In mathematical chemistry, the median eigenvalues of the adjacency matrix of a molecular graph are strictly related to orbital energies and molecular orbitals. In this regard, the difference between the occupied orbital of highest energy (HOMO) and the unoccupied orbital of lowest energy (LUMO) has been investigated (see Fowler and Pisansky in Acta Chim. Slov. 57:513-517, 2010). Motivated by the HOMO-LUMO separation problem, Jaklič et al. in (Ars Math. Contemp. 5:99-115, 2012) proposed the notion of HL-index that measures how large in absolute value are the median eigenvalues of the adjacency matrix. Several bounds for this index have been provided in the literature. The aim of the paper is to derive alternative inequalities to bound the HL-index. By applying majorization techniques and making use of some known relations, we derive new and sharper upper bounds for this index. Analytical and numerical results show the performance of these bounds on different classes of graphs.
The Hückel molecular orbital method (HMO) (see [3]) is a methodology for the determination of energies of molecular orbitals of π-electrons. It has been shown that π-electron energy levels are strictly related to graph eigenvalues. For this reason, graph spectral theory became a standard mathematical tool of HMO theory (see [4–6] and [7]). Among the various π-electron properties that can be directly expressed by means of graph eigenvalues, one of the most significant is the so-called HOMO-LUMO separation, based on the gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO). For more details as regards the HOMO-LUMO separation issue we refer the reader to [8–10] and [11].
In some recent works, Fowler and Pisanski ([1] and [12]) introduced an index of a graph that is related to the HOMO-LUMO separation. By analogy with the spectral radius, these authors proposed the notion of the HOMO-LUMO radius which measures how large in absolute value may be the median eigenvalues of the adjacency matrix of a graph. At the same time, an analogous definition is given in [2] that introduces the HL-index of a graph. Several bounds for this index have been proposed for some classes of graphs in [1] and [13]. Recently in [14] the authors provided some inequalities on the HL-index through the energy index.
The contribution of this paper is along those lines: we derive, through a methodology based on majorization techniques (see [15–17] and [18]), new bounds on the median eigenvalues of the normalized Laplacian matrix. Consequently, given the relation between the normalized Laplacian matrix and the adjacency matrix eigenvalues, we provide some new bounds for the HL-index. In particular, we employ a theoretical methodology proposed by Bianchi and Torriero in [19] based on nonlinear global optimization problems solved through majorization techniques. These bounds can also be quantified by using the numerical approaches developed in [20] and [21] and extended for the normalized Laplacian matrix in [22] and in [23].
Furthermore, another approach to derive new bounds makes use of the relation between HL-index and energy index. In particular, we take advantage of an existing bound on the energy index (see [24]) depending on additional information on the first eigenvalue of the adjacency matrix. This additional information is obtained here by using majorization techniques in order to provide new inequalities for the HL-index of bipartite and non-bipartite graphs.
The paper is organized as follows. In Section 2 some notations and preliminaries are given. Section 3 concerns the identification of new bounds for the HL-index. In particular, in Section 3.1, by the fact that the eigenvalues of normalized Laplacian matrix and adjacency matrix are related, we localize the eigenvalues of the normalized Laplacian matrix via majorization techniques. In Section 3.2 we find a tighter alternative upper bound for the HL-index by using the relation with energy index provided in [14] and an existing bound on energy index proposed in [24]. Section 4 shows how the bound determined in Section 3.2 improves those presented in the literature.
2 Notations and preliminaries
In this section we first recall some basic notions on graph theory (for more details refer to [25]) and on the HL-index. Considering a simple, connected and undirected graph \(G=(V,E)\) where \(V=\{1, 2, \ldots, n\}\) is the set of vertices and \(E\subseteq V\times V\) the set of edges, \({|E|=m}\). The degree sequence of G is denoted by \(\pi=(d_{1},d_{2},\ldots,d_{n})\) and it is arranged in non-increasing order \(d_{1}\geq d_{2}\geq \cdots \geq d_{n}\), where \(d_{i}\) is the degree of vertex i.
Let \(A(G)\) be the adjacency matrix of G and \(D(G)\) be the diagonal matrix of vertex degrees. The matrix \(L(G)=D(G)-A(G)\) is called Laplacian matrix of G, while \(\mathcal{L}(G)=D(G)^{-1/2}L(G)D(G)^{-1/2}\) is known as the normalized Laplacian. Let \(\lambda _{1}\geq \lambda _{2}\geq \cdots\geq \lambda _{n}\), \(\mu _{1}\geq \mu _{2}\geq \cdots\geq \mu_{n}\) and \(\gamma _{1}\geq \gamma_{2}\geq \cdots\geq \gamma_{n}\) be the set of real eigenvalues of \(A(G)\), \(L(G)\), and \(\mathcal{L}(G)\), respectively. The following properties of spectra of \(A(G)\) and \(\mathcal{L}(G)\) hold:
The eigenvalues involved in the HOMO-LUMO separation are \(\lambda_{H}\) and \(\lambda_{L}\), where \(H= \lfloor \frac{n+1}{2} \rfloor\) and \(L= \lceil \frac{n+1}{2} \rceil\).
We now recall the following results regarding nonlinear global optimization problems solved through majorization techniques. We refer the reader to [19] for more details as regards majorization techniques and for the proofs of Lemma 1 and Theorems 1 and 2 recalled in the following.
Let g be a continuous function, homogeneous of degree p, real, and strictly Schur-convex (see [28] for the definition of Schur-convex functions and related properties). Let us assume
We can now deduce upper and lower bounds for \(x_{h}\) (with \(h=1,\ldots,N\)) by solving the following optimization problems \(P(h)\) and \(P^{*}(h)\):
Theorem 1
The solution of the optimization problem\(P(h)\)is\((\frac{a}{N})\)if\(b=\frac{a^{p}}{N^{p-1}}\). If\(b \neq \frac{a^{p}}{N^{p-1}}\), the solution of the optimization problem\(P(h)\)is\(\alpha^{*}\)where
1.
for\(h>h^{*}\), \(\alpha^{*}\)is the unique root of the equation
The solution of the optimization problem\(P^{*}(h)\)is\((\frac{a}{N})\)if\(b=\frac{a^{p}}{N^{p-1}}\). If\(b \neq \frac{a^{p}}{N^{p-1}}\), the solution of the optimization problem\(P^{*}(h)\)is\(\alpha^{*}\)where
1.
for\(h=1\), \(\alpha^{*}\)is the unique root of the equation
3 Some new bounds for the HL-index via majorization techniques
3.1 Bounds on median eigenvalues of the normalized Laplacian matrix
In this section we localize median eigenvalues of the normalized Laplacian matrix by applying the methodology (see [19]) recalled in Section 2 (Theorems 1 and 2). The additional information on median eigenvalues and the interlacing between eigenvalues of normalized Laplacian and adjacency matrices turned out to be a handy tool for bounding the HL-index for both non-bipartite and bipartite graphs. According to [29] (see Theorem 2.2.1), the following relations hold:
From (11), we can easily derive the following bounds:
$$ 0 \leq R(G) \leq \left \{ \textstyle\begin{array}{l@{\quad}l} d_{1}\max (|1-\gamma_{\frac{n+2}{2}}|,|1-\gamma_{\frac{n}{2}}| ) & \text{ if $n$ is even,} \\ d_{1} (|1-\gamma_{\frac{n+1}{2}}| ) & \text{ if $n$ is odd.} \end{array}\displaystyle \right . $$
(14)
By applying majorization, we are able to bound the median eigenvalues \(\gamma_{i}\) (with \(i=\frac{n}{2},\frac{n+1}{2},\frac{n+2}{2}\)) considered in (14).
where the left inequality is attained for the complete graph \(G=K_{n}\).
It is noteworthy to see that \(S_{b}^{1}\) is derived from the general set S with \(a=n\), \(N=n-1\), \(b=b_{1}\), and1\(p=2\). By Lemma 1 we have for a non-complete graph
Considering \(\gamma_{\frac{n}{2}}\) for n even, by (16) we have \(h < h^{\ast}\). Hence, by solving equation (8) of Theorem 1, we can derive the unit root \(\alpha_{1}\) such as \(\gamma_{\frac{n}{2}}\leq \alpha_{1}\). Therefore
In virtue of (15), we have \(\frac{n}{n-1} \leq \alpha_{1} < 2\) with the left inequality attained only for the complete graph \(G=K_{n}\).
In a similar way, we can evaluate the value \(\beta_{1} \leq \gamma_{\frac{n}{2}} \), through the solution of equation (10) of Theorem 2 (where \(h< h^{\ast}+1\)), where
In this case, we derive \(\vert 1-\gamma_{\frac{n+2}{2}}\vert \leq \max(|1-\alpha_{2}|, |1-\beta_{2}|)\).
3.
For a graph with n odd number of vertices, we need to study only \(\gamma_{\frac{n+1}{2}}\) with \(\beta_{3} \leq \gamma_{\frac{n+1}{2}} \leq \alpha_{3}\), and we have by Theorem 1 and Theorem 2, respectively:
where the left inequality is attained for the complete graph \(G=K_{n}\).
The set \(S_{b}^{2}\) is derived from the general set S with \(a=n-2\), \(N=n-2\), \(b=b_{2}\), and \(p=2\). By Lemma 1 we have for a non-complete graph
In a similar way, we can evaluate the value \(\beta^{\mathrm{bip}}_{1} \leq \gamma_{\frac{n}{2}} \), where \(h < h^{\ast}+1\). Applying Theorem 2 entails
It is easy to show that \(|1-\alpha^{\mathrm{bip}}_{1}| \leq |1-\beta^{\mathrm{bip}}_{1}|\). Hence bound (17) follows. □
3.2 Bounds on \(R(G)\) through the energy index
In the following we obtain bounds on the HL-index starting from (2). Our aim is to bound the energy index making use of additional information on the first eigenvalue of \(A(G)\). In [24] the authors show that, if a tighter bound k on \(\lambda_{1}\) such as \(\lambda_{1} \geq k \geq \frac{2m}{n}\) is available, then the energy index for a non-bipartite graph is bounded as
$$ E(G)\leq k + \sqrt{(n-1) \bigl(2m-k^{2}\bigr)}, $$
We start by proving that the condition \((2m-k^{2})\geq0\) required in bound (22) is always satisfied for simple and connected graphs.
We have \(k \in ( \frac{2m}{n},\frac{1}{2} (n+\sqrt{4m-n^{2}} ) )\). Indeed, by the basic concepts of calculus it is easy to see that k increases when m increases and then \(h^{\ast}\) tends to 1. Hence, k is limited from above by \(\frac{1}{2} (n+\sqrt{4m-n^{2}} )\) and where \(\frac{1}{2} (n+\sqrt{4m-n^{2}} )\leq \sqrt{2m}\) the required condition is satisfied.
We now show how bound (22) improves bound (3) presented in [14].
We need to prove that the following inequality holds:
The function \(f(k)\) is a convex parabola that assumes negative values in the range of k we are interested in. Indeed we have \(f(\frac{2m}{n}) \leq 0\) and \(f(\sqrt{2m}) \leq 0\).
Both sides of (24) being negative we can apply some basic concepts of algebra, getting
The function \(t(k)=k^{2}n^{2}+k ( 4mn-8m ) +8m-8mn+4m^{2}\) is again a convex parabola with vertex \(( \frac{2m ( 2-n ) }{n^{2}},\frac{8m ( 2m-n^{2} ) ( n-1 )}{n^{2}} ) \). Both coordinates are less than zero (then \(\frac{2m (2-n ) }{n^{2}} < \frac{2m}{n}\)). Having \(t(\frac{2m}{n})=8m(n-2m)(1-n)\geq 0\), inequality (25) is satisfied.
Therefore, bound (22) performs better than or equal to bound (3).
Furthermore, we see that both bounds perform equally when \(h^{\ast}= \frac{n^{2}}{2m}\) (i.e.\(\frac{n^{2}}{2m}\) is an integer). It is noteworthy that:
(a)
when n is odd, \(\frac{n^{2}}{2m}\) is never an integer (\(\lfloor \frac{n^{2}}{2m} \rfloor \neq \frac{n^{2}}{2m}\));
(b)
when n is even, \(\frac{n^{2}}{2m}\) is an integer when \(m=\frac{n^{2}}{2x}\) with \(x|\frac{n^{2}}{2}\), \(2 \leq x \leq \frac{n}{2}\) (where \(x|\frac{n^{2}}{2}\) is shorthand for ‘x divides \(\frac{n^{2}}{2}\)’). In this case \(k=\frac{2m}{n}\) and we derive bound (3).
2.
Bipartite graphs
As for non-bipartite graphs, we start by proving that the condition \((m-k^{2})\geq0\) required in bound (23) is always satisfied for simple and connected graphs. In this case \(h^{\ast}\) tends to 2 for complete graphs (where \(m \leq \frac{n^{2}}{4}\)).
By some basic algebraic concepts, we see that \((m-k^{2})\geq0\) entails:
The function \(f(m)=m^{2}(h^{4}+2h^{3}-3h^{2}-4h+4)+m(2n^{2}h-4n^{2}-2n^{2}h^{2})+n^{4}\) is concave, decreasing, and non-negative on the interval \(m \in (\frac{n}{2},\frac{n^{2}}{4} )\). Therefore, the required condition is satisfied.
We now show how bound (23) improves bound (4) presented in [14].
We need to prove that the following inequality holds:
The left-hand side term of (26) can be represented by
$$ f(k)=k^{2}n^{2}-8km+4mn+4m^{2}-2mn^{2}. $$
The function \(f(k)\) is a convex parabola that assumes negative values in the range of k we are interested in. Indeed we have \(f(\frac{2m}{n}) \leq 0\) and \(f(\sqrt{m}) \leq 0\).
Both sides of (26) being negative, by some manipulations we obtain
$$ k^{2}n^{2}+4km(n-4)+4m^{2}+16m-8mn\geq 0 . $$
(27)
The function \(t(k)=k^{2}n^{2}+4km(n-4)+4m^{2}+16m-8mn\) is again a convex parabola with vertex \(( \frac{2m (4-n ) }{n^{2}},\frac{8m(4m-n^{2})(n-2)}{n^{2}} )\). Both coordinates are less than zero (then \(\frac{4m ( 4-n ) }{2n^{2}} < \frac{2m}{n}\)). Having \(t(\frac{2m}{n})=\frac{8m}{n}(n+2m(n-2))\geq 0\), inequality (27) is satisfied.
Hence bound (23) performs better than or equal to bound (4).
□
4 Numerical results
4.1 Non-bipartite graphs
In this subsection we compare our bounds ((12), (13), and (22)) with those existing in the literature. We initially compute them for graphs generated by using the Erdös-Rényi (ER) model \(G_{\mathrm{ER}}(n,q)\) (see [31–33] and [34]) where edges are included with probability q independent from every other edge. Graphs have been derived randomly by using the well-known package of R (see [35]) and by ensuring that the graph is connected. Table 1 compares alternative upper bounds of \(R(G)\) evaluated for simulated \(G_{\mathrm{ER}}(n,0.5)\) graphs with different number of vertices. It is noteworthy that bound (13) has the best performance for \(n=5\), while bound (22) is the sharpest one in all other cases. However, the improvements are very slight with respect to bound (3) when large graphs are considered.
We now extend the analysis to the same class of graphs but with varying the probability q. Two alternative values (\(q=0.1\) and \(q=0.9\), respectively) have been tested in Tables 2 and 3. In both cases bound (22) appears always as the tightest one. In particular, we observe significant differences with respect to other bounds when the number of edges increases. It is noticeable in Table 2 how the improvement is very remarkable even for very large graphs.
We now focus on bounds (3), (5), and (22) that share the advantage of being dependent only on n and/or m. In Section 3.2, it has been already proved that bound (22) is better than or equal to bound (3). This result is obviously confirmed by Figure 1, where it is interesting to highlight how bound (22) strongly improves the other bounds for large values of m.
Figure 1
Upper bounds of\(\pmb{R(G)}\)for non-bipartite graphs with different number of vertices and edges.
4.2 Bipartite graphs
We now report in Table 4 a comparison of alternative bounds derived for bipartite graphs, varying the number of vertices and edges. In some selected cases (i.e.\(n=10\), \(m=25\) and \(n=20\), \(m=91\)), our bound (17) performs better. In all the other cases, the tightest one is bound (23) as obtained in Section 3.2.
As for non-bipartite graphs, we can focus on bounds (4), (6), and (23) that can be computed without generating any graph. We observe in Figure 2 how the bound (23) provides a significant improvement for higher values of m. This result is also confirmed for very large graphs.
Figure 2
Upper bounds of\(\pmb{R(G)}\)for bipartite graphs with different number of vertices and edges.
5 Conclusions
In this paper alternative upper bounds on the HL-index are provided by means of majorization techniques. On one hand, we find new bounds for both non-bipartite and bipartite graphs by exploiting additional information on the median eigenvalues and the interlacing between the eigenvalues of normalized Laplacian and adjacency matrices. On the other hand, new bounds are derived by making use of the relation between the HL-index and the energy index. Analytical and numerical results show the performance of these bounds on different classes of graphs. In particular, the bound related to the energy index performs better with respect to the best-known results in the literature.
For values of \(p > 2\), b (and then \(b_{1}\) and \(b_{2}\) used in the sequel) depends on the graph’s structure and topology. So the procedure can be only numerically applied: we need to compute the eigenvalues of either adjacency or normalized Laplacian matrix, but this information allows one to directly obtain \(R(G)\). In this case, the evaluation of the bounds is useless.
Declarations
Acknowledgements
The authors are grateful to Anna Torriero and Monica Bianchi for useful advice and suggestions.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Authors’ Affiliations
(1)
Department of Mathematics and Econometrics, Catholic University
References
Fowler, PW, Pisansky, T: HOMO-LUMO maps for chemical graphs. Acta Chim. Slov. 57, 513-517 (2010)
Google Scholar
Jaklič, G, Fowler, WP, Pisanski, T: HL-index of a graph. Ars Math. Contemp. 5, 99-115 (2012)
MathSciNetMATHGoogle Scholar
Coulson, CA, O’Leary, B, Mallion, RB: Hückel Theory for Organic Chemists. Academic Press, London (1978)
Google Scholar
Dias, JR: Molecular Orbital Calculations Using Chemical Graph Theory. Springer, Berlin (1993)
View ArticleGoogle Scholar
Graovac, A, Gutman, I, Trinajstić, N: Topological Approach to the Chemistry of Conjugated Molecules. Springer, Berlin (1977)
View ArticleMATHGoogle Scholar
Gutman, I, Polansky, OE: Mathematical Concepts in Organic Chemistry. Springer, Berlin (1986)
View ArticleMATHGoogle Scholar
Gutman, I, Trinajstić, N: Graph theory and molecular orbits. Total π-electron energy of alternant hydrocarbons. Chem. Phys. Lett. 17, 535-538 (1972)
View ArticleGoogle Scholar
Graovac, A, Gutman, I: Estimation of the HOMO-LUMO separation. Croat. Chem. Acta 53, 45-50 (1980)
Google Scholar
Gutman, I: Note on a topological property of the HOMO-LUMO separation. Z. Naturforsch. 35a, 458-460 (1980)
MathSciNetGoogle Scholar
Gutman, I, Knop, JV, Trinajstić, N: A graph-theoretical analysis of the HOMO-LUMO separation in conjugated hydrocarbons. Z. Naturforsch. 29b, 80-82 (1974)
Google Scholar
Kiang, YS, Chen, ET: Evaluation of HOMO-LUMO separation and homologous linearity of conjugated molecules. Pure Appl. Chem. 55, 283-288 (1983)
View ArticleGoogle Scholar
Fowler, PW, Pisansky, T: HOMO-LUMO maps for fullerenes. MATCH Commun. Math. Comput. Chem. 64, 373-390 (2010)
MathSciNetMATHGoogle Scholar
Mohar, B: Median eigenvalues of bipartite planar graphs. MATCH Commun. Math. Comput. Chem. 70, 79-84 (2013)
MathSciNetMATHGoogle Scholar
Li, X, Li, Y, Shi, Y, Gutman, I: Note on the HOMO-LUMO index of graphs. MATCH Commun. Math. Comput. Chem. 70, 85-96 (2013)
MathSciNetMATHGoogle Scholar
Bianchi, M, Cornaro, A, Palacios, JL, Torriero, A: Bounding the sum of powers of normalized Laplacian eigenvalues of graphs through majorization methods. MATCH Commun. Math. Comput. Chem. 70(2), 707-716 (2013)
MathSciNetMATHGoogle Scholar
Bianchi, M, Cornaro, A, Palacios, JL, Torriero, A: New upper and lower bounds for the additive degree-Kirchhoff index. Croat. Chem. Acta 86(4), 363-370 (2013)
View ArticleMATHGoogle Scholar
Bianchi, M, Cornaro, A, Palacios, JL, Torriero, A: New bounds of degree-based topological indices for some classes of c-cyclic graphs. Discrete Appl. Math. 184, 62-75 (2015)
MathSciNetView ArticleMATHGoogle Scholar
Bianchi, M, Cornaro, A, Torriero, A: Majorization under constraints and bounds of the second Zagreb index. Math. Inequal. Appl. 16(2), 329-347 (2013)
MathSciNetMATHGoogle Scholar
Bianchi, M, Torriero, A: Some localization theorems using a majorization technique. J. Inequal. Appl. 5, 443-446 (2000)
MathSciNetMATHGoogle Scholar
Clemente, GP, Cornaro, A: Computing lower bounds for the Kirchhoff index via majorization techniques. MATCH Commun. Math. Comput. Chem. 73, 175-193 (2015)
MathSciNetGoogle Scholar
Cornaro, A, Clemente, GP: A new lower bound for the Kirchhoff index using a numerical procedure based on majorization techniques. Electron. Notes Discrete Math. 41, 383-390 (2013)
View ArticleGoogle Scholar
Clemente, GP, Cornaro, A: New bounds for the sum of powers of normalized Laplacian eigenvalues of graphs. Ars Math. Contemp. 11(2), 403-413 (2016)
Google Scholar
Clemente, GP, Cornaro, A: Novel bounds for the normalized Laplacian Estrada index and normalized Laplacian energy. MATCH Commun. Math. Comput. Chem. 77(3), 673-690 (2017)
Google Scholar
Bianchi, M, Cornaro, A, Torriero, A: A majorization method for localizing graph topological indices. Discrete Appl. Math. 161, 2731-2739 (2013)
MathSciNetView ArticleMATHGoogle Scholar
Gutman, I: The energy of a graph: old and new results. In: Betten, A, Kohnert, A, Laue, R, Wasserman, A (eds.) Algebraic Combinatorics and Applications, pp. 196-211 (2011)
Google Scholar
Marshall, AW, Olkin, I, Arnold, B: Inequalities: Theory of Majorization and Its Applications. Springer, Berlin (2011)
View ArticleMATHGoogle Scholar
Cavers, M, Fallat, S, Kirkland, S: On the normalized Laplacian energy and general Randić index \(R_{-1}\) of graphs. Linear Algebra Appl. 433, 172-190 (2010)
MathSciNetView ArticleMATHGoogle Scholar
Bianchi, M, Cornaro, A, Palacios, JL, Torriero, A: Bounds for the Kirchhoff index via majorization techniques. J. Math. Chem. 51(2), 569-587 (2013)
MathSciNetView ArticleMATHGoogle Scholar
Chung, FRK, Lu, L, Vu, V: The spectra of random graphs with given expected degrees. Proc. Natl. Acad. Sci. USA 100, 6313-6318 (2003)
MathSciNetView ArticleMATHGoogle Scholar
