Observations on the Lov\'asz $\theta$-Function, Graph Capacity, Eigenvalues, and Strong Products

TL;DR

This paper explores new properties and bounds related to the Lovász theta-function, especially for strongly regular and regular graphs, with applications to graph capacity, eigenvalues, and strong product parameters.

Contribution

It provides a closed-form expression for the Lovász theta-function of strongly regular graphs and new bounds for regular graphs based on eigenvalues, enhancing understanding of graph capacities and parameters.

Findings

Exact theta-function for strongly regular graphs

Bounds on eigenvalues of strong graph products

Improved bounds on chromatic numbers of strong products

Abstract

This paper provides new observations on the Lov\'{a}sz $\theta$-function of graphs. These include a simple closed-form expression of that function for all strongly regular graphs, together with upper and lower bounds on that function for all regular graphs. These bounds are expressed in terms of the second-largest and smallest eigenvalues of the adjacency matrix of the regular graph, together with sufficient conditions for equalities (the upper bound is due to Lov\'{a}sz, followed by a new sufficient condition for its tightness). These results are shown to be useful in many ways, leading to the determination of the exact value of the Shannon capacity of various graphs, eigenvalue inequalities, and bounds on the clique and chromatic numbers of graphs. Since the Lov\'{a}sz $\theta$-function factorizes for the strong product of graphs, the results are also particularly useful for parameters of strong products or strong powers of graphs. Bounds on the smallest and second-largest eigenvalues of strong products of regular graphs are consequently derived, expressed as functions of the Lov\'{a}sz $\theta$-function (or the smallest eigenvalue) of each factor. The resulting lower bound on the second-largest eigenvalue of a $k$-fold strong power of a regular graph is compared to the Alon--Boppana bound; under a certain condition, the new bound is superior in its exponential growth rate (in $k$). Lower bounds on the chromatic number of strong products of graphs are expressed in terms of the order and the Lov\'{a}sz $\theta$-function of each factor. The utility of these bounds is exemplified, leading in some cases to an exact determination of the chromatic numbers of strong products or strong powers of graphs. The present research paper is aimed to have tutorial value as well.