Graham–Pollak theorem

E748752

The Graham–Pollak theorem is a result in graph theory that states the edges of a complete graph on n vertices cannot be partitioned into fewer than n−1 complete bipartite subgraphs.

All labels observed (1)

Label Occurrences
Graham–Pollak theorem canonical 1

How this entity was disambiguated

Statements (35)

Predicate Object
instanceOf result in graph theory ⓘ
theorem ⓘ
appliesTo complete bipartite subgraphs ⓘ
complete graphs ⓘ
edge partitions ⓘ
assumes complete graph on n vertices ⓘ
simple undirected graphs ⓘ
conclusion Any partition of the edge set of K_n into complete bipartite graphs uses at least n−1 parts. ⓘ
equalityCase There exists a partition of the edges of K_n into exactly n−1 complete bipartite subgraphs. ⓘ
field graph theory ⓘ
hasProofMethod incidence matrix techniques ⓘ
linear algebra ⓘ
matrix rank arguments ⓘ
implies The bipartite dimension of the complete graph K_n equals n−1. ⓘ
invariant bipartite dimension of K_n equals n−1 regardless of the partition chosen. ⓘ
lowerBound n−1 complete bipartite subgraphs are necessary to partition the edges of K_n. ⓘ
namedAfter Henry O. Pollak ⓘ
linked to: Henry Pollak

Ronald Graham ⓘ
linked to: Ronald L. Graham
originalAuthors Henry O. Pollak ⓘ
linked to: Henry Pollak

Ronald L. Graham ⓘ
publishedIn Bell System Technical Journal ⓘ
relatedTo bipartite edge cover ⓘ
rank of adjacency matrices ⓘ
relatesConcept K_n ⓘ
bipartite dimension of a graph ⓘ
complete bipartite graph ⓘ
edge partition ⓘ
statement The edges of a complete graph on n vertices cannot be partitioned into fewer than n−1 complete bipartite subgraphs. ⓘ
tightness The bound n−1 is best possible. ⓘ
topic edge decompositions into bipartite graphs ⓘ
extremal graph theory ⓘ
graph decompositions ⓘ
usedFor computing bipartite dimension of complete graphs ⓘ
lower bounds in communication complexity (via related ideas) ⓘ
yearProved 1971 ⓘ

How these facts were elicited

Referenced by (1)

Full triples — surface form annotated when it differs from this entity's canonical label.

Ronald L. Graham → notableIdea → Graham–Pollak theorem ⓘ