Vertex separator

In graph theory, a vertex subset ${\displaystyle S\subset V}$ is a vertex separator for nonadjacent vertices ${\displaystyle a}$ and ${\displaystyle b}$ if the removal of ${\displaystyle S}$ from the graph separates ${\displaystyle a}$ and ${\displaystyle b}$ into distinct connected components.

Examples

A separator for a grid graph.

Consider a grid graph with r rows and c columns; the total number n of vertices is r*c. For instance, in the illustration, r = 5, c = 8, and n = 40. If r is odd, there is a single central row, and otherwise there are two rows equally close to the center; similarly, if c is odd, there is a single central column, and otherwise there are two columns equally close to the center. Choosing S to be any of these central rows or columns, and removing S from the graph, partitions the graph into two smaller connected subgraphs A and B, each of which has at most n/2 vertices. If r ≤ c (as in the illustration), then choosing a central column will give a separator S with r ≤ √n vertices, and similarly if c ≤ r then choosing a central row will give a separator with at most √n vertices. Thus, every grid graph has a separator S of size at most √n, the removal of which partitions it into two connected components, each of size at most n/2.[1]

On the left a centered tree, on the right a bicentered one. The numbers show each node's eccentricity.

To give another class of examples, every free tree T has a separator S consisting of a single vertex, the removal of which partitions T into two or more connected components, each of size at most n/2. More precisely, there is always exactly one or exactly two vertices, which amount to such a separator, depending on whether the tree is centered or bicentered.[2]

As opposed to these examples, not all vertex separators are balanced, but that property is most useful for applications in computer science, such as the planar separator theorem.

Minimal separators

Let S be an (a,b)-separator, that is, a vertex subset that separates two nonadjacent vertices a and b. Then S is a minimal (a,b)-separator if no proper subset of S separates a and b. More generally, S is called a minimal separator if it is a minimal separator for some pair (a,b) of nonadjacent vertices. The following is a well-known result characterizing the minimal separators:[3]

Lemma. A vertex separator S in G is minimal if and only if the graph ${\displaystyle G-S}$, obtained by removing S from G, has two connected components ${\displaystyle C_{1}}$ and ${\displaystyle C_{2}}$ such that each vertex in S is both adjacent to some vertex in ${\displaystyle C_{1}}$ and to some vertex in ${\displaystyle C_{2}}$.

The minimal separators also form an algebraic structure: For two fixed vertices a and b of a given graph G, an (a,b)-separator S can be regarded as a predecessor of another (a,b)-separator T, if every path from a to b meets S before it meets T. More rigorously, the predecessor relation is defined as follows: Let S and T be two (a,b)-separators in 'G'. Then S is a predecessor of T, in symbols ${\displaystyle S\sqsubseteq _{a,b}^{G}T}$, if for each ${\displaystyle x\in S\setminus T}$, every path connecting x to b meets T. It follows from the definition that the predecessor relation yields a preorder on the set of all (a,b)-separators. Furthermore, Template:Harvtxt proved that the predecessor relation gives rise to a complete lattice when restricted to the set of minimal (a,b)-separators in G.

Notes

1. Template:Harvtxt. Instead of using a row or column of a grid graph, George partitions the graph into four pieces by using the union of a row and a column as a separator.
2. Template:Harvtxt

References

|CitationClass=citation }}.

• {{#invoke:citation/CS1|citation

|CitationClass=citation }}.

• {{#invoke:Citation/CS1|citation

|CitationClass=journal }}