Next Article in Journal
EAMA: Efficient Adaptive Migration Algorithm for Cloud Data Centers (CDCs)
Next Article in Special Issue
Cactus Graphs with Maximal Multiplicative Sum Zagreb Index
Previous Article in Journal
Improved Bounds on Lorentz Symmetry Violation from High-Energy Astrophysical Sources
Previous Article in Special Issue
Topological Indices and f-Polynomials on Some Graph Products
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Bounds on the Arithmetic-Geometric Index

by
José M. Rodríguez
1,†,
José L. Sánchez
2,†,
José M. Sigarreta
2,*,† and
Eva Tourís
3,†
1
Departamento de Matemáticas, Universidad Carlos III de Madrid, Avenida de la Universidad 30, 28911 Madrid, Spain
2
Facultad de Matemáticas, Universidad Autónoma de Guerrero, Carlos E. Adame No.54 Col. Garita, 39650 Acalpulco, Mexico
3
Departamento de Matemáticas, Universidad Autónoma de Madrid, Ciudad Universitaria de Cantoblanco, 28049 Madrid, Spain
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work.
Symmetry 2021, 13(4), 689; https://doi.org/10.3390/sym13040689
Submission received: 19 March 2021 / Revised: 7 April 2021 / Accepted: 12 April 2021 / Published: 15 April 2021
(This article belongs to the Special Issue Analytical and Computational Properties of Topological Indices)

Abstract

:
The concept of arithmetic-geometric index was recently introduced in chemical graph theory, but it has proven to be useful from both a theoretical and practical point of view. The aim of this paper is to obtain new bounds of the arithmetic-geometric index and characterize the extremal graphs with respect to them. Several bounds are based on other indices, such as the second variable Zagreb index or the general atom-bond connectivity index), and some of them involve some parameters, such as the number of edges, the maximum degree, or the minimum degree of the graph. In most bounds, the graphs for which equality is attained are regular or biregular, or star graphs.

1. Introduction

In chemical graph theory, a topological descriptor is a function that associates each molecular graph with a real value. If it correlates well with some chemical property, then it is called a topological index. Since Winer’s work (see [1]), numerous topological indices have been defined and discussed, the growing interest in their study is because there are several applications in theoretical chemistry, especially in QSPR/QSAR research (see [2,3,4]).
In particular, vertex-degree-based topological indices belong to one of the largest and most studied classes of topological descriptors. The Randić index [5] and the Zagreb indices [6] are probably the best known such descriptors.
In [7,8,9], the first and second variable Zagreb indices are defined, for each α R , as
M 1 α ( G ) = u V ( G ) d u α , M 2 α ( G ) = u v E ( G ) ( d u d v ) α ,
where d u denotes the degree of u V ( G ) .
Note that, for α = 2 , α = 1 and α = 3 , the index M 1 α is the first Zagreb index M 1 , the inverse index I D , and the forgotten index F, respectively; also, for α = 1 , α = 1 / 2 and α = 1 , the index M 2 α is the second Zagreb index M 2 , the Randić index R and the modified Zagreb index.
The geometric-arithmetic index| G A is defined in [10] as
G A ( G ) = u v E ( G ) 2 d u d v d u + d v .
There are many papers studying the mathematical and computational properties of the G A index (see [10,11,12,13,14,15,16,17]).
In 2015, the arithmetic-geometric index [18] was defined as
A G ( G ) = u v E ( G ) d u + d v 2 d u d v .
The A G index of path graphs with pendant vertices attached was discussed in the papers [18,19]. Additionally, the arithmetic-geometric index of graphene, which is the most conductive and effective material for electromagnetic interference shielding, was computed in [20]. The paper [21] studied the spectrum and energy of arithmetic-geometric matrix, in which the sum of all elements is equal to 2 A G . Other bounds of the arithmetic-geometric energy appeared in [22,23]. The paper [24] studies extremal A G -graphs for various classes of simple graphs, and it includes inequalities that involve A G + G A , A G G A , A G · G A , and A G / G A . In [25,26,27,28], there are more bounds on the A G index and a discussion on the effect of deleting an edge from a graph on the arithmetic-geometric index.
Along the paper, we denote, by G, a simple graph without isolated vertices.
An important subject in the study of topological indices is to bind them in terms of some parameters. Reference [29] proves that many upper bounds of G A are not useful, and it shows the importance of obtaining upper bounds of G A less than the number of edges m. In a similar way, it is important to find lower bounds of A G that are greater than m. With this aim, in this paper we obtain several new lower bounds of A G , which are greater than m, and we characterize the extremal graphs.

2. Bounds Involving Other Indices

A graph is said biregular if it is bipartite and the degree of any vertex in one side of the bipartition is the maximum degree Δ and the degree of any vertex in the other side is the minimum degree δ .
One can check that the following result holds.
Lemma 1.
Let g be the function g ( x , y ) = x + y 2 x y with 0 < a x , y b . Then
1 g ( x , y ) a + b 2 a b .
The equality in the upper bound is attained if and only if either x = a and y = b , or x = b and y = a , and the equality in the lower bound is attained if and only if x = y .
The following inequalities follow from Lemma 1:
m A G ( G ) Δ + δ 2 Δ δ m .
The lower bound in (1) also follows from the inequalities G A ( G ) · A G ( G ) m 2 and G A ( G ) m , see [11],12]. The upper bound in (1) appears in [27].
The following result improves the lower bound in (1), see Remark 1.
Theorem 1.
If G is a graph with m edges, maximum degree Δ, and minimum degree δ, then
m + M 1 ( G ) 2 M 2 1 / 2 ( G ) 2 Δ A G ( G ) m + M 1 ( G ) 2 M 2 1 / 2 ( G ) 2 δ .
The equality in each bound is attained if and only if G is a regular graph.
Proof. 
We have
d u + d v 2 d u d v = 1 + d u d v 2 2 d u d v , A G ( G ) = m + u v E ( G ) d u d v 2 2 d u d v .
Because
u v E ( G ) d u d v 2 2 d u d v 1 2 Δ u v E ( G ) d u d v 2 = 1 2 Δ u v E ( G ) d u + d v 2 u v E ( G ) d u d v = M 1 ( G ) 2 M 2 1 / 2 ( G ) 2 Δ ,
we conclude
A G ( G ) m + M 1 ( G ) 2 M 2 1 / 2 ( G ) 2 Δ .
Because
u v E ( G ) d u d v 2 2 d u d v 1 2 δ u v E ( G ) d u d v 2 = 1 2 δ u v E ( G ) d u + d v 2 u v E ( G ) d u d v = M 1 ( G ) 2 M 2 1 / 2 ( G ) 2 δ ,
we conclude
A G ( G ) m + M 1 ( G ) 2 M 2 1 / 2 ( G ) 2 δ .
If G is regular, then both bounds are the same, and they are equal to A G ( G ) .
If the equality in some bound is attained, then we have either d u d v = Δ 2 for every u v E ( G ) or d u d v = δ 2 for every u v E ( G ) , so d u = Δ for every u V ( G ) or d u = δ for every u V ( G ) , and G is a regular graph. □
Remark 1.
Because Cauchy–Schwarz inequality gives
M 1 ( G ) 2 M 2 1 / 2 ( G ) = u v E ( G ) d u d v 2 = u v E ( G ) d u d v 2 1 m u v E ( G ) 1 2 1 m u v E ( G ) | d u d v | 2 ,
we have M 1 ( G ) 2 M 2 1 / 2 ( G ) 0 and, so, Theorem 1 improves the lower bound in (1).
The misbalance rodeg index [30] is
M R ( G ) = u v E ( G ) | d u d v | .
Theorem 1 and Remark 1 have the following consequence.
Corollary 1.
If G is a graph with m edges, maximum degree Δ, and minimum degree δ, then
m + M R ( G ) 2 2 Δ m A G ( G ) ,
and the equality is attained if and only if G is regular graph.
The following fact is elementary.
Lemma 2.
Let us consider the function f ( x , y ) = ( x y ) α with δ x , y Δ . Then
f ( x , y ) δ 2 α , if α 0 , f ( x , y ) Δ 2 α , if α 0 .
The following result provides bounds that relate the arithmetic-geometric and the second variable Zagreb indices.
Theorem 2.
If G is a graph with maximum degree Δ and minimum degree δ, and α R , then
A G ( G ) Δ δ 2 α 1 M 2 α ( G ) , i f α 1 / 2 , A G ( G ) Δ 2 α M 2 α ( G ) , i f α 1 / 2 ,
and the equality in each bound is attained for some fixed α if and only if G is regular.
Proof. 
We have
u v E ( G ) d u + d v 2 d u d v Δ u v E ( G ) ( d u d v ) α 1 / 2 ( d u d v ) α .
If α 1 / 2 , then Lemma 2 gives
A G ( G ) Δ δ 2 α 1 u v E ( G ) ( d u d v ) α .
If α 1 / 2 , then we have, by Lemma 2
A G ( G ) Δ 2 α u v E ( G ) ( d u d v ) α .
If G is regular, then A G ( G ) = m , M 2 α ( G ) = δ 2 α m = Δ 2 α m and Δ δ 2 α 1 = Δ 2 α , and the equality in each bound is attained.
If the equality is attained, then d u + d v = 2 Δ for every u v E ( G ) ; thus, d u = Δ for every u V ( G ) , and G is a regular graph. □
The symmetric division deg index
S D D ( G ) = u v E ( G ) d u 2 + d v 2 d u d v = u v E ( G ) d u d v + d v d u .
is another Adriatic index that appears in [30,31], see also [32].
We need the following inequality (see e.g., [14], Lemma 4) in the proof of Theorem 3 below.
Lemma 3.
Let ( X , μ ) be a measure space and f , g : X R measurable functions. If there exist positive constants ω , Ω with ω | g | | f | Ω | g | μ-a.e., then
f 2 g 2 1 2 Ω ω + ω Ω f g 1 .
If these norms are finite, the equality in the bound is attained if and only if ω = Ω and | f | = ω | g | μ-a.e. or f = g = 0 μ-a.e.
We have the following direct consequence.
Corollary 2.
If a j , b j 0 and ω b j a j Ω b j for 1 j m , then
j = 1 m a j 2 1 / 2 j = 1 m b j 2 1 / 2 1 2 Ω ω + ω Ω j = 1 m a j b j .
If a j > 0 for some 1 j m , then the equality holds if and only if ω = Ω and a j = ω b j for every 1 j m .
The following result provides an inequality relating the arithmetic-geometric and the symmetric division deg indices.
Theorem 3.
Let G be a graph with m edges, maximum degree Δ, and minimum degree δ. Subsequently,
2 Δ δ ( Δ + δ ) Δ + δ 2 m S D D ( G ) + 2 m A G ( G ) 1 2 m S D D ( G ) + 2 m .
The equality in the lower bound is attained if and only if G is a regular graph. The equality in the upper bound is attained if G is a regular or biregular graph.
Proof. 
Let us consider
a j : = d u + d v 2 d u d v , b j : = 1 .
We have, by Corollary 1,
1 a j b j Δ + δ 2 Δ δ .
Thus, Corollary 2 gives
u v E ( G ) 1 u v E ( G ) ( d u + d v ) 2 4 d u d v 1 4 Δ + δ 2 Δ δ + 2 Δ δ Δ + δ 2 u v E ( G ) d u + d v 2 d u d v 2 = 1 4 Δ + δ 2 2 Δ δ ( Δ + δ ) 2 A G ( G ) 2 .
Because
u v E ( G ) 1 = m , u v E ( G ) ( d u + d v ) 2 4 d u d v = 1 4 u v E ( G ) d u 2 + d v 2 d u d v + u v E ( G ) 1 2 = 1 4 S D D ( G ) + 1 2 m ,
we conclude
m 4 S D D ( G ) + 2 m 1 4 Δ + δ 2 2 Δ δ ( Δ + δ ) 2 A G ( G ) 2 , A G ( G ) 2 Δ δ ( Δ + δ ) Δ + δ 2 m S D D ( G ) + 2 m .
If the equality in this bound is attained, then Corollary 2 gives
1 = Δ + δ 2 Δ δ .
Thus, Corollary 1 gives Δ = δ , and, so, G is regular.
If G is regular, then
2 Δ δ ( Δ + δ ) Δ + δ 2 m S D D ( G ) + 2 m = 2 δ 2 δ 4 δ m 2 m + 2 m = m = A G ( G ) .
On the other hand, the Cauchy–Schwarz inequality gives
A G ( G ) 2 = u v E ( G ) d u + d v 2 d u d v 2 u v E ( G ) 1 u v E ( G ) ( d u + d v ) 2 4 d u d v .
Because
u v E ( G ) 1 = m , u v E ( G ) ( d u + d v ) 2 4 d u d v = 1 4 S D D ( G ) + 1 2 m ,
we conclude
A G ( G ) 2 m 4 S D D ( G ) + 2 m .
If G is regular or biregular, then
1 2 m S D D ( G ) + 2 m = 1 2 m Δ δ + δ Δ m + 2 m = m 2 Δ 2 + δ 2 + 2 Δ δ Δ δ = Δ + δ 2 Δ δ m = A G ( G ) .
The atom-bond connectivity index [33] is
A B C ( G ) = u v E ( G ) d u + d v 2 d u d v .
Furtula et al. [34] made a generalization of A B C index, defined as
A B C α ( G ) = u v E ( G ) d u + d v 2 d u d v α , where α R .
They showed that the A B C α defined in this way, for α = 3 , has better predictive power than the original A B C index.
The three following results relate the arithmetic-geometric and the general atom-bond connectivity indices.
Theorem 4.
Let G be a graph with maximum degree Δ and without isolated edges, and α > 0 . Then
A G ( G ) ( Δ 1 ) α ( Δ + 1 ) 2 Δ α + 1 2 A B C α ( G ) ,
and the equality in the inequality holds if and only if G is a union of stars S Δ + 1 .
Proof. 
Note that ( d u , d v ) ( 1 , 1 ) , since G does not have isolated edges, hence Δ 2 . First of all, we are going to compute the minimum value of
W ( x , y ) = x + y 2 x y α 2 x y x + y = 2 ( x + y 2 ) α ( x + y ) 1 x α + 1 2 y α + 1 2
on { 1 x y , 2 y Δ } . We have
W x = 2 y α + 1 2 α ( x + y 2 ) α 1 ( x + y ) 1 x α + 1 2 ( x + y 2 ) α ( x + y ) 2 x α + 1 2 + α + 1 2 ( x + y 2 ) α ( x + y ) 1 x α 1 2 = 2 y α + 1 2 x α 1 2 ( x + y 2 ) α 1 ( x + y ) 2 α ( x + y ) x ( x + y 2 ) x + α + 1 2 ( x + y 2 ) ( x + y ) = 2 y α + 1 2 x α 1 2 ( x + y 2 ) α 1 ( x + y ) 2 α ( x + y ) ( x + y 2 x ) + ( x + y 2 ) x + y 2 x = 2 y α + 1 2 x α 1 2 ( x + y 2 ) α 1 ( x + y ) 2 α ( x + y ) ( y 2 ) + 1 2 ( x + y 2 ) ( y x ) 0 ,
so, W ( x , y ) is strictly increasing on x [ 1 , y ] for every fixed y 2 and, so, W ( 1 , y ) W ( x , y ) . Consider
a ( y ) = W ( 1 , y ) = 2 ( y 1 ) α ( 1 + y ) 1 y α + 1 / 2 .
Subsequently,
a ( y ) = 2 α ( y 1 ) α 1 ( y + 1 ) 1 y α + 1 2 ( y 1 ) α ( y + 1 ) 2 y α + 1 2 + α + 1 2 ( y 1 ) α ( y + 1 ) 1 y α 1 2 = 2 ( y 1 ) α 1 ( y + 1 ) 2 y α 1 2 α ( y + 1 ) y ( y 1 ) y + α + 1 2 ( y 1 ) ( y + 1 ) = 2 ( y 1 ) α 1 ( y + 1 ) 2 y α 1 2 α ( y + 1 ) ( y 1 y ) + ( y 1 ) y + 1 2 y = 2 ( y 1 ) α 1 ( y + 1 ) 2 y α 1 2 α ( y + 1 ) 1 2 ( y 1 ) 2 < 0 ,
so, w is strictly decreasing on y [ 2 , Δ ] . Thus, we have a ( Δ ) a ( y ) = W ( 1 , y ) W ( x , y ) for every 1 x y , 2 y Δ and the equalities hold if and only if x = 1 and y = Δ . Therefore,
2 Δ α + 1 2 ( Δ 1 ) α ( Δ + 1 ) d u + d v 2 d u d v d u + d v 2 d u d v α for every u v E ( G ) ,
and the equality is attained if and only if d u = 1 and d v = Δ or vice versa for each edge u v E ( G ) , i.e., every connected component of G is a star S Δ + 1 . □
Remark 2.
The argument in the proof of Theorem 4 (with the same hypotheses) allows for obtaining the following lower bound of A G , but it is elementary:
( 2 Δ 2 ) α Δ 2 α A B C α ( G ) A G ( G ) ,
and the equality in the inequality holds if and only if G is regular.
We can improve Theorem 4 when δ 2 .
Theorem 5.
Let G be a graph with maximum degree Δ and minimum degree δ 2 , and α > 0 . Afterwards,
A G ( G ) max ( 2 δ 2 ) α δ 2 α , ( Δ + δ 2 ) α ( Δ + δ ) 2 ( Δ δ ) α + 1 2 A B C α ( G ) .
The equality in the inequality holds if G is regular.
Proof. 
Consider the notation in the proof of Theorem 4, and the function
c ( y ) = W ( δ , y ) = 2 δ α + 1 2 ( y + δ 2 ) α ( y + δ ) 1 y α + 1 2 ,
with 2 δ y Δ . The argument in the proof of Theorem 4 gives that c ( y ) = W ( δ , y ) W ( x , y ) for every δ x y Δ .
We have
c ( y ) = 2 δ α + 1 2 α ( y + δ 2 ) α 1 ( y + δ ) 1 y α + 1 2 ( y + δ 2 ) α ( y + δ ) 2 y α + 1 2 + α + 1 2 ( y + δ 2 ) α ( y + δ ) 1 y α 1 2 = 2 δ α + 1 2 ( y + δ 2 ) α 1 ( y + δ ) 2 y α 1 2 [ α ( y + δ ) y ( y + δ 2 ) y + α + 1 2 ( y + δ 2 ) ( y + δ ) = 2 δ α + 1 2 ( y + δ 2 ) α 1 ( y + δ ) 2 y α 1 2 [ α ( y + δ ) ( y + y + δ 2 ) + ( y + δ 2 ) y + y + δ 2 = 2 δ α + 1 2 ( y + δ 2 ) α 1 ( y + δ ) 2 y α 1 2 α ( y + δ ) ( δ 2 ) 1 2 ( y + δ 2 ) ( y δ ) .
Consider first the case δ = 2 . We have
c ( y ) = 2 δ α + 1 2 ( y + δ 2 ) α 1 ( y + δ ) 2 y α 1 2 α ( y + δ ) ( δ 2 ) 1 2 ( y + δ 2 ) ( y δ ) = δ α + 1 2 ( y + δ 2 ) α ( y + δ ) 2 y α 1 2 ( y δ ) 0 .
Thus, min y [ δ , Δ ] c ( y ) = c ( Δ ) .
Now, assume that δ 3 . Let us consider the second degree polynomial
P ( y ) = α ( y + δ ) ( δ 2 ) 1 2 ( y + δ 2 ) ( y δ ) .
Because
P ( 0 ) = α δ ( δ 2 ) 1 2 ( δ 2 ) ( δ ) = α + 1 2 δ ( δ 2 ) 0 ,
there exists at least a non-positive zero of P. Hence, there exists at most a zero of P in the interval [ δ , Δ ] . Additionally, P ( δ ) = 2 δ ( δ 2 ) > 0 .
Thus, there exists, at most, a zero of c in the interval [ δ , Δ ] and c ( δ ) > 0 . Consequently,
min y [ δ , Δ ] c ( y ) = min c ( δ ) , c ( Δ ) ,
for every δ 3 and, so, for every δ 2 . Therefore,
W ( x , y ) W ( δ , y ) c ( y ) min c ( δ ) , c ( Δ ) = min δ 2 α ( 2 δ 2 ) α , 2 ( Δ δ ) α + 1 2 ( Δ + δ 2 ) α ( Δ + δ ) 1 ,
for every δ x y Δ and, by symmetry, for every δ x , y Δ . Consequently,
min δ 2 α ( 2 δ 2 ) α , 2 ( Δ δ ) α + 1 2 ( Δ + δ 2 ) α ( Δ + δ ) d u + d v 2 d u d v d u + d v 2 d u d v α
for every u v E ( G ) , and
A G ( G ) max ( 2 δ 2 ) α δ 2 α , ( Δ + δ 2 ) α ( Δ + δ ) 2 ( Δ δ ) α + 1 2 A B C α ( G ) .
If G is regular, thus Δ = δ and
max ( 2 δ 2 ) α δ 2 α , ( Δ + δ 2 ) α ( Δ + δ ) 2 ( Δ δ ) α + 1 2 A B C α ( G ) max ( 2 δ 2 ) α δ 2 α , ( 2 δ 2 ) α 2 δ 2 δ 2 α + 1 δ 2 α ( 2 δ 2 ) α m = m = A G ( G ) ,
and the equality in the inequality holds. □
Now, we relate the arithmetic-geometric and general atom-bond connectivity indices with parameter greater than or equal to 1 / 2 .
Theorem 6.
If G is a graph with minimum degree δ 2 and maximum degree Δ, and β 1 / 2 , then
A G ( G ) Δ 2 2 Δ 2 β A B C β ( G ) ,
and the equality in the inequality is attained if and only if G is regular.
Proof. 
Define α = β 1 / 2 . As in the proof of Theorem 4, let us consider
W ( x , y ) = x + y 2 x y α 2 x y x + y = 2 ( x + y 2 ) α ( x + y ) 1 x α + 1 2 y α + 1 2
on { 2 δ x y Δ } . We have
W x = 2 y α + 1 2 x α 1 2 ( x + y 2 ) α 1 ( x + y ) 2 α ( x + y ) ( y 2 ) + 1 2 ( x + y 2 ) ( y x ) 2 y α + 1 2 x α 1 2 ( x + y 2 ) α 1 ( x + y ) 2 1 2 ( x + y ) ( y 2 ) + 1 2 ( x + y 2 ) ( y x ) = 2 y α + 1 2 x α 1 2 ( x + y 2 ) α 1 ( x + y ) 2 1 2 ( x 2 ) ( y x ) 0 ,
on { δ x y Δ } . Hence, W ( y , y ) W ( x , y ) when δ x y Δ . We define now
b ( y ) = W ( y , y ) = 2 α y 1 y 2 α .
Subsequently,
b ( y ) = α 2 α y 1 y 2 α 1 y 2 y 3 0 .
Consequently, b is a strictly decreasing function on δ y Δ , and
W ( Δ , Δ ) = b ( Δ ) b ( y ) = W ( y , y ) W ( x , y )
when δ x y Δ . Hence, by symmetry,
2 Δ 2 Δ 2 β = W ( Δ , Δ ) W ( x , y )
for each δ x , y Δ , and
2 Δ 2 Δ 2 β d u + d v 2 d u d v d u + d v 2 d u d v β for every u v E ( G ) ,
2 Δ 2 Δ 2 β A G ( G ) A B C β ( G ) .
Remark 3.
The arguments in the proof of Theorem 6 (with the same hypotheses) allow to obtain the following lower bound of A G , but it is elementary:
δ 2 2 δ 2 β A B C β ( G ) A G ( G ) ,
and the equality in the inequality holds if and only if G is regular.

3. General Bounds on the AG Index

In this section we obtain additional lower bounds of A G improving the lower bound in (1), which do not involve other topological indices. The two following bounds involve m and the minimum degree.
Theorem 7.
If G is a graph with m edges, minimum degree δ, maximum degree δ + 1 , and α is the number of edges u v with d u d v , then α is an even integer and
A G ( G ) = m + α 2 δ + 1 2 δ ( δ + 1 ) 1 .
Proof. 
Let D = u v E ( G ) : d u d v , then α is the cardinality of D. Because δ is the minimum degree of G and δ + 1 is its maximum degree, if u v D , then d u = δ and d v = δ + 1 or vice versa and, therefore,
d u + d v 2 d u d v = 2 δ + 1 2 δ ( δ + 1 ) .
If u v D c , then d u = d v = δ or d u = d v = δ + 1 , and therefore
d u + d v 2 d u d v = 1 .
Because there are exactly α edges in D and m α edges in D c , we have
A G ( G ) = u v E ( G ) d u + d v 2 d u d v = u v D c d u + d v 2 d u d v + u v D d u + d v 2 d u d v = u v D c 1 + u v D 2 δ + 1 2 δ ( δ + 1 ) = m α + α 2 δ + 1 2 δ ( δ + 1 ) .
Assume, for contradiction, that α is an odd integer.
Let G 1 be a subgraph of G induced by the n 1 vertices with degree δ in V ( G ) , and denote by m 1 the number of edges of G 1 . Handshaking Lemma gives n 1 δ α = 2 m 1 . Because α is an odd integer, δ is also an odd integer. Thus, δ + 1 is an even integer.
Let G 2 be a subgraph of G that is induced by the n 2 vertices with degree δ + 1 in V ( G ) , and denote, by m 2 , the number of edges of G 2 . Handshaking Lemma gives n 2 ( δ + 1 ) α = 2 m 2 , a contradiction, since α is an odd integer and δ + 1 is an even integer.
Thus, we conclude that α is an even integer. □
Theorem 8.
If G is a connected graph with m edges, minimum degree δ and maximum degree δ + 1 , then
A G ( G ) m + 2 δ + 1 δ ( δ + 1 ) 2 ,
and the equality is attained for each δ.
Proof. 
Let α be the number of edges u v E ( G ) with d u d v . Theorem 7 gives that α is an even integer. Because G is a connected graph, we have α 0 and so, α 2 . Since
2 δ + 1 2 δ ( δ + 1 ) > 1
and α 2 , Theorem 7 gives
A G ( G ) = m + α 2 δ + 1 2 δ ( δ + 1 ) 1 m + 2 2 δ + 1 2 δ ( δ + 1 ) 1 = m 2 + 2 δ + 1 δ ( δ + 1 ) .
Given a fixed δ , let us consider the complete graphs K δ + 1 and K δ + 2 with δ + 1 and δ + 2 vertices, respectively. Fix u 1 , u 2 V ( K δ + 1 ) and v 1 , v 2 V ( K δ + 2 ) , and denote by K δ + 1 and K δ + 2 the graphs obtained from K δ + 1 and K δ + 2 by deleting the edges u 1 u 2 and v 1 v 2 , respectively. Let Γ δ be the graph with V ( Γ δ ) = V ( K δ + 1 ) V ( K δ + 2 ) and E ( Γ δ ) = E ( K δ + 1 ) E ( K δ + 2 ) { u 1 v 1 } { u 2 v 2 } . Thus, Γ δ has δ 2 + 2 δ + 1 edges, minimum degree δ , maximum degree δ + 1 , and Theorem 7 give
A G ( Γ δ ) = δ 2 + 2 δ 1 + 2 δ + 1 δ ( δ + 1 ) .
Hence, the equality is attained for each δ . □
A chemical graph is a graph with Δ 4 .
Corollary 3.
If G is a connected chemical graph with m edges, minimum degree δ, and maximum degree δ + 1 , then Then
A G ( G ) m 2 + 7 3 6 .
Furthermore, the equality in the bound is attained.
Proof. 
Because G is a chemical graph, we have 1 δ 3 . Since
min 1 δ 3 2 δ + 1 δ ( δ + 1 ) = min 3 2 , 5 6 , 7 12 = 7 3 6 ,
Theorem 8 gives the desired inequality.
The graph Γ 3 in the proof of Theorem 8 provides that the equality is attained. □
We need some definitions. Let G be a graph with maximum degree Δ and minimum degree δ < Δ 1 . We denote, by α 0 , α 1 , α 2 , the cardinality of the subsets of edges
A 0 = { u v E ( G ) : d u = δ , d v = Δ } , A 1 = { u v E ( G ) : d u = δ , δ < d v < Δ } , A 2 = { u v E ( G ) : d u = Δ , δ < d v < Δ } ,
respectively.
We need the following result ([28], Theorem 5).
Lemma 4.
If G is a graph with m edges, maximum degree Δ, and minimum degree δ < Δ 1 , then
A G ( G ) Δ + δ 2 Δ δ m α 1 Δ + δ 2 Δ δ δ + Δ 1 2 δ ( Δ 1 ) α 2 Δ + δ 2 Δ δ Δ + δ + 1 2 Δ ( δ + 1 ) ,
A G ( G ) m + α 0 Δ + δ 2 Δ δ 1 + α 1 2 δ + 1 2 δ ( δ + 1 ) 1 + α 2 2 Δ 1 2 Δ ( Δ 1 ) 1 .
We are going to use Lemma 4 to obtain the following lower bound of A G involving m and the minimum and maximum degree.
Theorem 9.
Let G be a connected graph with m edges, maximum degree Δ and minimum degree δ < Δ 1 . Subsequently,
A G ( G ) m + min 2 δ + 1 2 δ ( δ + 1 ) + 2 Δ 1 2 Δ ( Δ 1 ) 2 , Δ + δ 2 Δ δ 1 .
The equality in the bound is attained.
Proof. 
Because G is connected, we have two possibilities: A 0 , or A 1 and A 2 .
In the first case, α 0 1 and, since
d u + d v 2 d u d v 1 ,
Lemma 4 gives
A G ( G ) m + α 0 Δ + δ 2 Δ δ 1 + α 1 2 δ + 1 2 δ ( δ + 1 ) 1 + α 2 2 Δ 1 2 Δ ( Δ 1 ) 1 m + Δ + δ 2 Δ δ 1 .
In the second case, α 1 , α 2 1 and Lemma 4 give
A G ( G ) m + α 0 Δ + δ 2 Δ δ 1 + α 1 2 δ + 1 2 δ ( δ + 1 ) 1 + α 2 2 Δ 1 2 Δ ( Δ 1 ) 1 m + 2 δ + 1 2 δ ( δ + 1 ) + 2 Δ 1 2 Δ ( Δ 1 ) 2 .
Let G be the graph in the figure.
Symmetry 13 00689 i001
We have m = 12 , Δ = 3 , δ = 1 , A 0 = , α 0 = 0 , A 1 = { u 2 u 3 } , α 1 = 1 , A 2 = { u 1 u 2 } and α 2 = 1 . Additionally, if u v A 0 A 1 A 2 , then d u = d v . Thus,
A G ( G ) = u v E ( G ) \ A 0 A 1 A 2 d u + d v 2 d u d v + u v A 0 Δ + δ 2 Δ δ + u v A 1 δ + d v 2 δ d v + u v A 2 Δ + d v 2 Δ d v = 10 + 2 δ + 1 2 δ ( δ + 1 ) + 2 Δ 1 2 Δ ( Δ 1 ) = 10 + 3 2 2 + 5 2 6 12.0813
The lower bound is
m + min 2 δ + 1 2 δ ( δ + 1 ) + 2 Δ 1 2 Δ ( Δ 1 ) 2 , Δ + δ 2 Δ δ 1 12 + min 3 2 2 + 5 2 6 2 , 2 3 1 12 + min { 0.0813 , 0.1547 } = 12.0813
and so, this graph attains the equality in the inequality. □

4. Conclusions

Topological indices have become a useful tool for the study of theoretical and practical problems in different areas of science. An important line of research that is associated with topological indices is to find optimal bounds and relations between known topological indices. In particular, to obtain bounds for the topological indices that are associated with invariant parameters of a graph.
We have the following nine results for the arithmetic-geometric index A G :
  • An upper and lower bound of A G based on the first and second variable Zagreb indices (Theorem 1).
  • An upper bound of A G that is based on the second variable Zagreb index M 2 a (Theorem 2).
  • An upper and lower bound of A G based on S D D (Theorem 3).
  • An upper bound of A G based on the general atom-bond connectivity index A B C a (Theorem 4).
  • Another upper bound of A G based on the general atom-bond connectivity index A B C a for graphs with minimum degree δ 2 (Theorem 5).
  • A further upper bound of A G based on the general atom-bond connectivity index A B C a for graphs with minimum degree δ 2 (Theorem 6).
  • An exact formula of A G based on the number of edges m and the minimum degree δ if the maximum degree is δ + 1 (Theorem 7).
  • A lower bound of A G based on the number of edges m and the minimum degree δ if the maximum degree is δ + 1 (Theorem 8). We provide a family of graphs for which the equality is attained.
  • A lower bound of A G that is based on the number of edges m, the minimum degree δ , and the maximum degree Δ (Theorem 9). We provide a graph for which the equality is attained.
Because the arithmetic-geometric index is useful from a practical point of view, to know extremal graphs for each bound involving this index allows for detecting chemical compounds that could satisfy desirable properties. Hence, these extremal graphs should correspond to molecules with a extremal value of a desired property correlated well with this index.
In the case of centrality indices, the generalization of degree has turned out to be a useful approach: the role of a more interconnected node can differ from a node that is connected to nodes having a lower degree [35]. We would like to purpose as a direction for future research to study similar problems for centrality indices.

Author Contributions

Investigation, J.M.R., J.L.S., J.M.S. and E.T. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by a grant from Agencia Estatal de Investigación (PID2019-106433GBI00/ AEI/10.13039/501100011033), Spain. The research of José M. Rodríguez was supported by the Madrid Government (Comunidad de Madrid-Spain) under the Multiannual Agreement with UC3M in the line of Excellence of University Professors (EPUC3M23), and in the context of the V PRICIT (Regional Programme of Research and Technological Innovation).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Acknowledgments

We would like to thank the reviewers by their careful reading of the manuscript and their suggestions which have improved this work.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Wiener, H. Structural determination of paraffin boiling points. J. Am. Chem. Soc. 1947, 69, 17–20. [Google Scholar] [CrossRef]
  2. Devillers, J.; Balaban, A.T. (Eds.) Topological Indices and Related Descriptors in QSAR and QSPR; Gordon and Breach: Amsterdam, The Netherlands, 1999. [Google Scholar]
  3. Karelson, M. Molecular Descriptors in QSAR/QSPR; Wiley-Interscience: New York, NY, USA, 2000. [Google Scholar]
  4. Todeschini, R.; Consonni, V. Handbook of Molecular Descriptors; Wiley-VCH: Weinheim, Germany, 2000. [Google Scholar]
  5. Randić, M. On characterization of molecular branching. J. Am. Chem. Soc. 1975, 97, 6609–6615. [Google Scholar] [CrossRef]
  6. Gutman, I.; Trinajstić, N. Graph theory and molecular orbitals. Total π–electron energy of alternant hydrocarbons. Chem. Phys. Lett. 1972, 17, 535–538. [Google Scholar] [CrossRef]
  7. Li, X.; Zheng, J. A unified approach to the extremal trees for different indices. MATCH Commun. Math. Comput. Chem. 2005, 54, 195–208. [Google Scholar]
  8. Li, X.; Zhao, H. Trees with the first smallest and largest generalized topological indices. MATCH Commun. Math. Comput. Chem. 2004, 50, 57–62. [Google Scholar]
  9. Miličević, A.; Nikolić, S. On variable Zagreb indices. Croat. Chem. Acta 2004, 77, 97–101. [Google Scholar]
  10. Vukičević, D.; Furtula, B. Topological index based on the ratios of geometrical and arithmetical means of end-vertex degrees of edges. J. Math. Chem. 2009, 46, 1369–1376. [Google Scholar] [CrossRef]
  11. Das, K.C. On geometric-arithmetic index of graphs. MATCH Commun. Math. Comput. Chem. 2010, 64, 619–630. [Google Scholar] [CrossRef] [Green Version]
  12. Das, K.C.; Gutman, I.; Furtula, B. Survey on Geometric-Arithmetic Indices of Graphs. MATCH Commun. Math. Comput. Chem. 2011, 65, 595–644. [Google Scholar]
  13. Das, K.C.; Gutman, I.; Furtula, B. On first geometric-arithmetic index of graphs. Discrete Appl. Math. 2011, 159, 2030–2037. [Google Scholar] [CrossRef] [Green Version]
  14. Martínez-Pérez, A.; Rodríguez, J.M.; Sigarreta, J.M. A new approximation to the geometric-arithmetic index. J. Math. Chem. 2018, 56, 1865–1883. [Google Scholar] [CrossRef]
  15. Mogharrab, M.; Fath-Tabar, G.H. Some bounds on GA1 index of graphs. MATCH Commun. Math. Comput. Chem. 2010, 65, 33–38. [Google Scholar]
  16. Rodríguez, J.M.; Sigarreta, J.M. Spectral properties of geometric-arithmetic index. Appl. Math. Comput. 2016, 277, 142–153. [Google Scholar] [CrossRef]
  17. Sigarreta, J.M. Bounds for the geometric-arithmetic index of a graph. Miskolc Math. Notes 2015, 16, 1199–1212. [Google Scholar] [CrossRef] [Green Version]
  18. Shegehall, V.S.; Kanabur, R. Arithmetic-geometric indices of path graph. J. Math. Comput. Sci. 2015, 16, 19–24. [Google Scholar]
  19. Shegehall, V.S.; Kanabur, R. Arithmetic-geometric indices of graphs with pendant vertices attached to the middle vertices of path. J. Math. Comput. Sci. 2015, 6, 67–72. [Google Scholar]
  20. Shegehall, V.S.; Kanabur, R. Computation of new degree-based topological indices of graphene. J. Math. 2016, 2016, 4341919. [Google Scholar] [CrossRef]
  21. Zheng, L.; Tian, G.-X.; Cui, S.-Y. On spectral radius and energy of arithmetic-geometric matrix of graphs. MACTH Commun. Math. Comput. Chem. 2020, 83, 635–650. [Google Scholar]
  22. Guo, X.; Gao, Y. Arithmetic-geometric spectral radius and energy of graphs. MACTH Commun. Math. Comput. Chem. 2020, 83, 651–660. [Google Scholar]
  23. Das, K.C.; Gutman, I. Degree-based energies of graphs. Linear Algebra Appl. 2018, 554, 185–204. [Google Scholar] [CrossRef]
  24. Vujošević, S.; Popivoda, G.; Vukićević, Ž.K.; Furtula, B.; Škrekovski, R. Arithmetic-geometric index and its relations with geometric-arithmetic index. Appl. Math. Comput. 2021, 391, 125706. [Google Scholar] [CrossRef]
  25. Carballosa, W.; Granados, A.; Méndez-Bermúdez, J.A.; Pestana, D.; Portilla, A. Computational properties of the arithmetic-geometric index. 2021. submitted. [Google Scholar]
  26. Cui, S.-Y.; Wang, W.; Tian, G.-X.; Wu, B. On the arithmetic-geometric index of graphs. MATCH Commun. Math. Comput. Chem. 2021, 85, 87–107. [Google Scholar]
  27. Milovanović, I.Ž.; Matejić, M.; Milovanović, E.I. Upper bounds for arithmetic-geometric index of graphs. Sci. Publ. State Univ. Novi Pazar Ser A Appl. Math. Inform. Mech. 2018, 10, 49–54. [Google Scholar] [CrossRef] [Green Version]
  28. Molina, E.; Rodríguez, J.M.; Sánchez, J.L.; Sigarreta, J.M. Inequalities on the arithmetic-geometric index. 2021. submitted. [Google Scholar]
  29. Milovanović, I.Ž.; Milovanović, E.I.; Matejić, M.M. On Upper Bounds for the Geometric–Arithmetic Topological Index. MATCH Commun. Math. Comput. Chem. 2018, 80, 109–127. [Google Scholar]
  30. Vukičević, D.; Gašperov, M. Bond Additive Modeling 1. Adriatic Indices. Croat. Chem. Acta 2010, 83, 243–260. [Google Scholar]
  31. Vukičević, D. Bond additive modeling 2. Mathematical properties of max-min rodeg index. Croat. Chem. Acta 2010, 83, 261–273. [Google Scholar]
  32. Furtula, B.; Das, K.C.; Gutman, I. Comparative analysis of symmetric division deg index as potentially useful molecular descriptor. Int. J. Quantum Chem. 2018, 118, e25659. [Google Scholar] [CrossRef]
  33. Estrada, E.; Torres, L.; Rodríguez, L.; Gutman, I. An atom-bond connectivity index. Modelling the enthalpy of formation of alkanes. Indian J. Chem. 1998, 37A, 849–855. [Google Scholar]
  34. Furtula, B.; Graovac, A.; Vukičević, D. Augmented Zagreb index. J. Math. Chem. 2010, 48, 370–380. [Google Scholar] [CrossRef]
  35. Csató, L. Measuring centrality by a generalization of degree. Central Europ. J. Oper. Res. 2017, 25, 771–790. [Google Scholar] [CrossRef] [Green Version]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Rodríguez, J.M.; Sánchez, J.L.; Sigarreta, J.M.; Tourís, E. Bounds on the Arithmetic-Geometric Index. Symmetry 2021, 13, 689. https://doi.org/10.3390/sym13040689

AMA Style

Rodríguez JM, Sánchez JL, Sigarreta JM, Tourís E. Bounds on the Arithmetic-Geometric Index. Symmetry. 2021; 13(4):689. https://doi.org/10.3390/sym13040689

Chicago/Turabian Style

Rodríguez, José M., José L. Sánchez, José M. Sigarreta, and Eva Tourís. 2021. "Bounds on the Arithmetic-Geometric Index" Symmetry 13, no. 4: 689. https://doi.org/10.3390/sym13040689

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop