Ingraph theory,thehypercube graphQnis the graph formed from the vertices and edges of ann-dimensionalhypercube.For instance, thecube graphQ3is the graph formed by the 8 vertices and 12 edges of a three-dimensional cube. Qnhas2nvertices,2n– 1nedges, and is aregular graphwithnedges touching each vertex.

Hypercube graph
The hypercube graphQ4
Vertices2n
Edges2n– 1n
Diametern
Girth4ifn≥ 2
Automorphismsn!2n
Chromatic number2
Spectrum
PropertiesSymmetric
Distance regular
Unit distance
Hamiltonian
Bipartite
NotationQn
Table of graphs and parameters

The hypercube graphQnmay also be constructed by creating a vertex for eachsubsetof ann-element set, with two vertices adjacent when their subsets differ in a single element, or by creating a vertex for eachn-digitbinary number,with two vertices adjacent when their binary representations differ in a single digit. It is then-foldCartesian productof the two-vertexcomplete graph,and may be decomposed into two copies ofQn– 1connected to each other by aperfect matching.

Hypercube graphs should not be confused withcubic graphs,which are graphs that have exactly three edges touching each vertex. The only hypercube graphQnthat is a cubic graph is the cubical graphQ3.

Construction

edit
Construction ofQ3by connecting pairs of corresponding vertices in two copies ofQ2

The hypercube graphQnmay be constructed from the family ofsubsetsof asetwithnelements, by making a vertex for each possible subset and joining two vertices by an edge whenever the corresponding subsets differ in a single element. Equivalently, it may be constructed using2nvertices labeled withn-bitbinary numbersand connecting two vertices by an edge whenever theHamming distanceof their labels is one. These two constructions are closely related: a binary number may be interpreted as a set (the set of positions where it has a1digit), and two such sets differ in a single element whenever the corresponding two binary numbers have Hamming distance one.

Alternatively,Qnmay be constructed from thedisjoint unionof two hypercubesQn− 1,by adding an edge from each vertex in one copy ofQn− 1to the corresponding vertex in the other copy, as shown in the figure. The joining edges form aperfect matching.

The above construction gives a recursive algorithm for constructing theadjacency matrixof a hypercube,An.Copying is done via theKronecker product,so that the two copies ofQn− 1have an adjacency matrix,whereis theidentity matrixindimensions. Meanwhile the joining edges have an adjacency matrix.The sum of these two terms gives a recursive function function for the adjacency matrix of a hypercube:

Another construction ofQnis theCartesian productofntwo-vertex complete graphsK2.More generally the Cartesian product of copies of a complete graph is called aHamming graph;the hypercube graphs are examples of Hamming graphs.

Examples

edit

The graphQ0consists of a single vertex, whileQ1is thecomplete graphon two vertices.

Q2is acycleof length4.

The graphQ3is the1-skeletonof acubeand is a planar graph with eightverticesand twelveedges.

The graphQ4is theLevi graphof theMöbius configuration.It is also theknight's graphfor atoroidalchessboard.[1]

Properties

edit

Bipartiteness

edit

Every hypercube graph isbipartite:it can becoloredwith only two colors. The two colors of this coloring may be found from the subset construction of hypercube graphs, by giving one color to the subsets that have an even number of elements and the other color to the subsets with an odd number of elements.

Hamiltonicity

edit
A Hamiltonian cycle on a tesseract with vertices labelled with a 4-bit cyclic Gray code

Every hypercubeQnwithn> 1has aHamiltonian cycle,a cycle that visits each vertex exactly once. Additionally, aHamiltonian pathexists between two verticesuandvif and only if they have different colors in a2-coloring of the graph. Both facts are easy to prove using the principle ofinductionon the dimension of the hypercube, and the construction of the hypercube graph by joining two smaller hypercubes with a matching.

Hamiltonicity of the hypercube is tightly related to the theory ofGray codes.More precisely there is abijectivecorrespondence between the set ofn-bit cyclic Gray codes and the set of Hamiltonian cycles in the hypercubeQn.[2]An analogous property holds for acyclicn-bit Gray codes and Hamiltonian paths.

A lesser known fact is that every perfect matching in the hypercube extends to a Hamiltonian cycle.[3]The question whether every matching extends to a Hamiltonian cycle remains an open problem.[4]

Other properties

edit

The hypercube graphQn(forn> 1):

  • is theHasse diagramof a finiteBoolean algebra.
  • is amedian graph.Every median graph is anisometric subgraph of a hypercube,and can be formed as a retraction of a hypercube.
  • has more than22n-2perfect matchings. (this is another consequence that follows easily from the inductive construction.)
  • isarc transitiveandsymmetric.The symmetries of hypercube graphs can be represented assigned permutations.
  • contains all the cycles of length4, 6,..., 2nand is thus abipancyclic graph.
  • can bedrawnas aunit distance graphin the Euclidean plane by using the construction of the hypercube graph from subsets of a set ofnelements, choosing a distinctunit vectorfor each set element, and placing the vertex corresponding to the setSat the sum of the vectors inS.
  • is an-vertex-connected graph,byBalinski's theorem.
  • isplanar(can bedrawnwith no crossings) if and only ifn≤ 3.For larger values ofn,the hypercube hasgenus(n− 4)2n− 3+ 1.[5][6]
  • has exactlyspanning trees.[6]
  • hasbandwidthexactly.[7]
  • hasachromatic numberproportional to,but the constant of proportionality is not known precisely.[8]
  • has as the eigenvalues of itsadjacency matrixthe numbers(−n,−n+ 2, −n+ 4,...,n− 4,n− 2,n)and as the eigenvalues of its Laplacian matrix the numbers(0, 2,..., 2n).Thekth eigenvalue has multiplicityin both cases.
  • hasisoperimetric numberh(G) = 1.

The familyQnfor alln> 1is aLévy family of graphs.

Problems

edit
Maximum lengths of snakes (Ls) and coils (Lc) in the snakes-in-the-box problem for dimensionsnfrom 1 to 4

The problem of finding thelongest pathor cycle that is aninduced subgraphof a given hypercube graph is known as thesnake-in-the-boxproblem.

Szymanski's conjectureconcerns the suitability of a hypercube as anetwork topologyfor communications. It states that, no matter how one chooses apermutationconnecting each hypercube vertex to another vertex with which it should be connected, there is always a way to connect these pairs of vertices bypathsthat do not share any directed edge.[9]

See also

edit

Notes

edit
  1. ^Watkins, John J. (2004),Across the Board: The Mathematics of Chessboard Problems,Princeton University Press, p. 68,ISBN978-0-691-15498-5.
  2. ^Mills, W. H. (1963), "Some complete cycles on then-cube ",Proceedings of the American Mathematical Society,14(4), American Mathematical Society: 640–643,doi:10.2307/2034292,JSTOR2034292.
  3. ^Fink, J. (2007), "Perfect matchings extend to Hamiltonian cycles in hypercubes",Journal of Combinatorial Theory, Series B,97(6): 1074–1076,doi:10.1016/j.jctb.2007.02.007.
  4. ^Ruskey, F. andSavage, C.Matchings extend to Hamiltonian cycles in hypercubeson Open Problem Garden. 2007.
  5. ^Ringel, G.(1955), "Über drei kombinatorische Probleme amn-dimensionalen Wiirfel und Wiirfelgitter ",Abh. Math. Sem. Univ. Hamburg,20:10–19,MR0949280
  6. ^abHarary, Frank;Hayes, John P.; Wu, Horng-Jyh (1988),"A survey of the theory of hypercube graphs"(PDF),Computers & Mathematics with Applications,15(4): 277–289,doi:10.1016/0898-1221(88)90213-1,hdl:2027.42/27522,MR0949280.
  7. ^Optimal Numberings and Isoperimetric Problems on Graphs, L.H. Harper,Journal of Combinatorial Theory,1, 385–393,doi:10.1016/S0021-9800(66)80059-5
  8. ^Roichman, Y. (2000), "On the Achromatic Number of Hypercubes",Journal of Combinatorial Theory, Series B,79(2): 177–182,doi:10.1006/jctb.2000.1955.
  9. ^Szymanski, Ted H. (1989), "On the Permutation Capability of a Circuit-Switched Hypercube",Proc. Internat. Conf. on Parallel Processing,vol. 1, Silver Spring, MD: IEEE Computer Society Press, pp. 103–110.

References

edit