Suppose a graph satisfying Ore's theorem were not Hamiltonian. Extend it, on the same vertices, to an edge-maximal non-Hamiltonian graph . Adding edges preserves the degree-sum hypothesis. The complete graph is Hamiltonian, so has a missing edge . By maximality, has a Hamiltonian cycle, which must use the new edge. Removing it gives a Hamiltonian path in , labelled .
If some has both edges and , the path can be closed into the cyclecontradicting the choice of . Thus the index sets and are disjoint subsets of . Their sizes are and . Consequently , contradicting the degree-sum condition for the nonadjacent pair . Hence the original graph is Hamiltonian.
Take a complete graph on vertices and add a fifth vertex joined only to and . The minimum degree is two, which is less than . The only nonadjacent pairs are and , and each has degree sum . Thus Ore's condition holds even though the minimum-degree condition fails. For example,is a Hamiltonian cycle in this graph.
Represent the two sides of the marriage problem as the parts of a bipartite graph, with an edge for each allowed pairing. A matching is a set of edges with no common endpoint; a complete matching of covers every vertex in . When the two parts have equal size it also covers , giving a perfect matching.
Hall's marriage theorem gives the exact finite-graph criterion:Here is the set of right-side neighbours of . Necessity follows because distinct matched vertices of need distinct neighbours; Hall's theorem supplies sufficiency.
Let . For the bound is immediate. Otherwise choose a vertex-minimal subgraph with chromatic number . Each vertex has degree at least in : if one had degree at most , a -colouring of the smaller graph obtained by deleting it could be extended to it using an absent neighbour colour, contradicting .
Root the tree at any vertex and order vertices so every parent precedes its children. The root has colour choices. Each of the other vertices has exactly one earlier adjacent vertex, its parent, and therefore exactly choices. There are no other restrictions because a tree has no cycles and no extra edges between those branches. Multiplication yieldsThe formula also includes the one-vertex tree.
Articles by others on the same topic
There are currently no matching articles.