European Journal of Combinatorics ( ) Contents lists available at ScienceDirect European Journal of Combinatorics journal homepage: www.elsevier.com/locate/ejc Proof of a conjecture concerning the direct product of bipartite graphs Richard H. Hammack Department of Mathematics Applied Mathematics, Virginia Commonwealth University, Richmond, VA 23284-214, USA a r t i c l e i n f o a b s t r a c t Article history: Available online xxxx We prove that if the direct product of two connected bipartite graphs has isomorphic components, then one of the factors admits an automorphism that interchanges its partite sets. his proves a conjecture made by Jha, Klavžar Zmazek in 1997 P. Jha, S. Klavzar, B. Zmazek, Isomorphic components of Kronecker product of bipartite graphs, Discussiones Mathematicae Graph heory 17 (1997) 32 38. 28 Elsevier Ltd. All rights reserved. 1. Introduction If G H are graphs (possibly with loops) then their direct product is the graph G H whose vertex set is the Cartesian product V(G) V(H) whose edges are all pairs (g, h)(g, h ) with gg E(G) hh E(H). It is a stard fact, first proved by Weichsel 8, that if G H are connected bipartite, then G H has exactly two components. hese components may or may not be isomorphic, depending on G H. For example, Fig. 1 shows two products of bipartite graphs, where in each case the two components are distinguished by solid dashed lines. In Fig. 1(a) the components are not isomorphic, in Fig. 1(b) they are. A number of authors have sought structural conditions on G H that characterize the condition of G H having isomorphic components. Jha, Klavžar Zmazek 6 observed that this condition seems to be related to a certain kind of symmetry in at least one of the factors. hey prove that if either G or H admits an automorphism that interchanges its partite sets, then G H has isomorphic components. hey conjecture that the converse is true. In 3 it is proved that the converse is true if G H are square-free. his note presents a general proof of the converse. o summarize our main result, we state the following definition theorem. (he definition was introduced in 6.) E-mail address: rhammack@vcu.edu. 195-6698/$ see front matter 28 Elsevier Ltd. All rights reserved. doi:1.116/j.ejc.28.9.15 European Journal of Combinatorics (28), doi:1.116/j.ejc.28.9.15
2 R.H. Hammack / European Journal of Combinatorics ( ) Fig. 1. Products of bipartite graphs. Definition 1. A connected bipartite graph has property π if it admits an automorphism that interchanges its partite sets. heorem 1. Suppose G H are connected bipartite graphs. he two components of G H are isomorphic if only if at least one of G or H has property π. As an example, neither G nor H in Fig. 1(a) has property π, indeed the components of G H are not isomorphic. By contrast, in Fig. 1(b) G has property π (the nontrivial automorphism switches the bipartition) the components of G H are isomorphic. One direction of heorem 1 is proved as follows. Suppose G or H (say G) has property π. hen there is an automorphism ϕ of G that interchanges its partite sets. By 6 (heorem 3.2), the map given by (g, h) (ϕ(g), h) restricts to an isomorphism of one component of G H to the other, so if one of G or H has property π, then G H has isomorphic components. he remainder of this note is devoted to proving the converse. 2. Preliminaries We assume that the reader is familiar with the basic properties of direct products graph homomorphisms. (For excellent surveys, see 5,4.) For convenience, we collect in this section some necessary definitions, ideas stard results. Observe that the direct product can be regarded as a product on digraphs. Given a digraph G, we let gg denote an arc pointing from g to g. hen G H is the digraph with arcs (g, h)(g, h ) directed from (g, h) to (g, h ) whenever there are arcs gg in G hh in H. Since any graph can be identified with a symmetric digraph (where each edge is replaced by a double arc) the direct product of graphs is a special case of the direct product of digraphs. hough our main result is about graphs, we will use digraphs where necessary in the proofs. Suppose G is a digraph whose vertices are ordered as g 1, g 2,..., g m. he adjacency matrix for G relative to this ordering is the m m matrix A for which a ij 1 or according to whether g i g j is or is not an arc of G. Graphs are identified with symmetric digraphs, so the adjacency matrix of a graph is symmetric, that is it satisfies A A, where indicates transpose. he tensor product of two matrices A B is the matrix A B obtained by replacing each entry a ij of A with the block a ij B. he rows columns of A B thus divide into blocks corresponding to the rows columns of A. he rows (columns) of A B are indexed by ordered pairs so that (i, j) indexes the row (column) corresponding to the jth row (column) of B in the ith row-block (column-block). herefore the entry of A B in row (i, j) column (k, l) equals a ik b jl. Although A B B A in general, there are permutation matrices M N for which M(A B)N B A. (Matrix M corresponds to the permutation that rearranges the row list (1, 1), (1, 2)..., (1, n), (2, 1), (2, 2)... (2, n),..., (m, 1), (m, 2)... (m, n) as (1, 1), (2, 1)..., (m, 1), (1, 2), (2, 2)... (m, 2),..., (1, n), (2, n)... (m, n), etc.) In fact, M(C D)N D C for any C D having the same sizes as A B, respectively. It is simple to verify that if digraphs G H have adjacency matrices A B, respectively, relative to vertex orderings g 1, g 2,..., g m h 1, h 2,..., h n, respectively, then G H has adjacency matrix European Journal of Combinatorics (28), doi:1.116/j.ejc.28.9.15
R.H. Hammack / European Journal of Combinatorics ( ) 3 A B relative to the ordering (g 1, h 1 ), (g 1, h 2 ),..., (g 1, h n ), (g 2, h 1 ), (g 2, h 2 ),..., (g 2, h n ),..., (g m, h 1 ), (g m, h 2 ),..., (g m, h n ) of its vertices. Proposition 1 (2, Lemma 8.1.1). Suppose G H are digraphs with adjacency matrices A B respectively. hen G H if only if there is a permutation matrix P for which PAP B. In using Proposition 1, we keep in mind the connection between the matrix P the corresponding isomorphism θ : G H, namely p ij 1 if only if θ(g i ) h j. For example, suppose G H are connected bipartite graphs, the vertices of each graph are ordered so that all vertices of one partite set are listed first, followed by vertices in the other partite set. Since θ preserves partite sets, P must have block form or. Q R R Q he following classic result of Lovász plays a major role in our main proof. Proposition 2 (Lovász, 7, heorem 6). Let A, B, C D be digraphs, suppose there exists a homomorphism D C. If C A C B, then D A D B. 3. Results We begin with some observations about the adjacency matrices of the direct product of two bipartite graphs. Suppose G H are connected bipartite graphs with bipartitions (G, G 1 ) ({w 1, w 2,... w m }, {x 1, x 2,..., x n }) (H, H 1 ) ({y 1, y 2,... y p }, {z 1, z 2,..., z q }), respectively. hen the adjacency matrices of G H relative to these vertex orderings have the forms A A B B, respectively. Now, it is simple to check that the two components of G H are induced on the vertex sets (G H ) (G 1 H 1 ) (G H 1 ) (G 1 H ), respectively. We now construct adjacency matrices for the two components. List the vertices of (G H ) (G 1 H 1 ) in the following order: (w 1, y 1 ), (w 1, y 2 ),..., (w 1, y p ), (w 2, y 1 ), (w 2, y 2 ),..., (w 2, y p ),............, (w m, y 1 ), (w m, y 2 ),..., (w m, y p ), (x 1, z 1 ), (x 1, z 2 ),..., (x 1, z q ), (x 2, z 1 ), (x 2, z 2 ),..., (x 2, z q ),............, (x n, z 1 ), (x n, z 2 ),..., (x n, z q ). Likewise, list the vertices of (G H 1 ) (G 1 H ) as (w 1, z 1 ), (w 1, z 2 ),..., (w 1, z q ), (w 2, z 1 ), (w 2, z 2 ),..., (w 2, z q ),............, (w m, z 1 ), (w m, z 2 ),..., (w m, z q ), (x 1, y 1 ), (x 1, y 2 ),..., (x 1, y p ), (x 2, y 1 ), (x 2, y 2 ),..., (x 2, y p ),............, (x n, y 1 ), (x n, y 2 ),..., (x n, y p ). As was observed in 1, it is simple to check that, relative to these vertex orderings, the two components of G H have adjacency matrices A B A B A B A B (1). (2) Now suppose the two components of G H are isomorphic. Any isomorphism θ between them preserves their partite sets, so either θ(g H ) G H 1 θ(g 1 H 1 ) G 1 H, or θ(g H ) G 1 H θ(g 1 H 1 ) G H 1. Proposition 1 ( the remark that follows it) applied to (2) guarantees permutation matrices Q R for which either Q A B Q R A B A B R A B European Journal of Combinatorics (28), doi:1.116/j.ejc.28.9.15
4 R.H. Hammack / European Journal of Combinatorics ( ) or R A B R Q A B Q A B A B Multiplying the matrices, we see that either Q (A B)R A B in the first case, or Q (A B)R A B in the second. Renaming R as R, we have the following lemma. A Lemma 1. Suppose connected bipartite graphs G H have adjacency matrices A B, respectively. If the components of G H are isomorphic, then there are permutation matrices B Q R for which either Q (A B)R A B or Q (A B)R A B. A Lemma 2. Let bipartite graph G have matrix. If there are permutation matrices Q R with Q AR A, then G has property π. Proof. Suppose Q AR A. hen R A R Q A Q A A A. By Proposition 1 the remark that follows it, we have an isomorphism from G to itself that reverses the partite sets, so G has property π. Now we can see how the proof of our main theorem will work. Suppose G H has isomorphic components. By Lemma 1, there are permutation matrices Q R for which either Q (A B)R A B or Q (A B)R A B. Roughly speaking, we want to cancel the common factor of B (or A) deduce that there are permutation matrices Q R for which Q AR A or Q BR B, from which Proposition 2 implies G or H has property π. he fact that this cancellation is justified follows from the next lemma. Lemma 3. Suppose A, A C are 1 matrices for which C O, A is square has at least one nonzero entry in each row. Suppose also there are permutation matrices Q R for which Q (C A)R C A. hen there are permutation matrices Q R for which Q AR A. Also, if Q (A C)R A C, then there are permutation matrices Q R for which Q AR A. Proof. We begin with the first statement. Suppose Q (C A)R C A. Let D be the digraph with adjacency matrix A, let D be the digraph with adjacency matrix A, let E be the digraph with adjacency matrix of block form C C. (3) (he lower left block is written as C rather than to indicate its size.) he adjacency matrices of E D E D are then C A C A respectively. Notice that the lower left block of zeros in the matrix for E D has the same size as C A. Since A is square, this block is also the same size as C A (C A), so this block can be multiplied by R. (Recall Q (C A)R C A, which means R (C A) Q (C A ).) By letting Q R be blocks in a larger permutation matrix, we have Q C A R Q Q (C A)R R C A. hen from Proposition 1 we deduce E D E D. European Journal of Combinatorics (28), doi:1.116/j.ejc.28.9.15.
R.H. Hammack / European Journal of Combinatorics ( ) 5 Recall that the graph E corresponds to the matrix C that we want to eliminate. In pursuit of this goal, we next show that the E in the expression E D E D can be replaced by a very simple graph K. Let K be the digraph on two vertices k k consisting of the single arc kk. We claim that K D K D : Since C O, we know from (3) that E has at least one arc ee. hen the map sending k to e k to e is a homomorphism K E. Proposition 2 now applies to give K D K D. he adjacency matrices for K D K D are A A A A K D K D, Proposition 1 guarantees a permutation matrix P for which A P P A A A., respectively. Since Since A has nonzero rows it follows that left-multiplying by P in (4) does not permute any of the rows Q of the upper blocks to rows of the lower blocks. herefore P has block form P, R so (4) yields Q AR A. We now have permutation matrices Q R with Q AR A. his proves the first part of the lemma. For the second part of the lemma it must be shown that Q (A C)R A C implies there are permutation matrices Q R for which Q AR A. Let M N be the permutation matrices for which M(A C)N C A M(A C)N C A. From Q (A C)R A C it follows that MQM (C A)N RN C A. hen the first part of the lemma guarantees there are permutation matrices Q R with Q AR A. heorem 1. Suppose G H are connected bipartite graphs. he two components of G H are isomorphic if only if at least one of G or H has property π. Proof. As was noted earlier, the necessity was proved in 6. We prove the sufficiency here. Suppose the two components of G H are isomorphic. Using the notation established in (1), along with Lemma 1, we get that either Q (A B)R A B or Q (A B)R A B. If Q (A B)R A B, it follows that A is square; since G is connected A has nonzero entries in each row column. hus Lemma 3 implies the existence of permutation matrices Q R for which Q AR A, so Lemma 2 implies G has property π. On the other h, if Q (A B)R A B, the same reasoning shows H has property π. References 1. Chow, he Q -spectrum spanning trees of tensor products of bipartite graphs, Proceedings of the American Mathematical Society 125 (11) (1997) 3155 3161. 2 C. Godsil, G. Royle, Algebraic Graph heory, in: GM Series No 27, Springer-Verlag, New York, 2. 3 R. Hammack, Isomorphic components of direct products of bipartite graphs, Discussiones Mathematicae Graph heory 26 (26) 231 248. 4 P. Hell, J. Nešetřil, Graphs Homomorphisms, in: Oxford Lecture Series in Mathematics, Oxford U. Press, New York, 24. 5 W. Imrich, S. Klavžar, Product Graphs; Structure Recognition, in: Wiley Interscience Series in Discrete Mathematics Optimization, Wiley Interscience, New York, 2. 6 P. Jha, S. Klavžar, B. Zmazek, Isomorphic components of Kronecker product of bipartite graphs, Discussiones Mathematicae Graph heory 17 (1997) 32 38. 7 L. Lovász, On the cancellation law among finite relational structures, Periodica Mathematica Hungarica 1 (2) (1971) 145 156. 8 P. Weichsel, he Kronecker product of graphs, Proceedings of the American Mathematical Society 13 (1962) 47 52. (4) European Journal of Combinatorics (28), doi:1.116/j.ejc.28.9.15