Elsevier

Applied Mathematics and Computation

Volume 341, 15 January 2019, Pages 70-79
Applied Mathematics and Computation

A note on extremal trees with degree conditions

https://doi.org/10.1016/j.amc.2018.08.026Get rights and content

Abstract

A fundamental question in the study of graph invariants asks for the extremal structures under certain constraints that maximize or minimize a graph invariant. In this note, we summarize some recent work on the extremal trees of distance-based and degree-based graph invariants under various degree conditions. We note that many of such extremal structures turned out to be identical for different but similar invariants. Such common extremal structures are investigated through the greedy trees and majorization between degree sequences. We show that many of the known extremal results can be obtained through this line of arguments. We also introduce some new extremal results as immediate consequences.

Introduction

The study of the so-called chemical indices and the associated extremal problems has been a big part of research in chemical graph theory, where the molecular structures of a chemical compound are analyzed through various graph invariants/indices that were found to correlate to properties of these compounds. Among numerous indices introduced in recent years, we are particularly interested in the distance-based and degree-based indices. Because of many applications related to acyclic structures, extremal structures among trees appear to be of interest. Furthermore, as the vertex degrees correspond to the valences of atoms, trees under various degree conditions have been extensively studied.

After presenting the necessary background information in this section, we will first survey some recent extremal results on trees under degree conditions in Section 2. Of particular interest to us are trees of a given degree sequence. We will first introduce the basic definitions and extremal structures in this section. We will then present several main extremal results among trees of a given degree sequence and discuss their immediate consequences in Section 3. Among these consequences are many previously established extremal results as well as some new ones.

All graphs considered in this paper are finite and simple. For a graph G, we use V(G), E(G), |V(G)|, e(G), Δ(G), β, α, γ to denote the vertex set, edge set, number of vertices, number of edges, maximum degree, matching number, independent number, domination number of G, respectively. For vertices x, y ∈ V(G), we will use deg(x) to denote the degree of x and d(x, y) to denote the distance between x and y. The distance function dG(x) is the sum of distances between x and all other vertices of G. Given a graph G, the branch vertex is a vertex of degree at least 3 and the pendent vertex or leaf is a vertex of degree one. A segment of a tree T is a subpath whose end vertices are branch vertices or leaves, and all internal vertices have degree two. Given a graph G with n vertices, the non-increasing sequence π=(deg(v1),deg(v2),,deg(vn)) is called the degree sequence of G.

The most well known distance-based index is probably the Wiener index, defined as the sum of distances between all pairs of vertices,W(T)={u,v}V(T)d(u,v).Related to the Wiener index is the Wiener Polarity index, WP(T), defined as the number of unordered pairs of vertices of distance 3 in T. Both of them were first introduced in [90] in 1947. Some recent work can be found in [7], [8], [9], [44], [45], [46], [49], [66], [67], [91], [94].

For a tree T and any function f defined on the distance between two vertices of T, a general distance-based graph invariant with respect to f can be defined asWf(T)={u,v}V(T)f(d(u,v)).When f(x)=x, this is the classic Wiener index. When f(x)=1x, this is the Harary index introduced in 1993 and denoted by H(T) [41], [70]. In the same year the hyper-Wiener index of a tree T was proposed in [72] and defined as WW(T)=Wf(T) with f(x)=x+x22.

The terminal Wiener index [34] of a tree T was defined as the sum of distances between leaves. This is denoted by TW(T)={u,v}L(T)d(u,v) where L(T) is the set of pendent vertices of T. A good survey on this index can be found in [33].

Some recent work has also been done on the eccentricity sum of a tree T [78]. That is, Ecc(T)=vV(T)eccT(v) where eccT(v)=maxuV(T)d(u,v) is the eccentricity of a vertex v ∈ V(T).

The Laplacian polynomial P(T, λ) of a tree T is also related to distance-based indices. It is defined as the characteristic polynomial of its Laplacian matrix L(T)=D(T)A(T): P(T,λ)=det(λInL(T))=k=0n(1)kckλnk. It has been shown in [32] that the coefficients cn2 and cn3 are equal to its Wiener index W(T) and modified hyper-Wiener index WWW(T), respectively. Related work can be found in [24], [25], [39], [40], [64], [82].

The first degree-based index is probably the Randić index [71], defined as R(T)=uvE(T)(deg(u)deg(v))12. One may see summaries of recent results in the survey papers [47], [65]. The generalization of this concept was proposed by Bollobás and Erdős [2] in 1998 and defined as ωα(T)=uvE(T)(deg(u)deg(v))α for α ≠ 0, which is also called the general Randić index or the connectivity index [17]. In fact, Randić also introduced ω1(T) which is known today as the Modified Zagreb index. Other examples of generalization of Randić index can be found in [69], [77].

Given a bivariate symmetric function f defined on the degrees of adjacent vertices in T, the connectivity function associated with f [88] is a natural generalization of the above mentioned degree-based indices, defined asRf(T)=uvE(T)f(deg(u),deg(v)).

It is easy to see that Rf(T) represents various degree-based indices with different choices of f. For instance, the second Zagreb index is M2(T)=Rf(T) with f(x,y)=xy. Note that the first Zagreb index M1(T) is simply the sum of the squares of degrees [36], [37]. A generalization of M1 is called the Zeroth-order general Randić index, denoted by 0Rα(T)=uV(T)degα(u) for α ≠ 0 or 1 [50]. When f(x,y)=(x+y2xy)12, the function Rf(T) is known as the atom-bond connectivity (or simply the ABC index) ABC(T) [23]. Other examples of such degree-based graph invariants includes the sum-connectivity index [99] χ(T)=Rf(T) with f(x,y)=(x+y)12, the general sum-connectivity index [100] χα(T)=Rf(T) with f(x,y)=(x+y)α, the reformulated Zagreb index [97] with f(x,y)=(x+y2)2, and the harmonic index [98] with f(x,y)=2x+y.

Among trees with a given degree sequence the so-called greedy structures have been known to be extremal with respect to many distance-based and degree-based indices. We first list some related definitions.

Definition 1.1 (Greedy Trees)

Given the degree sequence (d1,d2,,dn) such that idi=2(n1), the greedy tree is obtained through the following “greedy algorithm”:

  • (i)

    Label the vertex with the largest degree as v (the root);

  • (ii)

    Label the neighbors of v as v1,v2,, assign the largest degrees available to them such that deg(v1) ≥ deg(v2) ≥ ⋅⋅⋅;

  • (iii)

    Label the neighbors of v1 (except v) as v11,v12, such that they take all the largest degrees available and that deg(v11) ≥ deg(v12) ≥ ⋅⋅⋅, then do the same for v2,v3,;

  • (iv)

    Repeat (iii) for all the newly labeled vertices, always start with the neighbors of the labeled vertex with largest degree whose neighbors are not labeled yet.

As an example, a greedy tree with degree sequence (4,4,4,3,3,3,3,3,2,2,1,,1) is shown in Fig. 1.

A caterpillar is a tree whose removal of leaves results in a path. Given a degree sequence the greedy-caterpillar is defined as the following.

Definition 1.2 (Greedy Caterpillar)

A greedy caterpillar is a caterpillar where the path formed by the internal vertices can be labeled as v1v2vk such thatmin{d1,dk}max{d2,dk1}min{d2,dk1}max{d3,dk2},where di is the degree of vertex vi.

In the study of many extremal problems it is convenient to compare the greedy trees of different degree sequences. For this purpose a partial ordering of the degree sequences of trees of fixed order needs to be introduced.

Definition 1.3 (Majorization)

Given two non-increasing degree sequences π and π′ with π=(d1,d2,,dn) and π=(d1,d2,,dn), we say that π′ majors π, denoted by ππ′, if the following conditions are met:

  • (1)

    i=0kdii=0kdi for 1kn1,

  • (2)

    i=0ndi=i=0ndi.

Section snippets

Survey of extremal trees with given degree conditions

In this section we survey some of the recent work on extremal trees under various degree conditions with respect to different graph indices. For convenience we let T(deg(v1),deg(v2),,deg(vn)) denote a tree with vertices v1,v2,,vn and degree sequence (deg(v1),deg(v2),,deg(vn)).

In the next a few statements, let BTn,b and STn,s be the set of all trees of order n with exactly b branch vertices and s segments, respectively.

Theorem 2.1

Let TBTn,b:

  • (a)

    ([56]) for b[n/21], the greedy caterpillar T(3,,3b,2,,2,1,

Consequences of Greedy trees and majorization between degree sequences

In this section we first provide some general extremal results among trees with a given degree sequence. Then greedy trees of different degree sequences are compared. Last but not least, these results are applied to various distanced-based and degree-based indices in trees under different degree conditions.

Concluding remark

We provided a brief introduction to distance-based and degree-based graph indices and the associated extremal problems in trees. Recent work on the extremal trees with various degree conditions is surveyed. In particular, greedy trees of a given degree sequence and the comparison between greedy trees of different degree sequences were shown to be very useful tools. Following several key results many extremal trees with degree conditions follow as immediate consequences. These consequences

Acknowledgments

This work is partially supported by the National Natural Science Foundation (61872200), Natural Science Foundation of Tianjin (17JCQNJC00300), the Major Science and Technology Program of Big Data and Cloud Computing of Tianjin (15ZXDSGX00020), the Science and Technology Commission of Tianjin Binhai New Area (BHXQKJXM-PT-ZJSHJ-2017005) and the National Key Research and Development Program of China (2016YFC0400709).

The authors are also in debt to the anonymous referees for their valuable

References (100)

  • L. Feng et al.

    Harary index and graph properties

    Discrete Appl. Math

    (2017)
  • M. Fischermann et al.

    Wiener index versus maximum degree in trees

    Discrete Appl. Math

    (2002)
  • M. Goubko et al.

    Degree-based topological indices: optimal trees with given number of pendents

    Appl. Math. Comput

    (2014)
  • I. Gutman et al.

    Graph theory and molecular orbitals

    Total π-electron energy of alternant hydrocarbons, Chem. Phys. Lett

    (1972)
  • A. Ilić et al.

    Laplacian coefficients of trees with given number of leaves or vertices of degree two

    Linear Algebra Appl

    (2009)
  • F. Jelen et al.

    Superdominance order and distance of trees with bounded maximum degree

    Discrete Appl. Math

    (2003)
  • Y. Lan et al.

    Vertex-based and edge-based centroids of graphs

    Appl. Math. Comput

    (2018)
  • T. Li et al.

    A comparative analysis of graph distance measures based on topological indices and graph edit distance

    Inf. Sci.

    (2017)
  • C. Li et al.

    Identification of influential spreaders based on classified neighbors in real-world complex networks

    Appl. Math. Comput.

    (2018)
  • W. Lin et al.

    On the minimal ABC index of trees with k leaves

    Discrete Appl. Math

    (2017)
  • B. Liu et al.

    On the wiener polarity index of trees with maximum degree or given number of leaves

    Comput. Math. Appl

    (2010)
  • H. Liu et al.

    Trees of extremal connectivity index

    Discrete Appl. Math

    (2006)
  • M. Liu et al.

    The second Zagreb indices of unicyclic graphs with given degree sequences

    Discrete Appl. Math.

    (2014)
  • C. Magnant et al.

    Which tree has the smallest ABC index among trees with k leaves?

    Discrete Appl. Math

    (2015)
  • M. Randić

    Novel molecular descriptor for structure-property studies

    Chem. Phys. Lett

    (1993)
  • D. Rautenbach

    A note on trees of maximum weight and restricted degrees

    Discrete Math

    (2003)
  • D. Rautenbach

    Average distance and domination number revisited

    Discrete Appl. Math

    (2011)
  • Y. Shi

    Note on two generalizations of the Randić index

    Appl. Math. Comput

    (2015)
  • H. Smith et al.

    Eccentricity sums in trees

    Discrete Appl. Math

    (2016)
  • G. Su et al.

    Graphs with fixed number of pendent vertices and minimal zeroth-order general Randić index

    Appl. Math. Comput

    (2015)
  • F. Tian et al.

    Average distances and distance domination numbers

    Discrete Appl. Math

    (2009)
  • H. Wang

    Extremal trees with given degree sequence for the Randić index

    Discrete Math

    (2008)
  • W. Yuan et al.

    The second Zagreb indices of graphs with given degree sequences

    Discrete Appl. Math

    (2015)
  • J. Yue et al.

    On the generalized wiener polarity index of trees with a given diameter

    Discrete Appl. Math

    (2018)
  • L. Zhong

    The harmonic index for graphs

    Appl. Math. Lett

    (2012)
  • E.O.D. Andriantiana et al.

    Maximum wiener index of trees with given segment sequence

    MATCH Commun. Math. Comput. Chem.

    (2016)
  • B. Bollobás et al.

    Graphs of extremal weights

    Ars Combin.

    (1998)
  • B. Borovićanin

    On the extremal Zagreb indices of trees with given number of segments or given number of branching vertices

    MATCH Commun. Math. Comput. Chem.

    (2015)
  • B. Borovićanin et al.

    On the maximum and minimum Zagreb indices of trees with a given number of vertices of maximum degree

    MATCH Commun. Math. Comput. Chem.

    (2015)
  • K. Burns et al.

    A graph-theretic view of the united states postal service

    Graph Theory Combinnatorics Algorithms

    (1995)
  • Y. Chen et al.

    On wiener and terminal wiener indices of trees

    MATCH Commun. Math. Comput. Chem

    (2013)
  • Y. Chen et al.

    The terminal wiener index of trees with diameter or maximum degree

    Ars Combbin

    (2015)
  • K. Das et al.

    On Zagreb and Harary indices

    MATCH Commun. Math. Comput. Chem

    (2013)
  • Z. Du et al.

    Minimum wiener indices of trees and unicyclic graphs of given matching number

    MATCH Common. Math. Comput. Chem

    (2010)
  • Z. Du et al.

    Minimum sum-connectivity indices of trees and unicyclic graphs of a given matching number

    J. Math. Chem

    (2010)
  • Z. Du et al.

    Sum-connectivity indices of trees and unicyclic graphs of fixed maximum degree

    Mathematics

    (2012)
  • R.C. Entringer

    Bounds for the average distance-inverse degree product in trees

    Combinatorics Graph Theory Algorithms

    (1999)
  • E. Estrada et al.

    An atom-bond connectivity index: Modelling the enthalpy of formation of alkanes

    Indian J. Chem

    (1998)
  • L. Feng et al.

    Spectral radius and k-connectedness of a graph

    Monatsh Math

    (2018)
  • L. Gan et al.

    The ABC index of trees with given degree sequence

    MATCH Commun. Math. Comput. Chem

    (2012)
  • Cited by (0)

    View full text