where ¯f is defined by ¯f(m)=hfmf if f(m)=hf⊗mf for any m∈M.
Since for any a∈A,m∈M,
f(am)=af(m)=a(hf⊗mf)=hf⊗a(mf)=1⊗a(hfmf), |
we have
¯f(am)=hfa(mf)=a(hfmf). |
Thus ¯f∈EndA(M) and Φ is well-defined.
Since for any a∈A,f∈HomA(M,Hi⊗kM),
(af)(m)=a(f(m))=a(hf⊗mf)=hf⊗a(mf), |
we have
Φ(af)=1⊗¯af=1⊗a¯f=a(1⊗¯f)=aΦ(f). |
Thus Φ is an A-module homomorphism.
Since
(gf)(m)=g(f(m))=g(hf⊗mf), |
we have
Φ(gf)=1⊗¯gf=g(1⊗¯f)=gΦ(f). |
It means that Φ is an A♯HM-module homomorphism.
Since Φ is injective and
dimkHomA(M,Hi⊗kM)=dimkHi⊗kEndA(M), |
we have the homomorphism Φ is an A♯HM-module isomorphism.
Now we prove (ⅲ). Assume Hi⊗kM≅Hj⊗kM as A♯HM-modules for some i≠j. Then Hi⊗kEndA(M)≅Hj⊗kEndA(M) by the claim. Since EndA(M) is local, we have EndA(M)/radEndA(M)≅k as algebras. Then
Hi⊗kEndA(M)/radEndA(M)≅Hj⊗kEndA(M)/radEndA(M), |
which means Hi≅Hj as kHM-modules. We get a contradiction to i≠j.
Theorem 4.1. Let M be an indecomposable A-module M with HM={g∈G,gM≅M}. Suppose HM is abelian and RM={g1,g2,⋯,gs} is a complete set of left coset representatives of HM in G. Hi is defined in (4.1). Then we have
(i) A♯HM⊗AM≅r⨁i=1Hi⊗kM as A♯HM-modules.
(ii) For any A♯HM-module N, if M≅N as A-modules, then there exists a unique i∈{1,2⋯,r} such that N≅Hi⊗kM as A♯HM-modules.
Proof. (ⅰ) By Lemma 4.2, Hi⊗kM is a direct summand of A♯HM⊗A(Hi⊗kM) for any i∈{1,2⋯,r}. By Lemma 4.3(ⅰ), A♯HM⊗A(Hi⊗kM)≅A♯HM⊗AM. Then we have Hi⊗kM is a direct summand of A♯HM⊗AM. By Lemma 4.3(ⅲ), if i≠j, Hi⊗kM≇Hj⊗kM, then by Krull-Remak-Schmidt theorem, r⨁i=1Hi⊗kM is a direct summand of A♯HM⊗AM. And by Lemma 4.1(ⅲ), A♯HM⊗AM has at most r summands. It shows that
A♯HM⊗AM≅r⨁i=1Hi⊗kM |
as A♯HM-modules.
(ⅱ) For an A♯HM-module N, if N≅M as A-modules, then N is an indecomposable A♯HM-module. By Lemma 4.2, N is a direct summand of A♯HM⊗AN≅A♯HM⊗AM. By (ⅰ), Lemma 4.3 and the Krull-Remak-Schmidt theorem, it is easy to see that there exists a unique i∈{1,2⋯,r} such that N≅Hi⊗kM.
In this section, we give the main results which construct all induced indecomposable A♯G-modules from an indecomposable A-module and give the number of non-isomorphic indecomposable induced A♯G-modules from an indecomposable A-module M.
Lemma 5.1. Let G be a finite group with order n and M be an indecomposable A-module with HM={g∈G,gM≅M}. Suppose HM is abelian and RM={g1,g2,⋯,gs} is a complete set of left coset representatives of HM in G. Hi is defined in (4.1). Then
(i) A♯G⊗A♯HM(Hi⊗kM)≅⨁g∈RMgM as A-modules;
(ii) A♯G⊗A♯HM(Hi⊗kM) is an indecomposable A♯G -module;
(iii) A♯G⊗A♯HM(Hi⊗kM)≇A♯G⊗A♯HM(Hj⊗kM) as A♯G-modules, if i≠j;
(iv) A♯G⊗AM≅r⨁i=1A♯G⊗A♯HM(Hi⊗kM).
Proof. (ⅰ) Define f:gM→g⊗M such that f(m)=g⊗m for any g∈G,m∈M. Then f is bijection. Since for any a∈A,m∈A,
f(a⋅m)=f(g−1(a)m)=g⊗g−1(a)m=g(g−1(a))⊗m=a(g⊗m)=af(m), |
f is an A-module isomorphism. Therefore, gM≅g⊗M as A-modules.
Since A♯G⊗A♯HM(Hi⊗kM)≅⨁g∈RMg⊗Hi⊗kM as A♯HM-modules and Hi⊗kM≅M as A-modules, we have A♯G⊗A♯HM(Hi⊗kM)≅⨁g∈RMg⊗M≅⨁g∈RMgM as A-modules.
(ⅱ) By (ⅰ) and Lemma 2.1, A♯G⊗A♯HM(Hi⊗kM) is an indecomposable G-stable A-module. By Lemma 4.1, A♯G⊗A♯HM(Hi⊗kM) is an indecomposable A♯G-module.
(ⅲ) If i≠j, A♯G⊗A♯HM(Hi⊗kM)≅A♯G⊗A♯HM(Hj⊗kM) as A♯G-modules, by
A♯G⊗A♯HM(Hi⊗kM)≅⨁g∈RMg⊗Hi⊗kM |
as A♯HM-modules, we have h⊗Hi⊗kM is a direct summand of A♯G⊗A♯HM(Hj⊗kM)≅⨁g∈RMg⊗Hj⊗kM, where h∈RM.
If h⊗Hi⊗kM≅h⊗Hj⊗kM, then Hi⊗kM≅Hj⊗kM as A♯HM-modules. By Lemma 4.3(ⅲ), It is a contradiction.
If h⊗Hi⊗kM≅g⊗Hj⊗kM for some h≠g∈RM, by g⊗Hj⊗M≅gM, we have hM≅gM. It is also a contradiction.
(ⅳ) By Lemma 4.2, A♯G⊗A♯HM(Hi⊗kM) is a direct summand of
A♯G⊗AA♯G⊗A♯HM(Hi⊗kM). |
By (ⅰ),
A♯G⊗A♯HM(Hi⊗kM)≅⨁g∈RMgM |
is a direct summand of A♯G⊗A(⨁g∈RMgM). By (ⅱ) and Lemma 4.1(ⅱ), A♯G⊗A♯HM(Hi⊗kM) is a direct summand of A♯G⊗AM. By (ⅲ) and the Krull-Schmidt theorem, we have
r⨁i=1A♯G⊗A♯HM(Hi⊗kM) |
is a direct summand of A♯G⊗AM. By Lemma 4.1(ⅲ), A♯G⊗AM has at most r indecomposable summands. Therefore
A♯G⊗AM≅r⨁i=1A♯G⊗A♯HM(Hi⊗kM). |
Next, we construct all induced indecomposable A♯G-modules from an indecomposable A-module and give the number of non-isomorphic indecomposable A♯G-modules induced from the corresponding G-stable A-module.
Theorem 5.1. Let G be a finite group with order n and M be an indecomposable A-module with maximal stable subgroup HM={g∈G,gM≅M}. Suppose HM is abelian and RM={g1,g2,⋯,gs} is a complete set of left coset representatives of HM in G. Hi is defined in (4.1). Then for any A♯G-module N, if N≅⨁g∈RMgM as A-modules, there exists a unique i∈{1,2⋯,r} such that N≅A♯G⊗A♯HM(Hi⊗kM) as A♯G-modules. That is, there are r non-isomorphic indecomposable A♯G-modules induced from the same indecomposable G-stable A -module.
Proof. For any A♯G-module N, if N≅⨁g∈RMgM as A-modules, then by Lemma 4.2, N is a direct summand of
A♯G⊗AN≅A♯G⊗A⨁g∈RMgM. |
By Lemma 4.1(ⅱ), Lemma 5.1(ⅳ) and the Krull-Schmidt theorem, there exists a unique i∈{1,2⋯,r} such that N≅A♯G⊗A♯HM(Hi⊗kM). That is, there are r non-isomorphic indecomposable A♯G-modules induced from the same indecomposable G-stable A -module.
Theorem 5.2. Suppose G⊆Aut(A) is an abelian group. Any indecomposable A♯G-module is an indecomposable G-stable A-module. Conversely, for any indecomposable G-stable A-module, the corresponding canonical induced A♯G-module is indecomposable.
Proof. First, for any indecomposable A♯G-module M, by Lemma 2.1,
M≅t⨁j=1MjMj≅⨁g∈RNjgNj |
with Nj an indecomposable A-module, RNj={gj1,gj2,⋯,gjs} a complete set of left coset representatives of HNj in G. By Lemma 4.2, we have M is direct summand of
A♯G⊗AM≅t⨁j=1⨁g∈RNjA♯G⊗AgNj. |
Then by Lemma 4.1(ⅱ) and the Krull-Schmidt theorem, there exists j such that M is a direct summand of A♯G⊗ANj. Therefore, by Theorem 5.1, we have
M≅⨁g∈RNjgNj |
as A-modules. By Lemma 2.1, M is an indecomposable G-stable A-module.
Conversely, by Lemma 4.1 it is obvious that for any indecomposable G-stable A-module, the corresponding canonical induced A♯G-module is indecomposable.
According to Theorems 5.1 and 5.2, for a skew group algebra A♯G where A is a finite dimensional algebra and G⊆Aut(A) is abelian, all finite dimensional A♯G-modules can be obtained from G-stable modules. The number of non-isomorphic indecomposable A♯G-modules induced from the same G-stable A-module can be given. In this case, for any indecomposable A-module M, the G-stable A-module ⨁g∈RMgM has r non-isomorphic A♯G-module structures, where r=|HM|=n/s.
We give the relation between simple, projective and injective modules in modA and those in modA♯G.
Theorem 5.3. Suppose G⊆Aut(A) is an abelian group. Let M be an A♯G-module. Then
(i) M is simple if and only if there exists a simple A-module S such that M≅⨁g∈RSgS.
(ii) M is projective if and only if there exists an indecomposable projective A-module P such that M≅⨁g∈RPgP.
(iii) M is injective if and only if there exists an indecomposable injective A-module I such that M≅⨁g∈RIgI.
Proof. According to Theorems 5.1 and 5.2, we need only to prove that M is a simple (projective, injective) A♯G-module if and only if M is a simple (projective, injective) A-module.
(ⅰ) Assume M≅⨁g∈RSgS for some simple A-module S. If it is not simple, we have for its proper submodule N, N≅⨁g∈EgS, where E is a proper subset of RS. By Lemma 4.1, N is G-stable. It is a contradiction.
(ⅱ) By [8], M is an indecomposable projective A♯G -module if and only if M is an indecomposable projective A -module.
By duality, we get (ⅲ).
Let Q be a connected finite quiver without oriented cycles and σ∈Aut(Q) with order n. I. Reiten and Chrisine Riedtmann in [12] constructed the dual quiver with automorphism (˜Q,˜σ), where ˜Q is the Ext-quiver of KQ♯⟨σ⟩ and ˜σ is the automorphism of k˜Q induced from an admissible automorphism. Fix a primitive n-th root of unity ζ, we give the definition of dual quiver.
Definition 6.1. [12] Let G=⟨σ⟩ and E be a set of representatives of the G-orbits of vertices of Q. The dual quiver ˜Q is described as follows: Each vertex of Q in E whose G-orbit has size s gives rise to n/s vertices of ˜Q. Let ε and ε′ lie in E and have orbits of size s and s′, and let α:ε→σ−t(ε′) be an arrow of Q. If η0,η1,⋯,ηm−1 and η′0,η′1,⋯,η′m′−1 are the vertices of ˜Q arising from Gε and Gε′, there is an arrow from ημ to η′μ′ in ˜Q if and only if μ≡μ′+a module(m,m′), where a is defined by
σ[s,s′](α)=ζ[s,s′]aα. |
Then we have the following proposition.
Proposition 6.1. [12] the skew-group algebra KQ♯⟨σ⟩ is Morita equivalent to the path algebra k˜Q of dual quiver ˜Q.
Let A be a k-algebra. The corresponding stable module category has modules as objects, while its morphisms are equivalence classes of modulo homomorphisms factoring through projectives. The category of left A-modules is denoted by modA, and the corresponding stable category is denoted by mod_A with the set of morphisms denoted by Hom_(M,N) for any A-modules M and N. The stable category is additive.
Definition 6.2 [1] An A-module monomorphism u:M→N in mod A is minimal if every nonzero submodule X of N has a nonzero intersection with Imu. A monomorphism u:M→E in mod A is called an injective envolope of M if E is an injective module and u is a minimal monomorphism.
In this section we give the relationship between the stable category of a path algebra and corresponding smash product to be abelian.
Theorem 6.1. [10] Let A=KQ be a hereditary algebra. The stable category mod_A of left A-modules is abelian if and only if the injective envelope of AA is projective.
Theorem 6.2. [12] The hereditary algebra KQ is 1-Gorenstein if and only KQ♯⟨σ⟩ is.
Then we have the following theorem
Theorem 6.3. The stable category of a path algebra KQ is abelian if and only if the stable category of skew group algebra KQ♯⟨σ⟩ is.
Proof. Let I(A) denote the injective envelope of algebra A.
"if": Since the smash product kQ♯⟨σ⟩ is Morita to k˜Q, by the Theorem 6.1, I(kQ♯⟨σ⟩) is projective, which means that kQ♯⟨σ⟩ is 1-Gorenstein. By the Theorem 6.2, the hereditary algebra KQ is 1-Gorenstein, which means that I(kQ) is projective, . Hence the stable category of a path algebra KQ is abelian by Theorem 6.1.
"only if": By the Theorem 6.1, the stable category mod_kQ of left KQ-modules is abelian if and only if the injective envelope of I(kQ) is projective, which means that kQ is 1-Gorenstein. By the Theorem 6.2, kQ♯⟨σ⟩ is 1-Gorenstein, which means that I(kQ♯⟨σ⟩) is projective. Hence the stable category of a path algebra kQ♯⟨σ⟩ is abelian by Theorem 6.1.
This project is supported by the National Natural Science Foundation of China (No. 12131015, No. 12071422).
The authors declare that there is no conflict of interests regarding the publication of this paper.
[1] | I. Assem, D. Simson, A. Skowronski, Elements of the representation theory of associative algebras, Cambridge: Cambridge University Press, 2006. http://dx.doi.org/10.1017/CBO9780511614309 |
[2] | M. Auslander, I. Reiten, S. Smalo, Representation theory of Artin algebras, Cambridge: Cambridge University Press, 1995. http://dx.doi.org/10.1017/CBO9780511623608 |
[3] | O. Funes, M. Redondo, Skew group algebras of simply connected algebras, Ann. Sci. Math. Quebec, 26 (2002), 171–180. |
[4] | P. Gabriel, The universal cover of a representation-finite algebra, In: Representations of algebras, Berlin: Springer-Verlag, 1981, 68–105. http://dx.doi.org/10.1007/BFb0092986 |
[5] | A. Hubery, Representation of quivers respecting a quiver automorphism and a theorem of Kac, Ph. D Thesis, University of Leeds, 2002. |
[6] |
A. Hubery, Quiver representations respecting a quiver automorphism: a generalisation of a theorem of Kac, J. Lond. Math. Soc., 69 (2004), 79–96. http://dx.doi.org/10.1112/S0024610703004988 doi: 10.1112/S0024610703004988
![]() |
[7] |
B. Hou, S. Yang, Skew group algebras of deformed preprojective algebras, J. Algebra, 332 (2011), 209–228. http://dx.doi.org/10.1016/j.jalgebra.2011.02.007 doi: 10.1016/j.jalgebra.2011.02.007
![]() |
[8] | G. Liu, Classification of finite dimensional basic Hopf algebras and related topics, Ph. D Thesis, Zhejiang University, 2005. |
[9] |
F. Li, M. Zhang, Invariant properties of representations under cleft extensions, Sci. China Ser. A, 50 (2007), 121–131. http://dx.doi.org/10.1007/s11425-007-2026-8 doi: 10.1007/s11425-007-2026-8
![]() |
[10] |
A. Martsinkovsky, D. Zangurashvili, The stable category of a left hereditary ring, J. Pure Appl. Algebra, 219 (2015), 4061–4089. http://dx.doi.org/10.1016/j.jpaa.2015.02.007 doi: 10.1016/j.jpaa.2015.02.007
![]() |
[11] | R. Martínez-Villa, Skew group algebras and their Yoneda algebras, Math. J. Okayama Univ., 43 (2001), 1–16. |
[12] |
I. Reiten, C. Riedtmann, Skew group algebras in the representation theory of Artin algebras, J. Algebra, 92 (1985), 224–282. http://dx.doi.org/10.1016/0021-8693(85)90156-5 doi: 10.1016/0021-8693(85)90156-5
![]() |
[13] |
M. Zhang, F. Li, Representations of skew group algebras induced from isomorphically invariant modules over path algebras, J. Algebra, 321 (2009), 567–581. http://dx.doi.org/10.1016/j.jalgebra.2008.09.035 doi: 10.1016/j.jalgebra.2008.09.035
![]() |