Abstract
The eccentric connectivity index of a graph G is \(\xi ^c(G) = \sum _{v \in V(G)}\varepsilon (v)\deg (v)\), and the eccentric distance sum is \(\xi ^d(G) = \sum _{v \in V(G)}\varepsilon (v)D(v)\), where \(\varepsilon (v)\) is the eccentricity of v, and D(v) the sum of distances between v and the other vertices. A lower and an upper bound on \(\xi ^d(G) - \xi ^c(G)\) is given for an arbitrary graph G. Regular graphs with diameter at most 2 and joins of cocktail-party graphs with complete graphs form the graphs that attain the two equalities, respectively. Sharp lower and upper bounds on \(\xi ^d(T) - \xi ^c(T)\) are given for arbitrary trees. Sharp lower and upper bounds on \(\xi ^d(G)+\xi ^c(G)\) for arbitrary graphs G are also given, and a sharp lower bound on \(\xi ^d(G)\) for graphs G with a given radius is proved.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In this paper, we consider simple and connected graphs. If \(G = (V(G), E(G))\) is a graph and \(u, v \in V(G)\), then the distance \(d_G(u, v)\) between u and v is the number of edges on a shortest u, v-path. The eccentricity of a vertex and its total distance are distance properties of central interest in (chemical) graph theory; they are defined as follows. The eccentricity \(\varepsilon _G(v)\) of a vertex v is the distance between v and a farthest vertex from v, and the total distance \(D_G(v)\) of v is the sum of distances between v and the other vertices of G. Even more fundamental property of a vertex in (chemical) graph theory is its degree (or valence in chemistry), denoted by \(\deg _G(v)\). (We may skip the index G in the above notations when G is clear.) Multiplicatively combining two out of these three basic invariants naturally leads to the eccentric connectivity index \(\xi ^c(G)\), the eccentric distance sum \(\xi ^d(G)\), and the degree distance DD(G), defined as follows:
\(\xi ^c\) was introduced by Sharma, Goswami, and Madan [18], \(\xi ^d\) by Gupta, Singh, and Madan [7], and DD by Dobrynin and Kochetova [6] and by Gutman [8]. These three topological indices are well investigated, selected contributions to the eccentric connectivity index are [10, 13, 24], see also [21] for its generalization; to the eccentric distance sum [1, 14, 17, 23]; and to the degree distance [15, 19, 20]. The three invariants were also compared to other invariants, cf. [2,3,4,5, 25]. For information on additional topological indices based on eccentricity, see [16].
In [11], the eccentric distance sum and the degree distance are compared, while in [26] the difference between the eccentric connectivity index and the (not defined here) connective eccentricity index is studied. The primary motivation for the present paper, however, is the papers [12, 28] in which \(\xi ^d(G) - \xi ^c(G)\) was investigated. In [28], Zhang, Li, and Xu, besides other results on the two indices, determined sharp upper and lower bounds on \(\xi ^d(G) - \xi ^c(G)\) for graphs G of given order and diameter 2. Parallel results were also derived for subclasses of diameter 2 graphs with specified one of the minimum degree, the connectivity, the edge-connectivity, and the independence number. Hua, Wang, and Wang [12] extended the last result to general graphs. More precisely, they characterized the graphs that attain the minimum value of \(\xi ^d(G) - \xi ^c(G)\) among all connected graphs G of given independence number. They also proved a related result for connected graphs with given matching number.
In this paper, we continue the investigation along the lines of [12, 28] and proceed as follows. In the rest of this section, definitions and some observations needed are listed. In Section 2, we give a lower and an upper bound on \(\xi ^d(G) - \xi ^c(G)\) and in both cases characterize the equality case. The upper bound involves the Wiener index, the first Zagreb index, as well as the degree distance of G. In Section 3, we focus on trees and first prove that among all trees T with given order and diameter, \(\xi ^d(T)-\xi ^c(T)\) is minimized on caterpillars. Using this result, we give a lower bound on \(\xi ^d(T)-\xi ^c(T)\) for all trees T with given order, the bound being sharp precisely on stars. We also give a sharp upper bound on \(\xi ^d(T)-\xi ^c(T)\) for trees T with given order. In the last section, we give a sharp lower bound and a sharp upper bound on \(\xi ^d(G)+\xi ^c(G)\), compare \(\xi ^d(G)\) with \(\xi ^c(G)\) for graphs G with not too large maximum degree, and give a sharp lower bound on \(\xi ^d(G)\) for graphs G with a given radius.
1.1 Preliminaries
The order and the size of a graph G will be denoted by n(G) and m(G), respectively. The star of order \(n\ge 2\) is denoted by \(S_n\); in other words, \(S_n = K_{1,n-1}\). If \(n\ge 2\), then the cocktail party graph \(CP_{2n}\) is the graph obtained from \(K_{2n}\) by removing a perfect matching. The join \(G\oplus H\) of graphs G and H is the graph obtained from the disjoint union of G and H by connecting by an edge every vertex of G with every vertex of H. The maximum degree of a vertex of G is denoted by \(\Delta (G)\). A graph G is regular if all vertices have the same degree. The first Zagreb index [9] \(M_1(G)\) of G is the sum of the squares of the degrees of the vertices of G. The Wiener index [22] W(G) of G is the sum of distances between all pairs of vertices in G.
The diameter \(\mathrm{diam}(G)\) and the radius \(\mathrm{rad}(G)\) of a graph G are the maximum and the minimum vertex eccentricity in G, respectively. A graph G is self-centered if all vertices have the same eccentricity. It this eccentricity is d, we further say that G is d-self-centered. The eccentricity \(\varepsilon (G)\) of G is
The eccentric connectivity index of G can be equivalently written as
and the eccentric distance sum as
Finally, for a positive integer n we use the convention \([n] = \{1,\ldots , n\}\).
2 The Difference on General Graphs
In this section, we give some sharp upper and lower bounds on \(\xi ^d(G) - \xi ^c(G)\) for an arbitrary graph G. The bounds are in terms of the eccentricity, the Wiener index, the first Zagreb index, the degree distance, the maximum degree, the size, and the order of G. To state the result, we recall from [26] the following definitions. Let \(G^*(n, k)\), \(0 \le k \le \lfloor n/2 \rfloor \), be the connected graph of order \(n \ge 4\) obtained from \(K_n\) by deleting k pairwise independent edges. Set further \(G^*(n) = \{G^*(n, k):\ 0 \le k \le \lfloor n/2 \rfloor \}\). We note in passing that the graphs \(G^*(n, k)\), \(1 \le k \le \lfloor n/2 \rfloor \), are named generalized cocktail party graphs in [27]. Then, we recall the following result to be used a couple of times later on.
Lemma 2.1
[26, Lemma 3.1] Let G be a connected graph with \(n(G)\ge 5\). Then, \(\varepsilon (v) = n(G) - \deg (v)\) for any vertex \(v\in V(G)\) if and only if \(G \epsilon G^*(n)\).
Theorem 2.2
If G is a connected graph, then the following hold.
-
(i)
\(\xi ^d(G) - \xi ^c(G) \ge 2\big (n(G)-1-\Delta (G)\big )\varepsilon (G)\). Moreover, the equality holds if and only if G is a regular graph with \(\mathrm{diam}(G) \le 2\).
-
(ii)
\(\xi ^d(G) - \xi ^c(G) \le 2n(G)\big (W(G)-m(G)\big ) + M_1(G) - DD(G)\). Moreover, the equality holds if and only if \(G\in \{P_4\} \cup G^*(n)\).
Proof
(i) Let v be a vertex of G. If w is not adjacent to v, then \(d(v,w)\ge 2\) and consequently \(D(v) - \deg (v) \ge 2(n(G)-1-\Delta (G))\). Thus:
The equality holds if and only if \(D(v) - \deg (v) = 2(n(G)-1-\Delta (G))\) for every vertex v. As the last equality in particular holds for a vertex of maximum degree, we infer that G must be regular. Then, the condition \(D(v) - \deg (v) = 2(n(G)-1-\Delta (G))\) simplifies to
Suppose that \(\mathrm{diam}(G) = d\), and let \(x_i\), \(i\in \{2,\ldots , d\}\), be the number of vertices at distance i from v. Then, \(n(G) = 1 + \Delta (G) + x_2 + \cdots + x_d\) and \(D(v) = \Delta (G) + 2x_2 + \cdots + dx_d\). Plugging these equalities into (3) yields
which implies that \(x_3 = \cdots = x_d = 0\), that is, \(\mathrm{diam}(G) = 2\). Finally, if \(\mathrm{diam}(G) = 2\), then \(D(v) = \Delta (G) + 2(n(G)-\Delta (G)-1)\), so (3) is fulfilled for every regular graph of diameter 2. Clearly, (3) is also fulfilled for graphs of diameter 1, that is, complete graphs.
(ii) If \(v \in V(G)\), then clearly \(\varepsilon (v) \le n(G)-\deg (v)\). Then, we deduce that
The equality in the above computation holds if and only if \(\varepsilon (v) = n(G) - \deg (v)\) holds for all \(v \in V (G)\). So suppose that G is a graph for which \(\varepsilon (v) = n(G) - \deg (v)\) holds for all \(v \in V (G)\) and distinguish the following two cases.
Suppose first that \(\mathrm{diam}(G)\le 2\). If the order of a graph is at most 4, then the equality holds only for \(P_4\). If the order is at least 5, then the statement holds by Lemma 2.1. Suppose second that \(\mathrm{diam}(G)\ge 3\). Let P be a diametral path in G, and let v and \(v'\) be its endpoints. Since \(\varepsilon (v) = n(G) - \deg (v)\) and \(|V(P) \setminus N[v]| = \varepsilon (v) - 1\), it follows that \(n(G) = 1 + \deg (v) + |V(P) \setminus N[v]|\). The latter means that \(V(G) = N[v]\cup V(P)\). Since \(\mathrm{diam}(G) = \varepsilon (v) \ge 3\) it follows that \(\deg (v') = 1\). Since we have also assumed that \(\varepsilon (v') = n(G) - \deg (v')\) holds we see that \(\varepsilon (v') = n(G) -1\) which in turn implies that G is a path. Among the paths \(P_n\), \(n\ge 4\), the path \(P_4\) is the unique one which fulfills the condition \(\varepsilon (v) = n - \deg (v)\) for all \(v \in V (P_n)\). \(\square \)
3 The Difference on Trees
In this section, we turn our attention to \(\xi ^d(T) - \xi ^c(T)\) for trees T, and in particular on extremal trees regarding this difference.
Theorem 3.1
Among all trees T with given order and diameter, \(\min \{ \xi ^d(T)-\xi ^c(T)\}\) is achieved on caterpillars.
Proof
Fix the order and diameter of trees to be considered. Let T be an arbitrary tree that is not a caterpillar with this fixed order and diameter. Let P be a diametral path of T connecting x to y. Then, the eccentricity of each vertex w of T is equal to \(\max \{d(w,x), d(w,y)\}\). Let \(z\ne x,y\) be a vertex of P, and let \(T_z\) be a maximal subtree of T which contains z but no other vertex of P. We may assume that z can be selected such that \(T_z\) contains a vertex at distance \(k \ge 2\) from z, for otherwise T is a caterpillar. Let u be vertex of \(T_z\) with \(d(u,z)=k-1\), and let v be the neighbor of u with \(d(v,z)=k-2\). Let \(S = N(u)\setminus \{v\}\) and let \(s = |S|\). Note that \(s > 0\). Let \(T'\) be the tree obtained from T by replacing the edges between u and the vertices of S with the edges between v and the vertices of S.
Claim A: \(\xi ^d(T) - \xi ^c(T) > \xi ^d(T') - \xi ^c(T')\).
Set \(X_d = \xi ^d(T) - \xi ^d(T')\) and \(X_c = \xi ^c(T) - \xi ^c(T')\). To prove the claim, it is equivalent to show that \(X_d - X_c > 0\).
For a vertex \(w \in V(G)\setminus (S \cup \{u\})\), we have \(D_{T'}(w) = D_{T}(w) - s\) and \(\varepsilon _{T'}(w)\le \varepsilon _{T}(w)\). Moreover if \(w \in S\), then \(\varepsilon _{T'}(w)= \varepsilon _{T}(w) -1\) and \(D_T(w) = D_{T'}(w) + n-s-2\). With these facts in hand, we can compute as follows.
Similarly, but shorter, we get that \(X_c = 2s\). Thus,
This proves Claim A. If \(T'\) is not a caterpillar, we can repeat the construction as many times as required to arrive at a caterpillar. Since at each step the value of \(\xi ^d - \xi ^c\) is decreased, the minimum of this difference is indeed achieved on caterpillars. \(\square \)
Theorem 3.2
If T is a tree of order \(n\ge 3\), then
Moreover, equality holds if and only if \(T = S_n\).
Proof
Let \(n\ge 3\) be a fixed integer. By Theorem 3.1, it suffices to consider caterpillars. More precisely, let T be a caterpillar of order n and with \(\mathrm{diam}(T) = d \ge 3\). Then, we wish to prove that \(\xi ^d(T) - \xi ^c(T) > \xi ^d(S_n) - \xi ^c(S_n) = 4 n^2-12 n+8\). The latter equality is straightforward to check; for the strict inequality, we proceed as follows.
Let \(w, z\in V(T)\) be two adjacent vertices of eccentricities \(d-1\) and \(d-2\), respectively. Let \(S = N(w)\setminus \{z\}\) and set \(s = |S|\). As \(\varepsilon (w) = d-1\), we have \(s\ge 1\). Let further \(S_1 = V(G) \setminus (S\cup \{w,z\})\). Construct now a tree \(T'\) from T by replacing the edges between w and the vertices of S with the edges between z and the vertices of S. Note that \(\deg _T(w) = \deg _{T'}(w)+s\) and \(\deg _T(z) = \deg _{T'}(z)- s\), while the other vertices have the same degree in T and \(T'\). Further, it is straightforward to verify the following relations:
Setting \(X_d = \xi ^d(T)-\xi ^d(T')\), we have:
Similarly, setting \(X_c = \xi ^c(T)-\xi ^c(T')\), we have
Therefore,
that is, \(\xi ^d(T)-\xi ^c(T) > \xi ^d(T')-\xi ^c(T')\). Repeating the above transformation until \(S_n\) is constructed finishes the argument. \(\square \)
To bound the difference \(\xi ^d(T) - \xi ^c(T)\) for an arbitrary tree T from above, we first recall the following result.
Lemma 3.3
[14, Theorem 2.1] Let w be a vertex of graph G. For nonnegative integers p and q, let G(p, q) denote the graph obtained from G by attaching to vertex w pendant paths \(P = wv_1\cdots v_p\) and \(Q = wu_1 \cdots u_q\) of lengths p and q, respectively. Let \(G(p+q,0)=G(p,q)- wu_1+v_pu_1\). If \(r = \varepsilon _G(w)\) and \(r\ge p\ge q \ge 1\), then
Lemma 3.4
Let G, p, q, G(p, q), and \(G(p+q,0)\) be as in Lemma 3.3. Then,
Proof
Let \(\deg '(v)\) and \(\varepsilon '(v)\) (resp. \(\deg (v)\) and \(\varepsilon (v)\)) denote the degree and the eccentricity of v in \(G(p+q,0)\) (resp. G(p, q)). Then, we have:
Moreover, the degrees of vertices in \(G(p+q,0)\) do not decrease. Calculating the difference of contributions of vertices in \(\xi ^c\) for \(G(p+q,0)\) and G(p, q), we can estimate the difference \(X_c = \xi ^c(G(p+q,0)) - \xi ^c(G(p,q))\) as follows:
\(\square \)
Theorem 3.5
If T is a tree of order n, then
Moreover, equality holds if and only if \(T=P_n\).
Proof
The right side of the above inequality is equal to \(\xi ^d(P_n)-\xi ^c(P_n)\). (The value of \(\xi ^d(P_n)\) has been determined in [14], while it is straightforward to deduce \(\xi ^c(P_n)\). Combining the two formulas, the polynomials from the right hand side of the inequality are obtained.) Suppose now that \(T \ne P_n\). Then, there is always a vertex w of degree at least 3 such that we can apply Lemmas 3.3 and 3.4. Setting
we have:
and the result follows. \(\square \)
4 Further Comparison
In this concluding section, we give sharp lower and upper bounds on \(\xi ^d(G)+\xi ^c(G)\), compare \(\xi ^d(G)\) with \(\xi ^c(G)\) for graphs G with \(\Delta (G) \le \frac{2}{3}(n-1)\), and give a sharp lower bound on \(\xi ^d(G)\) for graphs G with a given radius.
Theorem 4.1
If G is a connected graph, then the following hold.
-
(i)
\(\xi ^d(G) + \xi ^c(G)\le 2(n(G)-1) \varepsilon (G) + 2 \mathrm{diam}(G) \big ( W(G) + m(G) - 2{n(G) \atopwithdelims ()2} \big )\).
-
(ii)
\(\xi ^d(G) + \xi ^c(G)\ge 2(n(G)-1)\varepsilon (G) + 2 \mathrm{rad}(G) \big ( W(G) + m(G) - 2{n(G) \atopwithdelims ()2} \big ) \).
Moreover, each of the equalities holds if and only if G is a self-centered graph.
Proof
(i) Partition the pairs of vertices of G into neighbors and non-neighbors, and using (1), we can compute as follows:
The inequality above becomes equality if and only if \(\varepsilon (v) = \mathrm{diam}(G)\) for every \(v \in V(G)\). That is, the equality holds if and only if G is a self-centered graph.
(ii) This inequality as well as its equality case is proved along the same lines as (i). The only difference is that the inequality \(\varepsilon (u)+\varepsilon (v) \le 2\mathrm{diam}(G)\) is replaced by \(\varepsilon (u)+\varepsilon (v) \ge 2\mathrm{rad}(G)\). \(\square \)
In our next result, we give a relation between \(\xi ^d(G)\) and \(\xi ^c(G)\) for graph G with maximum degree at most \(\frac{2}{3}(n(G)-1)\).
Theorem 4.2
If G is a graph with \(\Delta (G) \le \frac{2}{3}(n-1)\), then \(\xi ^d(G) \ge 2\xi ^c(G)\). Moreover, the equality holds if and only if G is 2-self-centered, \(\frac{2}{3}(n(G)-1)\)-regular graph.
Proof
Set \(n = n(G)\) and let v be a vertex of G. Since \(\deg (v) < n-1\) we have \(\varepsilon (v) \ge 2\). Therefore, \(D(v) \ge 2(n-1) - \deg (v) \) with equality holding if and only if \(\varepsilon (v) = 2\). Using the assumption that \(\deg (v) \le \frac{2}{3}(n-1)\), equivalently, \( 2n-2 \ge 3\deg (v)\), we infer that \( \varepsilon (v)D(v) \ge 2\varepsilon (v)\deg (v)\). Summing over all vertices of G, the inequality is proved. Its derivation also reveals that the equality holds if and only if \(\deg (v) =\frac{2}{3}(n-1)\) and \(\varepsilon (v)=2\) for each vertex \(v \in V(G)\). \(\square \)
A graph \(G=C(n; a_0,a_1, \ldots , a_k)\) is called a circulant if \(V(G)=[n]\) and \(E(G)=\{(i,j): |i-j|\in \{a_0, a_1, \ldots , a_k\} \pmod n\}\), where \(1\le a_0<a_1<\cdots <a_k\le n/2\). Consider the circulants \(G_k = C(3k+1; 1, 2,\ldots , k)\), \(k\ge 1\), and note that \(G_1 = C_4\). Then, it can be checked that \(G_k\) is a 2-self-centered, \(\frac{2}{3}(n(G_k)-1) = 2k\)-regular graph. Hence, \(G_k\), \(k\ge 1\), is an infinite series of graphs that attain the equality in Theorem 4.2.
To conclude the paper, we give a lower bound on the eccentric distance sum in terms of the radius of a given graph. Interestingly, the cocktail-party graphs are again among the extreme graphs.
Theorem 4.3
If G is a graph with \(\mathrm{rad}(G) = r\), then
Equality holds if and only if G is a complete graph or a cocktail-party graph.
Proof
Set \(n = n(G)\) and let \(v\in V(G)\). Let P be a longest path starting in v. Separately considering the neighbors of v, the last \(\varepsilon (v) - 2\) vertices of P, and all the other vertices, we can estimate that
Since \(n-\deg (v) \ge \varepsilon (v)\) for every vertex \(v \in V(G)\), we have \(D(v) \ge n+ \varepsilon (v) + \frac{\varepsilon (v)^2 - 3\varepsilon (v)}{2}-1\). Consequently, having the fact \(\varepsilon (v)\ge r\) in mind, we get \(D(v)\ge n-1 + {r \atopwithdelims ()2}\). Multiplying this inequality by \(\varepsilon (v)\) and summing over all vertices of G the claimed inequality is proved.
From the above derivation, we see that the equality can hold only if \(\varepsilon (v)=r=n-\deg (v)\) holds for every \(v\in V(G)\). Then, \(\mathrm{diam}(G)\le 2\) by Lemma 2.1. For the equality, we must also have \(D(v) = n-1 + {r \atopwithdelims ()2}\) for every v. If \(r = 2\) this means that \(D(v) = n\), and hence, \(\deg (v) = n-2\). It follows that G is a cocktail-party graph. And if \(r = 1\), then we get a complete graph. \(\square \)
References
Chen, S., Li, S., Wu, Y., Sun, L.: Connectivity, diameter, minimal degree, independence number and the eccentric distance sum of graphs. Discrete Appl. Math. 247, 135–146 (2018)
Dankelmann, P., Morgan, M.J., Mukwembi, S., Swart, H.C.: On the eccentric connectivity index and Wiener index of a graph. Quaest. Math. 37, 39–47 (2014)
Das, KCh., Trinajstić, N.: Relationship between the eccentric connectivity index and Zagreb indices. Comput. Math. Appl. 62, 1758–1764 (2011)
Das, K.Ch., Nadjafi-Arani, M.J.: Comparison between the Szeged index and the eccentric connectivity index. Discrete Appl. Math. 186, 74–86 (2015)
Das, K.Ch., Su, G., Xiong, L.: Relation between degree distance and Gutman index of graphs. MATCH Commun. Math. Comput. Chem. 76, 221–232 (2016)
Dobrynin, A.A., Kochetova, A.A.: Degree distance of a graph: a degree analogue of the Wiener index. J. Chem. Inf. Comput. Sci. 34, 1082–1086 (1994)
Gupta, S., Singh, M., Madan, A.K.: Eccentric distance sum: a novel graph invariant for predicting biological and physical properties. J. Math. Anal. Appl. 275, 386–401 (2002)
Gutman, I.: Selected properties of the Schultz molecular topological index. J. Chem. Inf. Comput. Sci. 34, 1087–1089 (1994)
Gutman, I., Trinajstić, N.: Graph theory and molecular orbitals. Total \(\pi \)-electron energy of alternant hydrocarbons. Chem. Phys. Lett. 17, 535–538 (1972)
Hauweele, P., Hertz, A., Mélot, H., Ries, B., Devillez, G.: Maximum eccentric connectivity index for graphs with given diameter. Discrete Appl. Math. 268, 102–111 (2019)
Hua, H., Wang, H., Hu, X.: On eccentric distance sum and degree distance of graphs. Discrete Appl. Math. 250, 262–275 (2018)
Hua, H., Wang, H., Wang, M.: The difference between the eccentric distance sum and eccentric connectivity index. Ars Combin. 144, 3–12 (2019)
Ilić, A., Gutman, I.: Eccentric connectivity index of chemical trees. MATCH Commun. Math. Comput. Chem. 65, 731–744 (2011)
Ilić, A., Yu, G., Feng, L.: On the eccentric distance sum of graph. J. Math. Anal. Appl. 381, 590–600 (2011)
Li, S., Song, Y., Zhang, H.: On the degree distance of unicyclic graphs with given matching number. Graphs Combin. 31, 2261–2274 (2015)
Madan, A.K., Dureja, H.: Eccentricity based descriptors for QSAR/QSPR. In: Gutman, I., Furtula, B. (eds.) Novel Molecular Structure Descriptors - Theory and Applications II, pp. 91–138. Univ. Kragujevac, Kragujevac (2010)
Miao, L., Pang, J., Xu, S.: On the extremal values of the eccentric distance sum of trees with a given maximum degree. Discrete Appl. Math. 284, 375–383 (2020)
Sharma, V., Goswami, R., Madan, A.K.: Eccentric connectivity index: A novel highly discriminating topological descriptor for structure - property and structure - activity studies. J. Chem. Inf. Comput. Sci. 37, 273–282 (1997)
Tomescu, A.I.: Unicyclic and bicyclic graphs having minimum degree distance. Discrete Appl. Math. 156, 125–130 (2008)
Wang, H., Kang, L.: Further properties on the degree distance of graphs. J. Comb. Optim. 31, 427–446 (2016)
Vetrik, T., Masre, M.: General eccentric connectivity index of trees and unicyclic graphs. Discrete Appl. Math. 284, 301–315 (2020)
Wiener, H.: Structural determination of paraffin boiling points. J. Am. Chem. Soc. 69, 17–20 (1947)
Xie, Y.-T., Xu, S.-J.: On the maximum value of the eccentric distance sums of cubic transitive graphs. Appl. Math. Comput. 359, 194–201 (2019)
Xu, K., Das, K.C., Liu, H.: Some extremal results on the connective eccentricity index of graphs. J. Math. Anal. Appl. 433, 803–817 (2016)
Xu, K., Li, X.: Comparison between two eccentricity-based topological indices of graphs. Croat. Chem. Acta 89, 499–504 (2016)
Xu, K., Alizadeh, Y., Das, K.Ch.: On two eccentricity-based topological indices of graphs. Discrete Appl. Math. 233, 240–251 (2017)
Xu, K., Das, K.Ch., Gu, X.: Comparison and extremal results on three eccentricity-based invariants of graphs. Acta Math. Sin. (Engl. Ser.) 36, 40–54 (2020)
Zhang, H., Li, S., Xu, B.: Extremal graphs of given parameters with respect to the eccentricity distance sum and the eccentric connectivity index. Discrete Appl. Math. 254, 204–221 (2019)
Acknowledgements
Sandi Klavžar acknowledges the financial support from the Slovenian Research Agency (research core Funding P1-0297 and Projects J1-9109, J1-1693, N1-0095).
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Rosihan M. Ali.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Alizadeh, Y., Klavžar, S. On the Difference Between the Eccentric Connectivity Index and Eccentric Distance Sum of Graphs. Bull. Malays. Math. Sci. Soc. 44, 1123–1134 (2021). https://doi.org/10.1007/s40840-020-01015-5
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40840-020-01015-5