The association of integers, conjugate pairs and tightness with the eigenvalues of graphs provide the motivation for the following definitions. A class of graphs, with the property, that for each graph (member) of the class, there exists a pair a,b of non-zero, distinct, eigenvalues, whose sum (or product) yields the same integer, either as a fixed constant, or a function of an inherent aspect of the graph (such as its size), is said to be sum-eigen*(a+b)*pair balanced (or product-eigen*(a.b)*pair balanced, respectively). For example, complete graphs on n vertices, are eigen-bi-balanced with sum-eigen*(n-2)*pair balanced and product-eigen*(1-n)*pair balanced, and since a,b are non-zero their reciprocals (which affect the tightness of a graph ) are defined, so that this class has the eigen-balanced ratio of 1/a+1/b=(a+b)/(a.b)= (n-2)/(1-n) =f(n) which is asymptotic to the constant value of -1. The absolute value of the integral of f(n) multiplied by the average degree yields the area (n-1)(n-ln(n-1)) – we show that this is the maximum area for most known classes of eigen-bi-balanced graphs. We investigate the effect of this asymptotic ratio on the energy of the molecular representation of graphs. Cycles are generally neither sum-eigen-pair, nor product-eigen-pair balanced, while paths are only sum- eigen-pair balanced. In this paper, we introduce a class of graphs, involving q cliques each of size q, and show that this class is eigen-bi-balanced with respect to the sum -1 and product 1-q so that it has ratio 1/(q-1) asymptotic to 0, and has area q(2q+2ln(q-1)), and discuss its eigen-bi-balanced criticality.
[v1] 2013-05-08 10:15:41
Unique-IP document downloads: 65 times
Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.