English

Graph functionality

Combinatorics 2018-07-06 v1

Abstract

Let G=(V,E)G=(V,E) be a graph and AA its adjacency matrix. We say that a vertex yVy \in V is a function of vertices x1,,xkVx_1, \ldots, x_k \in V if there exists a Boolean function ff of kk variables such that for any vertex zV{y,x1,,xk}z \in V - \{y, x_1, \ldots, x_k\}, A(y,z)=f(A(x1,z),,A(xk,z))A(y,z)=f(A(x_1,z),\ldots,A(x_k,z)). The functionality fun(y)fun(y) of vertex yy is the minimum kk such that yy is a function of kk vertices. The functionality fun(G)fun(G) of the graph GG is maxHminyV(H)fun(y)\max\limits_H\min\limits_{y\in V(H)}fun(y), where the maximum is taken over all induced subgraphs HH of GG. In the present paper, we show that functionality generalizes simultaneously several other graph parameters, such as degeneracy or clique-width, by proving that bounded degeneracy or bounded clique-width imply bounded functionality. Moreover, we show that this generalization is proper by revealing classes of graphs of unbounded degeneracy and clique-width, where functionality is bounded by a constant. This includes permutation graphs, unit interval graphs and line graphs. We also observe that bounded functionality implies bounded VC-dimension, i.e. graphs of bounded VC-dimension extend graphs of bounded functionality, and this extension is also proper.

Keywords

Cite

@article{arxiv.1807.01749,
  title  = {Graph functionality},
  author = {Bogdan Alecu and Aistis Atminas and Vadim Lozin},
  journal= {arXiv preprint arXiv:1807.01749},
  year   = {2018}
}