Complexity of Restricted Star Colouring

TL;DR

This paper investigates the computational complexity of restricted star colouring, proving NP-completeness results for various graph classes and providing efficient algorithms for specific cases like trees and chordal graphs.

Contribution

It establishes NP-completeness of restricted star colouring for planar bipartite graphs and 3-star colourable graphs, and offers polynomial-time algorithms for trees and chordal graphs.

Findings

NP-complete for planar bipartite graphs with max degree k

NP-complete for 3-star colourable graphs

Polynomial-time algorithms for trees and chordal graphs

Abstract

Restricted star colouring is a variant of star colouring introduced to design heuristic algorithms to estimate sparse Hessian matrices. For $k\in\mathbb{N}$, a $k$-restricted star colouring ($k$-rs colouring) of a graph $G$ is a function $f:V(G)\to{0,1,\dots,k-1}$ such that (i)$f(x)\neq f(y)$ for every edge $xy$ of G, and (ii) there is no bicoloured 3-vertex path ($P_3$) in $G$ with the higher colour on its middle vertex. We show that for $k\geq 3$, it is NP-complete to test whether a given planar bipartite graph of maximum degree $k$ and arbitrarily large girth admits a $k$-rs colouring, and thereby answer a problem posed by Shalu and Sandhya (Graphs and Combinatorics, 2016). In addition, it is NP-complete to test whether a 3-star colourable graph admits a 3-rs colouring. We also prove that for all $\epsilon > 0$, the optimization problem of restricted star colouring a 2-degenerate bipartite graph with the minimum number of colours is NP-hard to approximate within $n^{(1/3)-\epsilon}$. On the positive side, we design (i) a linear-time algorithm to test 3-rs colourability of trees, and (ii) an $O(n^3)$-time algorithm to test 3-rs colourability of chordal graphs.