arxiv: v1 [math.co] 21 Dec 2018

Transcription

1 HOPF ALGEBRA STRUCTURE ON GRAPH SHOUJUN XU, XIAOMENG WANG, AND XING GAO arxiv: v1 [math.co] 21 Dec 2018 Abstract. In this paper, we construct a bialgebraic and further a Hopf algebraic structure on top of subgraphs of a given graph. Further, we give the dual structure of this Hopf algebraic structure. We study the algebra morphisms induced by graph homomorphisms, and obtain a functor from a graph category to an algebra category. Contents 1. Introduction 1 2. Bialgebra on subgraphs of a given graph An algebraic structure on subgraphs of a graph A coproduct on subgraphs of a graph A combinatorial description of the coproduct G A coalgebraic structure on subgraphs of a given graph A bialgebra structure on subgraphs of a given graph 9 3. Hopf algebra on subgraphs of a graph Hopf algebraic structure on subgraphs of a graph The dual Hopf algebra Algebra morphisms induced from graph homomorphisms 13 References Introduction The Hopf algebraic theory has a long history and broad applications in mathematics and physics [3, 9, 15, 23, 25, 28, 29, 31, 33]. The Hopf algebra of rooted forests originated from the work of A. Dür [6] and had a very rich theory. For example, A. Connes and D. Kreimer introduced the Connes-Kreimer Hopf algebra of rooted forests in [3, 24] to study the renormalization of perturbative quantum field theory. In [3, 4, 5], they introduced the commutative Hopf algebras of graphs to study the combinatorial structure of renormalization. It is also related to the Loday- Ronco Hopf algebra [27] and Grossman-Larson Hopf algebra [14] of rooted trees. L. Foissy has done a lot of work on the Hopf algebra of rooted forests [9, 10, 11, 12]. As related results, various infinitesimal bialgebras on docorated rooted forests have been established in [17, 18, 32], via different Hochschild 1-cocycle conditions. However, the Hopf algebra on a given graph does not have similar results with the Hopf algebra of rooted trees. A totally assigned graph is a graph with a total order on its edges. In [7], G. E. Duchamp et al. built a Hopf algebra of totally assigned graphs. L. Foissy [13] constructed a Hopf algebra on Date: December 24, Mathematics Subject Classification. 16W99, 16T05, 16T10, 16T30. Key words and phrases. Hopf algebra, graph category, algebra category, covariant functor. 1

2 2 SHOUJUN XU, XIAOMENG WANG, AND XING GAO all graphs to insert chromatic polynomial into the theory of combinatorial Hopf algebra, and gave the new proof of some classical results. The concept of graph algebras was introduced by Schmitt [30]. Namely, a graph algebra is a commutative, cocommutative, graded, connected Hopf algebra, whose basis elements correspond to finite graphs. Humpert and Martin [21] obtained a new nonrecursive formula for the antipode of a graph algebra. In [26], the author introduced the 4-bialgebra of graphs that satisfies some relations. The study of Hopf algebras of graphs is extensive, but the research on Hopf algebras of subgraphs has no result yet. Subgraph plays an important role in the study of graph theory. For example, we can study the prefect graph by the independence number and clique number of the induced subgraph. We can determine if a graph is an interval map by the induced subgraph and many more. In the present paper, we construct a Hopf algebraic structure on subgraphs of a given graph. In particular, a combinatorial description of the coproduct is given. Narrowing ourself to the algebraic part of the aforementioned Hopf algebra on subgraphs of a given graph, we obtain an algebra morphism from a graph homomorphism. Using the language of categories, we obtain a covariant functor from the category of graphs to the category of algebras. Here is the structure of the paper. In Section 2, we first review the definitions of free monoid and free module. Then we proceed to give an algebraic structure on subgraphs of a graph (Lemma 2.7). Next, we construct a coalgebraic structure on subgraphs of a graph (Lemma 2.17), thereby we construct a bialgebra on the subgraph (Theorem 2.21). Finally, to make the coproduct G more explicit, we describe it in a combinatorial method (Equation (5)). In Section 3, continuing the line in Section 2, a Hopf algebra structure on subgraphs of graph is given (Theorem 3.5). Further, we study the dual Hopf algebra structure. We conclude with a description of the dual Hopf algebra(lemmas 3.7, 3.9). This section is also devoted to an algebra morphism indeuced by a graph homomorphism (Theorem 3.12). As an application, we obtain a functor from graph category to algebra category (Theorem 3.15). Notation. Graphs considered in this paper are connected and undirected graphs without multiple edges and loops. We will be working over a unitary commutative base ring k. By an algebra we mean a unitary associative k-algebra and by a coalgebra we mean a counitary coassociative k-coalgebra, unless otherwise stated. Linear maps and tensor products are taken over k. Denote by [n] the set{1, 2,,n}. For an algebra A, we view A Aas an A-bimodule via (1) a (b c) := ab c and (b c) a := b ca. 2. Bialgebraonsubgraphsofagivengraph In this section, we build a bialgebraic structure on top of subgraphs of a given graph An algebraic structure on subgraphs of a graph. Let us first review some notations on graphs and algebras, which are used throughout the remainder of the paper. For a graph G, denote by V(G) and E(G) the vertex set and edge set of G, respectively. Definition 2.1. [8] Let G be a graph. (a) If there is a path between any two vertices of G, then G is connected. (b) A subgraph of a graph G is simply a graph, all of whose vertices belong to V(G) and all of whose edges belong to E(G). (c) For S V(G), the induced subgraph G[S ] in G by S is the graph whose edge set consists of all of the edges in G that have both endpoints in S.

3 HOPF ALGEBRA STRUCTURE ON GRAPH 3 The concepts of free monoid and free module are needed later. Definition 2.2. [19] (a) A semigroup is a nonempty set S together with a binary operation : S S S which is associative: (x y) z= x (y z) for all x, y, z S. A semigroup S is called a monoid if it contains an element 1 with the property that x 1=1 x= x for all x S. (b) Let (S, S, 1 S ) and (T, T, 1 T ) be monoids. A mapφ : S T is called a monoids morphism if φ(x S y)=φ(x) Tφ(y) and φ(1 S )=1 T for all x, y S. (c) The free semigroup on a set X is a semigroup S together with a map i : X S with the property that, for any semigroup T and mapφ : X T, there exists a unique morphism ψ : S T such thatψ i=φ. Adjoining an identity 1 to S, the free monoid on X is obtained. Definition 2.3. [22] (a) Let k be a ring. A (left) k-module is an additive abelian group M together with a function k M M, (k, x) kx such that for all k, l k and x, y M: (i) k(x+y)=kx+ky. (ii) (k+l)x=kx+lx. (iii) k(lx) = (kl)x. (b) Let M and N be modules over ring k. A map f : M Nis a k-module morphism if f (x+y)= f (x)+ f (y) and f (kx)=k f (x) for all x, y M and k k. (c) The free k-module on a set X is a k-module F together with a set map i : X Fwith the property that, for any k-module M and map f : X M, there exists a unique k-module morphism f : F M such that f i= f. A k-module M is free if and only if it has a k-linear basis [22]. In the rest of this paper, let G be a graph andgthe set of nonempty connected subgraphs of G. Let M(G)= G be the free monoid ongin which the multiplication is the concatenation, denoted by m G and usually suppressed. The unit in M(G) is the empty graph, denote by ½. An element F in M(G) is a noncommutative product of connected subgraph ing: F=Γ 1 Γ n, where n 0 andγ 1,,Γ n G. Here we employ the convention that F= ½ whenever n=0. We define bre(f) := n to be the breath of F. For example, bre(½)=0, bre( 1)=1, bre( )=4. Denote by H(G) := km(g) the free k-module spanned by M(G). Then H(G) is closed under the multiplication m G. Example 2.4. Here are some examples of H(G). (a) If G= 1, theng={ 1 }, M(G)= 1 and H(G)=k 1. (b) If G= 1 2, then G={ 1, 2, 1 2 }, M(G)= 1, 2, 1 2 and H(G)=k 1, 2, 1 2.

4 4 SHOUJUN XU, XIAOMENG WANG, AND XING GAO 1 (c) If G= 4, then 2 3 G={ 1, 2, 3, 4, 2 1, 3 4, 1 4, 2 3, , 2 1 3, 2 4, , 2 3 4, , , , }, M(G)= 1, 2, 3, 4, 2 1, 3 4, 1 4, 2 3, , 2 1 3, 2 4, , 2 3 4, , , , , H(G)=k 1, 2, 3, 4, 2 1, 3 4, 1 4, 2 3, , 2 3, 2 4, , 2 3 4, , , , The following is the concepts of tensor products and algebras. Definition 2.5. [2, p. 24] (a) Let M, N, P be three k-modules. A mapping f : M N P is said to be k-bilinear if for each x M the mapping y f (x, y) of N into P is k-linear, and for each y Nthe mapping x f (x, y) of M into P is k-linear. (b) Let M, N be k-modules. Then there exists a pair (T, g) consisting of an k-module T and a k-bilinear mapping g : M N T, with the following property: for any k-module P and any k-bilinear mapping f : M N P, there exists a unique k-linear mapping f : T P such that f=f g. The module T is called the tensor product of M and N, and is denoted by M N. Definition 2.6. [28, Definition 1.1] An algebra (A, m, u) over k is a k-module A together with morphisms of k-modules m : A A A, called the multiplication, and u : k A, called the unit, such that the diagrams A A A m id A A A and id A m A A m A u id A id k A A u A A A k m α l α r A are commutative. Hereα l andα r are the isomorphisms defined by α l : k A A, k x kx, α r : A k A, x k kx, k k, x A. m The empty graph ½ can also be viewed as a linear map ½ : k H(G) given by 1 k ½. Lemma 2.7. The triple (H(G), m G, ½) is an algebra. Proof. It follow from F 1 (F 2 F 3 )=(F 1 F 2 )F 3 for F 1, F 2, F 3 H(G) and ½F= F= F½ for F H(G). This completes the proof.

5 HOPF ALGEBRA STRUCTURE ON GRAPH A coproduct on subgraphs of a graph. In this subsection, we construct a coproduct on subgraphs of a graph. Let us first recall Definition 2.8. [28, Definition 2.1] A coalgebra (C,,ε) over k is a k-module C together with morphisms of k-modules := C C C, called the coproduct, andε := C k, called the counit, such that the diagrams C C C id C C C id C C C C and ε id C id C ε k C C C C k β l β r C are commutative. Hereβ l andβ r are the isomorphisms defined by β l : C k C, x 1 k x, β r : C C k, x x 1 k, x C. Definition 2.9. [1, p. 54] Let (C,,ε) be a coalgebra. A k-linear subspace D of a coalgebra C is called a k-subcoalgebra of C. If D is a coalgebra with the restriction of andε. Now we define a coproduct G on H(G). By linearity, we only need to define G (F) for basis elements F M(G). If bre(f)=0, then F= ½ and define (2) G (F) := G (½) := ½ ½. If bre(f)=1, then F Gand we define (3) G (F) := V 1 V 2 =V(F) F[V 1 ] F[V 2 ], where F[V i ] are induced subgarphs of F by vertex sets V i with i=1, 2. Note that F[V 1 ], F[V 2 ] M(G) and G (F) H(G) H(G). If bre(f)>1, then F=Γ 1 Γ n for some n 2andΓ 1,...,Γ n G, and define (4) G (F) := G (Γ 1 Γ n ) := G (Γ 1 ) G (Γ n ). Again note that G (F) H(G) H(G) by G (Γ i ) H(G) H(G) for each i=1,, n. Let us give some examples for better insight into G. Example Let G= 1 2. Then 1 2 and are in M(G), and G ( 1 2 )= , G ( )= G ( 1 2 ) G ( 1 2 ) =

6 6 SHOUJUN XU, XIAOMENG WANG, AND XING GAO A combinatorial description of the coproduct G. This subsection is devoted to a combinatorial description of the coproduct G given in Subsection 2.2. Let G be a graph and F=Γ 1 Γ n M(G) withγ 1,,Γ n G. Note that V(F) is a multiset. For example, F= and V(F)={1, 2, 1, 2}. Now we define (5) G (F) := F[U] F[V], U V=V(F) where U V means that U and V can be expressed as U= U 1 U n and V= V 1 V n, with U i V i = V(Γ i ), i=1,, n. Further, the coproduct G can be extended to H(G)=kM(G) by linearity. Let us compute some examples for better understanding of Eq. (5). Example LetΓ= If F= 1 2 is a subgraph ofγ, then V(F)={1, 2} ={1} {2}={2} {1}= {1, 2}, and so G ( 1 2)= 1 2 ½+½ If F= 3 4 is a subgraph ofγ, then and so Note G ( 3 4 V(F)={3, 4} ={3} {4}={4} {3}= {3, 4}, G ( 3 4 )= 3 4 ½+½ )= ½ Example ConsiderΓ= Then and ½ = G ( 3 4 ) G ( 1 2). G 1 2 )= 1 2 ½+½ , ( 5) G 2 = 1 25 ½+½ , 1 ( G 1 2 ) 1 25 = ½ ½

7 HOPF ALGEBRA STRUCTURE ON GRAPH = G ( ( 1 2 5) ) G 2. 1 Now we prove that the combinatorial definition of G in Eq. (5) is the same as the one given in Subsection 2.2. For this, it suffices to show that G in Eq. (5) satisfies Eqs. (2) (4). For the cases of Eqs. (2) and (3), it follows directly from the definition of G in Eq. (5). For the case of Eq. (4), we have Lemma Let G be given in Eq. (5) andγ 1,,Γ n Gwith n 2. Then G (Γ 1 Γ n )= G (Γ 1 ) G (Γ n ). Proof. We have G (Γ 1 Γ n )= (Γ 1 Γ n )[U] (Γ 1 Γ n )[V] (by Eq. (5)) = = = U V=V(Γ 1 Γ n ) U i V i = V(Γ i ) i=1,, n U i V i = V(Γ i ) i=1,, n U 1 V 1 =V(Γ 1 ) (Γ 1 Γ n )[U 1 U n ] (Γ 1 Γ n )[V 1 V n ] Γ 1 [U 1 ] Γ n [U n ] Γ 1 [V 1 ] Γ n [V n ] Γ 1 [U 1 ] Γ 1 [V 1 ] = G (Γ 1 ) G (Γ n ) (by Eq. (5)). This completes the proof. Γ n [U n ] Γ n [V n ] U n V n =V(Γ n ) So we conclude Proposition The G given in Eq. (5) coincides with the G given in Subsection 2.2. Remark The coproduct G is cocommutative by the Definition A coalgebraic structure on subgraphs of a given graph. In this subsection, we obtain a coalgebra structure on subgraphs of a given graph G. The following result shows that H(G) is closed under the coproduct G. Lemma Let G be a graph and F H(G). Then G (F) H(G) H(G). Proof. By linearity, it is suffices to consider basis elements F M(G). Then by Eq. (5), we have G (F)= F[U] F[V]. U V=V(F) Then F[U], F[V] H(G) by definition and so G (F) H(G) H(G).

8 8 SHOUJUN XU, XIAOMENG WANG, AND XING GAO Define a linear mapε G : H(G) kby taking { 1k, if F= ½, (6) ε G (F) := 0, if F ½. Lemma Let G be a graph. Then the triple (H(G), G,ε G ) is a coalgebra. Proof. We first show the coassociativity (id G ) G (F)=( G id) G (F) for F H(G). By linearity, it suffices to consider basis elements F M(G). We have (id G ) G (F)= (id G ) F[U] F[V] (by Eq. (5)) U V=V(F) = F[U] G (F[V]) U V=V(F) = F[U] (F[V])[V ] (F[V])[V ] (by Eq. (5)) U V=V(F) V V =V(F[V]) = F[U] F[V ] F[V ] U V=V(F) V V =V(F[V]) = F[U] F[V ] F[V ] (by V(F[V])=V) U V V =V(F) = (F[U V ])[U] (F[U V ])[V ] F[V ] U V V =V(F) = F[W][U] F[W][V ] F[V ] W V =V(F) U V =W = F[W][U] F[W][V ] F[V ] (by V(F[W])=W) W V =V(F) U V =V(F[W]) = G (F[W]) F[V ] (by U V = W) W V =V(F) = ( G id) F[W] F[V ] W V =V(F) = ( G id) G (F). Thus G is coassociative. Next we prove the counicity ofε G. For F M(G), (ε G id) G (F)= (ε G id) F[U] F[V] (by Eq. (5)) U V=V(F) = (ε G id) ½ F+ F ½+ F[U] F[V] U V= V(F) U, V V(F)

9 HOPF ALGEBRA STRUCTURE ON GRAPH 9 =ε G (½) F+ε G (F) ½+ = 1 k F= β l (F) (by Eq. (6)). U V= V(F) U, V V(F) ε G (F[U]) F[V] With the same argument, we can show (id ε G ) G =β r. Hereβ l andβ r are the isomorphisms defined in Definition 2.8. This completes the proof. Corollary Let G be a graph and G be a subgraph of graph G. Then the triple (H(G ), G,ε G ) is a subcoalgebra of coalgebra (H(G), G,ε G ). Proof. It follows from the Lemma Corollary Let G be a graph. Then (H( ), G,ε G ) is a simple subcoalgebra, for any V(G). Proof. The triple (H( ), G,ε G ) is a subcoalgebra by Corollary Let (H, G,ε G ) be a subgraph of (H( ), G,ε G ) and H 0, HH( ). Then for any F H, we have G (F)= F[U] F[V]= F[V(F)\ ]+ F[U] F[V]H H. U V=V(F) Our proof is complete. U V= V(F) U 2.5. A bialgebra structure on subgraphs of a given graph. Combing the algebraic structure obtained in Subsection 2.1 and the coalgebraic structure obtained in Subsection 2.4, we build a bialgebraic structure on top of subgraphs of a given graph. Let us review the concept of bialgebras. Definition (a) [16, p. 49] Let (H, m H, u H ) and (L, m L, u L ) be two algebras. A map f : H L is called an algebra morphism if m L ( f f )= f m H and u L = f u H. (b) [16, p. 51] A bialgebra is a quintuple (H, m, u,,ε), where (H, m, u) is an algebra and (H,,ε) is a coalgebra such that : H H H andε : H k are morphisms of algebras. Now, we arrive at our main result in this subsection. Theorem Let G be a graph. Then the quintuple (H(G), m G, ½, G,ε G ) is a bialgebra. Proof. The triple (H(G), m G, ½) is an algebra by Lemma 2.7, and the triple (H(G), G,ε G ) is a coalgebra by Lemma Further, the coproduct G is an algebra morphism by Eq. (4). We are left to prove thatε G is an algebra morphism. For this, let F 1, F 2 M(G). Then by Eq. (6) { 1k, if F ε G (F 1 F 2 )= 1 and F 2 are empty graphs, 0, others, and { 1k, if F ε G (F 1 )ε G (F 2 )= 1 and F 2 are empty graphs, 0, others. Thus,ε G (F 1 F 2 )=ε G (F 1 )ε G (F 2 ) andε G is an algebra morphism. This completes the proof.

10 10 SHOUJUN XU, XIAOMENG WANG, AND XING GAO 3. Hopfalgebraonsubgraphsofagraph In this section, we construct a Hopf algebra structure on subgraphs of a given graph G Hopf algebraic structure on subgraphs of a graph. Let us review some concepts needed later. Definition 3.1. [1, p. 61] Let (C,,ε) be a coalgebra and (A, m, u) an algebra. Hom(C, A), f g := m ( f g) is said to be the convolution of f and g. For f, g Definition 3.2. [1, p. 61] Let (H, m, u,,ε) be a bialgebra. Then Hom(H, H) is a k-algebra with structure maps m Hom(H,H) ( f g)= f g and u Hom(H,H) (k)=ku ε. When the identity map id H of H is a regular element of Hom(H, H) with respect to multiplication on Hom(H, H), the inverse S of id H is called the antipode of H. The antipode S satisfies the following condition: S id H = id H S= u ε. A bialgebra with an antipode is called a Hopf algebra. Definition 3.3. [16, Definition 2.3.1] A bialgebra (H, m, u,, ε) is called a graded bialgebra if there are k-submodules H (n), n 0, of H such that (a) H= H (n) ; n=0 (b) H (p) H (q) H (p+q), p, q 0; (c) (H (n) ) H (p) H (q), n 0. p+q=n Elements of H (n) are said to have degree n. H is called connected if H (0) = k and kerε= n 1 H (n). Lemma 3.4. [3, 12, 16] Any connected graded bialgebra H is a Hopf algebra. The antipode S is given by: S (x)= (e id H ) k (x). It is also defined by S (1 H )=1 H and recursively by any of the two formulas: S (x)= x S (x )x, Here x, x are from k 0 (x) S (x)= x x S (x ), x kerε. (x) (x)= x 1+1 x+ x x. (x)

11 HOPF ALGEBRA STRUCTURE ON GRAPH 11 We proceed to prove that H(G) is a connected graded bialgebra. For this, denoted by { H(G) (n) } = k F M(G) V(F) =n for n 0, where V(F) is the multiset of vertices of F. Now we arrive at one of our main results in this section. Theorem 3.5. Let G be a graph. Then (a) The quintuple (H(G), m G, ½, G,ε G ) is a connected graded bialgebra. (b) The quintuple (H(G), m G, ½, G,ε G ) is a Hopf algebra with antipode S= k 0(e id H ) k. Proof. (a). By the definition of H(G), we obtain H(G)= H(G) (n). Let F 1 H(G) (p) and F 2 H(G) (q) with p, q 0. Then whence By Eq. (5), n=0 V(F 1 ) = p, V(F 2 ) =q and H(G) (p) H(G) (q) H(G) (p+q). G (H(G) (n) ) V(F 1 F 2 ) = pq, H(G) (p) H(G) (q). p+q=n Thus, the quintuple (H(G), m G, ½, G,ε G ) is a graded bialgebra. Further, H(G) (0) = k{½}=k and kerε= H(G) (n) by Eq. (6). Therefore, the quintuple (H(G), m G, ½, G,ε G ) is a connected graded bialgebra. (b). The quintuple (H(G), m G, ½, G,ε G ) is a bialgebra by Theorem Further, it is a Hopf algebra with antipode S= k 0(e id H ) k by Lemma 3.4 and Item (a) The dual Hopf algebra. In this subsection, we consider the dual Hopf algebra of (H(G), m G, ½, G,ε G ). The following elementary result is fundamental [12]. Lemma 3.6. Let H= n N H (n) be a graded space. Then (a) The graded dual H is n N H (n). Note that H is also a graded space, and H H. (b) H H is also a graded space, with (H H)(n)= n H(i) H(n i) for all n N. Moreover, (H H) H H. Where the symbol is isomomorphism. The quintuple (H(G), m G, ½, G,ε G ) is a graded Hopf algebra by Theorem 3.5. Then its graded dual inherits also a graded Hopf algebra, writed (H(G), G,ε G, m G, ½ ), by Lemma 3.6. For any F M(G), we define the following element of the graded dual H(G), { H(G) k, (7) Z F = G M(G) δ F,G, whereδ F,G is Kronecker function. Then (Z F ) F M(G) is a basis of H(G). n 1 i=1

12 12 SHOUJUN XU, XIAOMENG WANG, AND XING GAO Lemma 3.7. Let G be graph. The product of Z F and Z G is given by (8) Z F Z G = Z H, for F, G, H M(G), where H is the induced subgraph in M(G) by V(F) and V(G). H Proof. Let G be a graph. We consider the basis elements of H(G). For any F, G, H M(G), we have Z F Z G (H)= (Z F Z G )( G (H)) = (Z F Z G ) H[U] H[V] U V=V(H) = Z F (H(U)) Z G (H(V)) U V=V(H) = δ F,H(U) δ G,H(V), U V=V(H) this equation not equal to 0 if and only if H is the induced subgraph in M(G) by V(F) and V(G). Example 3.8. Let G be a graph in Item (c) of Example 2.4. We consider the product of H(G). (a) Let F= 1 and G= 2 1. Then (b) Let F= 3 and G= 2 1. We have Z 1 Z 2 1 = Z Z Z 3 Z 2 1 = Z Z Z The coproduct m G of (H(G), G,ε G, m G, ½ ) is given by Lemma 3.9. Let G be graph. The coproduct of Z Γ1 Γ n is given by (9) mg (Z Γ 1 Γ n )= Z Γi Γ j Z Γk Γ l, forγ 1,,Γ n G, i j,k l, where{1,, n}={i,,j} {k,, l}. Proof. We consider the basis elements of H(G). For anyγ 1,,Γ n, F, G M(G), we have m G (Z Γ 1 Γ n )(F G)=Z Γ1 Γ n (FG)=δ Γ1 Γ n,fg, this equation not equal to 0 if and only if FG=Γ 1 Γ n by commutative, as desired. Example Let G be a graph in Item (c) of Example 2.4. We consider the coproduct of H(G). LetΓ 1 = 1 andγ 2 = 2 1. Then Z = Z Z ½ + Z ½ Z Z 1 Z Z 2 1 Z 1.

13 HOPF ALGEBRA STRUCTURE ON GRAPH Algebra morphisms induced from graph homomorphisms. In this subsection, we are going to consider algebra morphisms induced by graph homomorphisms. For this, let us recall the concept of a graph homomorphism. Definition [20, p. 4] Let G 1 and G 2 be two graphs. A homomorphism of G 1 to G 2, written as f : G 1 G 2, is a mapping f : V(G 1 ) V(G 2 ) such that ( f (u), f (v)) E(G 2 ) whenever (u, v) E(G 1 ). Next, we give the main result in this subsection. Let f : G 1 G 2 be a graph homomorphism. Then f can be restricted to any subgraphs of G 1, still denoted by f. We define the linear map H( f ) : H(G 1 ) H(G 2 ) given by H( f )(½ G1 ) := ½ G2 and (10) H( f )(F) := H( f )(Γ 1 Γ n ) := f (Γ 1 ) f (Γ n ) for F=Γ 1 Γ n M(G 1 ) with n 1. Theorem The H( f ) : H(G 1 ) H(G 2 ) is an algebra morphism. Proof. Let F 1 =Γ 1 Γ m and F 2 =Γ 1 Γ n M(G 1). Then by Eq. (10), H( f )(F 1 F 2 )= H( f )(Γ 1 Γ m Γ 1 Γ n )= f (Γ 1) f (Γ m ) f (Γ 1 ) f (Γ n ) = H( f )(Γ 1 Γ m )H( f )(Γ 1 Γ n ) = H( f )(F 1 )H( f )(F 2 ). Thus H( f ) is an algebra morphism. This completes the proof. As an application of Theorem 3.12, we obtain a functor from the graph category to the algebra category. Let us recall Definition [22, Definition 7.1] A category is a classcof objects (denoted A, B, C, ) together with (a) a class of disjoint sets, denoted hom(a, B), one for each pair of objects inc; (an element f of hom(a, B) is called a morphism from A to B and is denoted f : A B) (b) for each triple (A, B, C) of objects ofcafunction: hom(b, C) hom(a, B) hom(a, C); (for morphisms f : A B, g : B C, this function is written (g, f ) g f and g f : A Cis called the composite of f and g); all subject to the two axioms: (i) Associativity. If f : A B, g : B C, h : C D are morphisms ofc, then h (g f )=(h g) f. (ii) Identity. For each object B ofcthere exists a morphism 1 B : B B such that for any f : A B, g : B C, 1 B f=f and g 1 B = g. Definition [22, Definition 1.1.] Let C and D be categories. A covariant functor T from C tod(denoted T :C D) is a pair of functions (both denoted by T), an object function that assigns to each object C ofcan object T(C) ofdand a morphism function which assigns to each morphism f : C D ofcamorphism T( f ) : T(C) T(D)

14 14 SHOUJUN XU, XIAOMENG WANG, AND XING GAO ofd, such that (a) T(1 C )=1 T(C) for every identity morphism 1 C ofc; (b) T(g f )=T(g) T( f ) for any two morphisms f, g ofcwhose composite g f is defined. LetGbe the class of all graphs, for G 1, G 2 G, hom(g 1, G 2 ) is the set of all graph homomorphisms f : G 1 G 2. ThenGis a category, called the graph category. LetHbe the k-algebra category whose objects are all k-algebra; the hom(a 1, A 2 ) is the set of all k-algebra morphisms ϕ : A 1 A 2. Now we arrive at out main result of this section. Theorem LetGbe the graph category andhthe k-algebra category. Define H( ) :G H, G H(G), f H( f ), where G is a graph ing, f : G 1 G 2 is a graph homomorphism and H( f ) : H(G 1 ) H(G 2 ) is the algebra morphism obtained in Theorem Then H( ) is a covariant functor. Proof. By Lemma 2.7 and Theorem 3.12, it suffices to prove Items (a) and (b) in Definition For Item (a), let F=Γ 1 Γ n M(G). Then by Eq. (10), H(1 G )(F)= H(1 G )(Γ 1 Γ n )= 1 G (Γ 1 ) 1 G (Γ n )=Γ 1 Γ n = 1 H(G) (Γ 1 ) 1 H(G) (Γ n )= 1 H(G) (Γ 1 Γ n )= 1 H(G) (F). For Item (b), let f : G 1 G 2 and g : G 2 G 3, be graph homomorphisms. For F=Γ 1 Γ n M(G 1 ), H(g) H( f )(F)= H(g) H( f )(Γ 1 Γ n )= H(g)( f (Γ 1 ) f (Γ n )) = g( f (Γ 1 )) g( f (Γ n )) = (g f )(Γ 1 ) (g f )(Γ n ) = H(g f )(Γ 1 Γ n ) = H(g f )(F), where the third equation employs the fact that the image of a connected subgraph of a graph homomorphism is a connected subgraph. Thus H(g) H( f )(F)=H(g f )(F). This completes the proof. Acknowledgments: This work was supported by the National Natural Science Foundation of China (Grant No , and ), Fundamental Research Funds for the Central Universities (Grant No. lzujbky ), the Natural Science Foundation of Gansu Province (Grant No. 17JR5RA175). References [1] E. Abe, Hopf algebras, Cambridge university press, [2] M. F. Atiyah and I. G. Macdonald, Introduction to commutative algebra, Addison-Wesley publishing company, [3] A. Connes and D. Kreimer, Hopf algebras, renormalization and non-commutative geometry, Comm. Math. Phys. 199(1) (1998), [4] A. Connes and D. Kreimer, Renormalization in quantum field theory and the Riemann-Hilbert problem. I. The Hopf algebra structure of graphs and the main theorem, Comm. Math. Phys. 210 (2000),

15 HOPF ALGEBRA STRUCTURE ON GRAPH 15 [5] A. Connes and D. Kreimer, Renormalization in quantum field theory and the Riemann-Hilbert problem. II. The β-function, diffeomorphisms and the renormalization group, Comm. Math. Phys. 216 (2001), [6] A. Dür, Möbius functions, incidence algebras and power series representations, Lecture Notes in Mathematics 1202, Springer-Verlag, Berlin [7] G. E. Duchamp, L. Foissy, N. Hoang-Nghia, D. Manchon and A. Tanasa, A combinatorial noncommutative Hopf algebra of graphs, arxiv: v3 [math. CO] 29 Sep [8] R. Diestel, Graph Theory, Springer-Verlag, [9] L. Foissy, Finite-dimensional comodules over the Hopf algebra of rooted trees, J. Algebra 255(1) (2002), [10] L. Foissy, Les algèbres de Hopf des arbres enracinés décorés I, Bull. Sci. Math. 126 (2002), [11] L. Foissy, The infinitesimal Hopf algebra and the poset of planar forests, J. Algebra. Comb. 30 (2009), [12] L. Foissy, Introduction to Hopf Algebra of Rooted Trees, available at foissy. free. fr/pageperso/preprint3. pdf. [13] L. Foissy, Chromatic polynomials and bialgebras of graphs, arxiv: V2 [math.ra] 30 Nov [14] R. Grossman and R.G. Larson, Hopf-algebraic structure of families of trees, J. Algebra 126(1) (1989), [15] L. Guo, Algebraic Birkhoff decomposition and its applications, Automorphic Forms and the Langlands Program, International Press, 2008, [16] L. Guo, An Introduction to Rota-Baxter Algebra, International Press, [17] X. Gao and X. W, Infinitesimal unitary Hopf algebras and planar rooted forests, J. Algebra. Comb.(2018) [18] X. Gao and Y. Zhang, Planar rooted forests and weighted infinitesimal Hopf algebras: an operated algebra approach, arxiv: [19] J. M. Howie, Fundementals of semigroup theory, oxfore university press, [20] P. Hell and J. Nešetřil, Graphs and Homomorphisms, oxfore university press, [21] B. Humpert and J. L. Martin, The incidence hopf algebra of graphs, SIAM J. Discrete math. (2012), pp [22] T. W. Hungerford, Algebra, Springer science+business media, Inc, [23] C. Kreimer, Quantum groups, in Graduate Texts in Mathematics (Springer, 1994), Vol [24] D. Kreimer, On the Hopf algebra structure of perturbative quantum field theories, Adv. Theor. Math. Phys. 2 (1998), [25] D. Kreimer, On Overlapping Divergences, Comm. Math. Phys. 204 (1999), [26] S. K. Lando, On a Hopf algebra in graph theory, J. Comb. Theory B 80, , (2000). [27] J.-L. Loday and M.O. Ronco, Hopf algebra of the planar binary trees, Adv. Math. 139(2) (1998), [28] J. W. Milnor and J. C. Moore, On the structure of Hopf algebras, Annals of Mathematics 2 (2012), [29] I. Moerdijk, On the Connes-Kreimer construction of Hopf algebras, Contemp. Math. 271 (2001), [30] W. R. Schmitt, Incidence Hopf algebra, J. Pure Appl. Algebra 96, (1994), pp [31] M. Sweedler, Hopf algebras, Benjamin, New York, [32] Y. Zhang, D. Chen, X. Gao and Y. F. Luo, Weighted infinitesimal bialgebras, weihgted cocycles, and Pre-Lie algebras, arxiv: [33] T.J Zhang, X. Gao and L. Guo, Hopf algebras of rooted forests, cocyles and free Rota-Baxter algebras, Journal. Math. Phys 57 (2016),

16 16 SHOUJUN XU, XIAOMENG WANG, AND XING GAO School of Mathematics and Statistics, Lanzhou University, Lanzhou, Gansu , P. R. China address: School of Mathematics and Statistics, Lanzhou University, Lanzhou, Gansu , P. R. China address: School of Mathematics and Statistics, Key Laboratory of Applied Mathematics and Complex Systems, Lanzhou University, Lanzhou, Gansu , P. R. China address: