Defective coloring

From testwiki
Revision as of 14:19, 1 February 2025 by imported>BunnysBot (History: Fix CW Errors with GenFixes (T1))
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Template:Short description Template:See also

In graph theory, a mathematical discipline, coloring refers to an assignment of colours or labels to vertices, edges and faces of a graph. Defective coloring is a variant of proper vertex coloring. In a proper vertex coloring, the vertices are coloured such that no adjacent vertices have the same colour. In defective coloring, on the other hand, the vertices are allowed to have neighbours of the same colour to a certain extent.

History

Defective coloring was introduced nearly simultaneously by Andrews and Jacobson,[1] Harary and Jones and Cowen, Cowen and Woodall.[2] Surveys of this and related colorings are given by Marietjie Frick.[3] Cowen, Cowen and Woodall [2] focused on graphs embedded on surfaces and gave a complete characterization of all k and d such that every planar is (k, d)-colorable. Namely, there does not exist a d such that every planar graph is (1, d)- or (2, d)-colorable; there exist planar graphs which are not (3, 1)-colorable, but every planar graph is (3, 2)-colorable. Together with the (4, 0)-coloring implied by the four color theorem, this solves defective chromatic number for the plane. Poh [4] and Goddard [5] showed that any planar graph has a special (3,2)-coloring in which each color class is a linear forest, and this can be obtained from a more general result of Woodall. For general surfaces, it was shown that for each genus g0, there exists a k=k(g) such that every graph on the surface of genus g is (4, k)-colorable.[2] This was improved to (3, k)-colorable by Dan Archdeacon.[6] For general graphs, a result of László Lovász from the 1960s, which has been rediscovered many times [7][8][9] provides a O(∆E)-time algorithm for defective coloring graphs of maximum degree ∆.

Definitions and terminology

Defective coloring

A (kd)-coloring of a graph G is a coloring of its vertices with k colours such that each vertex v has at most d neighbours having the same colour as the vertex v. We consider k to be a positive integer (it is inconsequential to consider the case when k = 0) and d to be a non-negative integer. Hence, (k, 0)-coloring is equivalent to proper vertex coloring.[10]

d-defective chromatic number

The minimum number of colours k required for which G is (k, d)-colourable is called the d-defective chromatic number, χd(G).[11]

For a graph class G, the defective chromatic number of G is minimum integer k such that for some integer d, every graph in G is (k,d)-colourable. For example, the defective chromatic number of the class of planar graphs equals 3, since every planar graph is (3,2)-colourable and for every integer d there is a planar graph that is not (2,d)-colourable.

Impropriety of a vertex

Let c be a vertex-coloring of a graph G. The impropriety of a vertex v of G with respect to the coloring c is the number of neighbours of v that have the same color as v. If the impropriety of v is 0, then v is said to be properly colored.[2]

Impropriety of a vertex-coloring

Let c be a vertex-coloring of a graph G. The impropriety of c is the maximum of the improprieties of all vertices of G. Hence, the impropriety of a proper vertex coloring is 0.[2]

Example

Example of defective colouring of a cycle on five vertices when d = 0, 1, 2

An example of defective colouring of a cycle on five vertices, C5, is as shown in the figure. The first subfigure is an example of proper vertex colouring or a (k, 0)-coloring. The second subfigure is an example of a (k, 1)-coloring and the third subfigure is an example of a (k, 2)-coloring. Note that,

χ0(C5)=χ(C5)=3

χ1(C5)=3

χ2(Cn)=1;n

Properties

  • It is enough to consider connected graphs, as a graph G is (k, d)-colourable if and only if every connected component of G is (k, d)-colourable.[2]
  • In graph theoretic terms, each colour class in a proper vertex coloring forms an independent set, while each colour class in a defective coloring forms a subgraph of degree at most d.[12]
  • If a graph is (k, d)-colourable, then it is (k′, d′)-colourable for each pair (k′, d′) such that k′k and d′d.[2]

Some results

Every outerplanar graph is (2,2)-colorable

Proof: Let G be a connected outerplanar graph. Let v0 be an arbitrary vertex of G. Let Vi be the set of vertices of G that are at a distance i from v0. Let Gi be Vi, the subgraph induced by Vi. Suppose Gi contains a vertex of degree 3 or more, then it contains K1,3 as a subgraph. Then we contract all edges of V0V1...Vi1 to obtain a new graph G. It is to be noted that V0V1...Vi1 of G is connected as every vertex in Vi is adjacent to a vertex in Vi1. Hence, by contracting all the edges mentioned above, we obtain G such that the subgraph V0V1...Vi1 of G is replaced by a single vertex that is adjacent to every vertex in Vi. Thus G contains K2,3 as a subgraph. But every subgraph of an outerplanar graph is outerplanar and every graph obtained by contracting edges of an outerplanar graph is outerplanar. This implies that K2,3 is outerplanar, a contradiction. Hence no graph Gi contains a vertex of degree 3 or more, implying that G is (k, 2)-colorable. No vertex of Vi is adjacent to any vertex of Vi1 or Vi+1,i1, hence the vertices of Vi can be colored blue if i is odd and red if even. Hence, we have produced a (2,2)-coloring of G.[2]

Corollary: Every planar graph is (4,2)-colorable. This follows as if G is planar then every Gi (same as above) is outerplanar. Hence every Gi is (2,2)-colourable. Therefore, each vertex of Vi can be colored blue or red if i is even and green or yellow if i is odd, hence producing a (4,2)-coloring of G.

Graphs excluding a complete minor

For every integer t there is an integer N such that every graph G with no Kt minor is (t1,N)-colourable.[13]

Computational complexity

Defective coloring is computationally hard. It is NP-complete to decide if a given graph G admits a (3,1)-coloring, even in the case where G is of maximum vertex-degree 6 or planar of maximum vertex-degree 7.[14]

Applications

An example of an application of defective colouring is the scheduling problem where vertices represent jobs (say users on a computer system), and edges represent conflicts (needing to access one or more of the same files). Allowing a defect means tolerating some threshold of conflict: each user may find the maximum slowdown incurred for retrieval of data with two conflicting other users on the system acceptable, and with more than two unacceptable.[15]

Notes

Template:Reflist

References

Template:Refbegin

Template:Refend

  1. Template:Cite journal
  2. 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 Template:Cite journal
  3. Template:Cite book
  4. Template:Cite journal
  5. Template:Cite journal
  6. Template:Cite journal
  7. Template:Cite journal
  8. Template:Cite journal
  9. Template:Cite journal
  10. Template:Cite journal
  11. Template:Cite journal
  12. Cowen, L. J., Goddard, W., and Jesurum, C. E. 1997. Coloring with defect. In Proceedings of the Eighth Annual ACM-SIAM Symposium on Discrete Algorithms (New Orleans, Louisiana, United States, January 05–07, 1997). Symposium on Discrete Algorithms. Society for Industrial and Applied Mathematics, Philadelphia, PA, 548–557.
  13. Template:Cite journal
  14. Template:Cite journal
  15. Template:Cite journal