
During the Omni-media era, the in-depth advancement of intelligent process endowed public opinion information (referred to as public opinion) with unique spreading characteristics, and put forward new and higher requirements for its governance. Against this background, we proposed an improved public opinion propagation model coupling the possible factors to grasp its spreading rules. Then, the spreading characteristics of public opinion and its governance timing-intensity-effect in online social networks (OSN) were discussed through numerical simulations. Our results showed that the propagation of public opinion shows faster speed and is more dependent on netizens' attributes in open OSN with a wider scope and depends more on information content in closed OSN. During the governance process of public opinion propagation, the regulators' strategies should have priority: Governance timing ≻ governance proportion ≻ punishment intensity. Based on research findings, targeted countermeasures and decision-making references were provided for the regulators to reasonably guide the evolution trend of public opinion.
Citation: Jiakun Wang, Xiaotong Guo, Yun Li, Liu Chun. An improved propagation model of public opinion information and its governance in online social networks under Omni-media era[J]. Electronic Research Archive, 2024, 32(12): 6593-6617. doi: 10.3934/era.2024308
[1] | Gaofeng Du, Chenghua Gao, Jingjing Wang . Spectral analysis of discontinuous Sturm-Liouville operators with Herglotzs transmission. Electronic Research Archive, 2023, 31(4): 2108-2119. doi: 10.3934/era.2023108 |
[2] | Fei-fan Li, Ji-jun Ao . A dissipative third-order boundary value problem with distributional potentials and eigenparameter-dependent boundary conditions. Electronic Research Archive, 2025, 33(5): 3378-3393. doi: 10.3934/era.2025149 |
[3] | Chenghua Gao, Enming Yang, Huijuan Li . Solutions to a discrete resonance problem with eigenparameter-dependent boundary conditions. Electronic Research Archive, 2024, 32(3): 1692-1707. doi: 10.3934/era.2024077 |
[4] | Vladimir Rovenski . Willmore-type variational problem for foliated hypersurfaces. Electronic Research Archive, 2024, 32(6): 4025-4042. doi: 10.3934/era.2024181 |
[5] | Ting-Ying Chang, Yihong Du . Long-time dynamics of an epidemic model with nonlocal diffusion and free boundaries. Electronic Research Archive, 2022, 30(1): 289-313. doi: 10.3934/era.2022016 |
[6] | Qian He, Wenxin Du, Feng Shi, Jiaping Yu . A fast method for solving time-dependent nonlinear convection diffusion problems. Electronic Research Archive, 2022, 30(6): 2165-2182. doi: 10.3934/era.2022109 |
[7] | Pshtiwan Othman Mohammed, Hari Mohan Srivastava, Dumitru Baleanu, Ehab E. Elattar, Y. S. Hamed . Positivity analysis for the discrete delta fractional differences of the Riemann-Liouville and Liouville-Caputo types. Electronic Research Archive, 2022, 30(8): 3058-3070. doi: 10.3934/era.2022155 |
[8] | Mufit San, Seyma Ramazan . A study for a higher order Riemann-Liouville fractional differential equation with weakly singularity. Electronic Research Archive, 2024, 32(5): 3092-3112. doi: 10.3934/era.2024141 |
[9] | J. F. Toland . Path-connectedness in global bifurcation theory. Electronic Research Archive, 2021, 29(6): 4199-4213. doi: 10.3934/era.2021079 |
[10] | Nan Deng, Meiqiang Feng . New results of positive doubly periodic solutions to telegraph equations. Electronic Research Archive, 2022, 30(3): 1104-1125. doi: 10.3934/era.2022059 |
During the Omni-media era, the in-depth advancement of intelligent process endowed public opinion information (referred to as public opinion) with unique spreading characteristics, and put forward new and higher requirements for its governance. Against this background, we proposed an improved public opinion propagation model coupling the possible factors to grasp its spreading rules. Then, the spreading characteristics of public opinion and its governance timing-intensity-effect in online social networks (OSN) were discussed through numerical simulations. Our results showed that the propagation of public opinion shows faster speed and is more dependent on netizens' attributes in open OSN with a wider scope and depends more on information content in closed OSN. During the governance process of public opinion propagation, the regulators' strategies should have priority: Governance timing ≻ governance proportion ≻ punishment intensity. Based on research findings, targeted countermeasures and decision-making references were provided for the regulators to reasonably guide the evolution trend of public opinion.
In 2008, Bhattacharya et al. [5] and Bell et al. [4] discovered that bipartite chain graphs whose largest eigenvalues within the connected bipartite graph is maximal, and named therein as double nested graphs. After that, many scholars began to study some mathematical properties of chain graphs. Andelic et al. provide that some upper and lower bounds on index of chain graphs [3]. And Alazemi et al. proved that any chain graph has its least positive eigenvalue greater than 12 [2]. Hence Zhang et al. proposed that upper bounds on Laplacian spectral radius of chain graphs [13]. Das et al. studied the energy and Laplacian energy of chain graphs [8]. In this paper, we further study some bounds of energy and Laplacian energy of chain graphs.
We consider finite undirected connected graphs without loops and multiple edges. Let G be a such graph with vertex set V(G)={v1,v2,…,vn} and edge set E(G), where |E(G)|=m. Let di be the degree of the vertex vi for i=1,2,…,n. The minimum vertex degrees of G are denoted by δ(G). Let NG(vi) be the adjacent set of the vertex vi, then di=|NG(vi)|. If G has distinct vertices vi and vj with NG(vi)=NG(vj), then vi and vj are duplicates and (vi,vj) is a duplicate pair.
Let A(G) be the adjacency matrix of G, and rank(G) be the rank of the adjacency matrix A(G). Let λ1≥λ2≥⋯≥λn the eigenvalues of A(G). We denote S(G)={λ1,λ2,…,λn} as the spectrum of G. The energy of graph G is defined as [11]
ε(G)=n∑i=1|λi|. |
For its basic properties and application, including various lower and upper bounds, see the [17], the recent paper [1,7,8,11,12,20] and the references cited therein.
The Laplacian matrix of graph G is defined as L(G)=D(G)−A(G), where D(G) is the diagonal matrix of vertex degrees. The matrix L(G) has non-negative eigenvalues μ1≥μ2≥⋯≥μn−1≥μn=0, and the Laplacian spectrum of graph G be denoted by LS(G)={μ1,μ2,…,μn}. The Laplacian energy of G is defined as [10]
LE(G)=n∑i=1|μi−2mn|. |
It can also be defined as
LE(G)=2Sσ(G)−4mσn, | (1.1) |
where σ(1⩽σ⩽n) be the largest positive integer such that μσ≥2mn and Sk(G)=k∑i=1μi.
For its basic properties, including various lower and upper bounds, see [7,8,10,18,19] and the references cited therein. The Laplacian energy found applications not only in theoretical organic chemistry [12,21], but also in image processing [22] and information theory [16].
In the class of bipartite graphs of fixed order and size those having maximal spectral radius of adjacency/Laplacian/signless Laplacian matrix are chain graphs. Thus, they can be significant in modeling some bipartite networks with large spectral radius. Their applications involve ecological networks, in which graphs with nested properties are considered [14] and are used in some applications for economic network modeling.
We now introduce the structure of a (connected) chain graph. The vertex set of any chain graph consists of two color classes, which are U and V. Both of them are divided into h non-empty units U1,U2,…,Uh and V1,V2,…,Vh, respectively. All the vertices in Us are joined by edges to all vertices in ⋃h+1−sk=1Vk, for s=1,2,…,h. Therefore, if ui∈Us+1 and uj∈Us, then NG(ui)⊂NG(uj), or if vi∈Vt+1 and vj∈vt, then NG(vi)⊂NG(vj).
If ns=|Us| and ms=|Vs| for s=1,2,…,h, then G is denoted by G(m1,…,mh;n1,…,nh), as shown in Figure 1. And
m=m1h∑i=1ni+m2h−1∑i=1ni+⋯+mhn1=h∑i=1aimi, |
m=n1h∑i=1mi+n2h−1∑i=1mi+⋯+nhm1=h∑i=1bini, |
where
ai=h+1−i∑k=1nk,bi=h+1−i∑k=1mk. |
Moreover,
n=h∑k=1mk+h∑k=1nk. |
The second smallest Laplacian eigenvalue of a graph is well known as the algebraic connectivity. It has been proved that the second smallest Laplacian eigenvalue μn−1=0 if and only if G is disconnected. The algebraic connectivity is often applied in theoretical chemistry, control theory, combinatorial optimization and other fields [15].
As usual, Kn, Kp,q(p+q=n) and K1,n−1, denote, respectively, the complete graph, the complete bipartite graph and the star on n vertices. For other undefined notations and terminology from graph theory, the readers are referred to [6].
The paper is organized as follows. In Section 2, we list some previously known results. In Section 3, we get some upper and lower bounds on ε(G) of a chain graph G. In Section 4, we establish an upper bound on LE(G) of the chain graphs in terms of vertex cover number. In Section 5, we attain the maximal Laplacian energy of the bicyclic chain graph G by comparing the algebraic connectivity.
This section lists some known results to be used in this paper.
Lemma 2.1. [8] Let B be a p×p real symmetric matrix and Bk be its leading k×k submatrix. Then for i=1,2,…,k,
λp−i+1(B)⩽λk−i+1(Bk)⩽λk−i+1(B), |
where λi(B) is the i-th largest eigenvalue of B.
Lemma 2.2. [9] Let G be a graph with vertices {v1,v2,…,vk}⊆V(G) having same set of adjacent vertices, then G has at least k−1 equal eigenvalues 0.
Lemma 2.3. [18] Let G≆Kn. Then μn−1⩽δ(G).
Lemma 2.4. [10] Let A and B be real symmetric matrices of order n. Then for any 1⩽k⩽n,
k∑i=1λi(A+B)⩽k∑i=1λi(A)+k∑i=1λi(B), |
where λi(M) denotes the i-th largest eigenvalue of the matrix M.
Lemma 2.5. [1] If G is a connected bipartite graph of rank r, then
ε(G)≥√(r+1)2−5. |
Lemma 2.6. [11] If G is a connected bipartite graph of rank r, then
LE(G)≥2(ε(G)−r). |
Lemma 2.7. [8] Let G≅G(m1,…,mh;n1,…,nh) be a chain graph of order n. Then
ε(G)≥2√n−1, |
with equation holds if and only if G≅K1,n−1.
Lemma 2.8. [8] Let G be a graph with vertex set V(G)={v1,v2,…,vn}. If G has k−1 duplicate pairs (vi,vi+1), where i=1,2,…,k−1, then G has at least k−1 equal Laplacian eigenvalues and they are all equal to the cardinality of the neighbor set.
Theorem 3.1. Let G≅G(m1,…,mh;n1,…,nh) be a chain graph of order n. Then
ε(G)⩽2√hm | (3.1) |
with equation holds if and only if G≅Kn1,m1, where n1+m1=n.
Proof. By Lemma 2.2, the eigenvalue 0 with multiplicity h∑i=1(ni+mi−2) of A(G), and the remaining eigenvalues are the eigenvalues of the following matrix,
C=(00⋯00m1m2⋯mh−1mh00⋯00m1m2⋯mh−10⋮⋮⋱⋮⋮⋮⋮⋱⋮⋮00⋯00m1m2⋯0000⋯00m10⋯00n1n2⋯nh−1nh00⋯00n1n2⋯nh−1000⋯00⋮⋮⋱⋮⋮⋮⋮⋱⋮⋮n1n2⋯0000⋯00n10⋯0000⋯00). |
Let λ1≥λ2≥⋯≥λ2h be the eigenvalues of C. Then
ε(G)=2h∑i=1|λi|. |
Since G be a bipartite graph, we have λi and −λi are eigenvalues of G. Thus we have
ε(G)=2h∑i=1λi. |
Obviously,
2h∑i=1λ2i=Tr(C2)=2h∑i=1h−i+1∑j=1mjni=2m, |
that is,
h∑i=1λ2i=h∑i=1h−i+1∑j=1mjni=m. |
So
ε(G)=2√h∑i=1λ2i+2∑1⩽i<j⩽hλiλj≤2√h∑i=1λ2i+h∑i=1(h−1)λ2i=2√hh∑i=1λ2i=2√hm. |
First we assume that h=1. Then G≅Kn1,m1, where n1+m1=n. So S(G)={±√m1n1,0,⋯,0} and ε(G)=2√m1n1=2√m. Hence the equation holds in (3.1).
Next we assume that h≥2. By the definition of chain graph, G(1,1;1,1), that is, P4 is an induced subetaaph of G. By Lemma 2.1, we get λ2(G)≥λ2(P4)>0. Since G is connected, by Perron-Frobenius theorem we have λ1(G)>λ2(G). Hence the inequality 2∑1⩽i<j⩽hλiλj≤h∑i=1(h−1)λ2i is strict. This completes the proof.
Theorem 3.2. Let G≅G(m1,…,mh;n1,…,nh) be a chain graph of order n. Then
ε(G)≥√(2h+1)2−5. | (3.2) |
Proof. By calculating the matrix C in the proof of Theorem 3.1, we get
det(C)=(−1)hh∏i=1mini≠0. |
Therefore, all the eigenvalues of matrix C are non-zero. Hence r(G)=2h. Using Lemma 2.5, we can get result in (3.2).
In this section, we give an upper bound on LE(G) of chain graphs in terms of vertex cover number. Also, the lower bound follows from a known lower bound for Laplacian energy of any graph in terms of rank and energy.
Theorem 4.1. Let G≅G(m1,…,mh;n1,…,nh) be a chain graph of order n, and a1≥b1. Then
LE(G)≤{2(m+b1)−4mn,if2mn≥b1,2b1(n−2)−2m+8mn,if2mn<b1, | (4.1) |
with equation holds if and only if G≅K1,n−1.
Proof. Let Γ={v11,v12,…,v1m1,v21,v22,…,v2m2,…,vh1,vh2,…,vhmh} be a vertex cover set of the graph G, where vij is the j-th vertex in Vi. Hence {vi1,vi2,…,vimi}∈Vi. We can assume that Gij are spanning subetaaphs of G such that V(G)=V(Gi1)=V(Gi2)=⋯=V(Gimi), and the edge set of Gij is defined as
E(Gij)={vijUk:Uk⊆NG(vij)}. |
Since |NG(vi1)|=|NG(vi2)|=⋯=|NG(vimi)|=ai,
Gij=K1,ai∪(n−ai−1)K1, |
we have
E(Kmi,ai)=E(Gi1)∪E(Gi2)∪⋯∪E(Gim1), |
so
L(Kmi,ai)=L(Gi1)+L(Gi2)+⋯+L(Gim1),i=1,2,…,h. |
By Figure 1,
E(G)=E(Km1,a1)∪E(Km2,a2)∪⋯∪E(Kmi,ai), |
then we can see easily that
L(G)=L(Km1,a1)+L(Km2,a2)+⋯+L(Kmi,ai). |
Note that
Sk(Gi1)=Sk(Gi2)=⋯=Sk(Gimi)≤ai+k, |
where Sk(G) is the sum of the k largest Laplacian eigenvalues of graph G.
By Lemma 2.4, we get
Sk(G)≤m1Sk(G11)+m2Sk(G21)+⋯+mhSk(Gh1)≤m1(a1+k)+m2(a2+k)+⋯+mh(ah+k)=h∑i=1miai+kh∑i=1mi=m+kb1. |
So from (1.1), we get
LE(G)=2Sσ(G)−4mσn≤2(m+σb1)−4mσn=2m+2σ(b1−2mn). |
Since G is connected, 1≤σ≤n−1. So it suffices to consider the following two cases.
Case1. 2mn≥b1.
Then we have
LE(G)≤2m+2b1−4mn=2(m+b1)−4mn. |
Case2. 2mn<b1.
By Lemma 2.3, we get μn−1≤δ(G)≤2mn. Thus it must be 1≤σ≤n−2. Hence
LE(G)≤2m+2(n−2)(b1−2mn)=2b1(n−2)−2m+8mn. |
Next we prove that the equality holds.
If G≅K1,n−1, we get b1=m1=1,n1=n−1, and S(G)={0,1n−2,n}. Then
LE(K1,n−1)=n∑i=1|μi−2mn|=2n−4(n−1)n=2(m+b1)−4mn. |
Theorem 4.2. Let G≅G(m1,…,mh;n1,…,nh) be a chain graph of order n. Then
LE(G)≥4(√n−1−h). | (4.2) |
Proof. By Theorem 3.2, we get r(G)=2h. Using Lemmas 2.6 and 2.7, we get result in (4.2).
Let G be a connected bicyclic chain graph. We have m=n+1, and h=2 or h=3. If h=2, then G≅G(1,1;3,n−5) or G≅G(1,2;2,n−5). If h=3, then G≅G(1,2,k−3;1,1,n−k−2), where 4≤k≤n−3 (Figure 2). In this section, we will attain the maximal Laplacian energy of all connected bicyclic chain graphs.
Lemma 5.1. Let G be a connected bicyclic chain graph (n≥8).
(1) If G≅G(1,1;3,n−5), then LE(G)=6+2(n−4)(n+1)n−2μn−1.
(2) If G≅G(1,2;2,n−5), then LE(G)=10+2(n−6)(n+1)n−2μn−1.
(3) If G≅G(1,2,k−3;1,1,n−k−2), where 4≤k≤n−3, then LE(G)=10+2(n−6)(n+1)n−2μn−1.
Proof. (1) Let G≅G(1,1;3,n−5). By Lemma 2.8, we conclude that 2,2,1,1,⋯,1⏟n−6 are the Laplacian eigenvalues of G and the remaining Laplacian eigenvalues of G are satisfying the equation f1(x)=0, where f1(x) is the characteristic polynomial of the matrix
A1=(n−20−35−n03−30−1−120−1001), |
that is, f1(x)=x(x3−(4+n)x2+(5n−2)x−3n).
Let h1(x)=x3−(4+n)x2+(5n−2)x−3n. Then we obtain h1(0)=−3n<0, h1(1)=n−5>0, h1(2)=3n−12>0, h1(n−1)=−3<0 and limx→∞h1(x)=∞. Thus the Laplacian eigenvalues of G are μ1,μ2,2,2,1,1,…,1⏟n−6,μn−1,0, where μ1≥n−1, 2≤μ2≤n−1, μn−1<1 and μ1+μ2+μn−1=n+4.
Therefore
LE(G)=n∑i=1|μi−2(n+1)n|=6+2(n−4)(n+1)n−2μn−1. | (5.1) |
(2) Let G≅G(1,2;2,n−5). By Lemma 2.8, we conclude that 3,2,1,1,⋯,1⏟n−6 are the Laplacian eigenvalues of G and the remaining Laplacian eigenvalues of G are satisfying the equation f2(x)=0, where f2(x) is the characteristic polynomial of the matrix
A2=(n−30−25−n02−20−1−230−1001), |
that is, f2(x)=x(x3−(3+n)x2+(5n−8)x−2n).
Let h2(x)=x3−(3+n)x2+(5n−8)x−2n. Then we obtain h2(0)=−2n<0, h2(1)=2n−10>0, h2(3)=4n−24>0, h2(n−2)=−4<0 and limx→∞h2(x)=∞. Thus the Laplacian eigenvalues of G are μ1,μ2,3,2,1,1,…,1⏟n−6,μn−1,0, where μ1≥n−2, 3≤μ2≤n−2, μn−1<1 and μ1+μ2+μn−1=n+3.
Therefore
LE(G)=n∑i=1|μi−2(n+1)n|=10+2(n−6)(n+1)n−2μn−1. | (5.2) |
(3) Let G≅G(1,2,k−3;1,1,n−k−2). When 4≤k≤⌈n2⌉, by Lemma 2.8, we conclude that 2,1,1,⋯,1⏟n−7 are the Laplacian eigenvalues of G and the remaining laplacian eigenvalues of G are satisfying equation f3(x)=0, where f3(x) is the characteristic polynomial of the matrix
A3=(n−k00−1−12+k−n020−1−10001−100−1−23−kk00−1−20030−100001), |
that is
f3(x)=x(x−1)(x4−(n+6)x3+(kn+5n−k2+10)x2−(4kn+5n−4k2+12)x+6n). | (5.3) |
Let g(x)=x4−(n+6)x3+(kn+5n−k2+10)x2−(4kn+5n−4k2+12)x+6n. Then we obtain g(0)=6n>0, g(1)=3k2−3kn+5n−7<0, g(2)=4(k−2)(2+k−n)<0, g(k)=−(k−2)(k−3)(2k−n)≥0. Since when n is odd, g(x) is same for k=⌈n2⌉ and k=⌊n2⌋, we take a smaller value k=⌊n2⌋. g(n−k)=(2+k−n)(2k−n)(−n+3+k)≤0 and limx→∞g(x)=∞. Thus the Laplacian eigenvalues of G are μ1,μ2,μ3,2,1,1,⋯,1⏟n−7,μn−1,0, where μ1≥n−k, k≤μ2≤n−k, 2<μ3<k, μn−1<1.
Since n∑i=1μi=2m=2(n+1)=2n+2, we get μ1+μ2+μ3+μn−1=n+6, that is, μ1+μ2+μ3=n+6−μn−1.
Therefore
LE(G)=n∑i=1|μi−2(n+1)n|=10+2(n−6)(n+1)n−2μn−1. | (5.4) |
When ⌈n2⌉<k<n−3, letting k=n−k in the Eq (5.3) we get the same characteristic polynomial, so it is equal to the Laplacian energy when 4≤k≤⌈n2⌉.
When k=n−3, f3(x)=x(x−1)(x−3)(x3−(3+n)x2+(5n−8)x−2n), so it is equal to the Laplacian energy of G(1,2;2,5).
This completes the proof.
Lemma 5.2. Let Gn,k≅G(1,2,k−3;1,1,n−k−2), where 4≤k≤⌈n2⌉. Then μn−1(Gn,k)≥μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)), with equation holds if and only if k=⌈n2⌉. In particular, if n is odd, then μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=μn−1(G(1,2,⌈n2⌉−4;1,1,⌊n2⌋−1)).
Proof. If k=⌈n2⌉, then μn−1(Gn,k)=μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)). By Lemma 5.1, we obtain that μ1,μ2,μ3,μn−1 are the roots of the equation P(Gn,k,x)=0, where
P(Gn,k,x)=x4−(n+6)x3+(kn+5n−k2+10)x2−(4kn+5n−4k2+12)x+6n, |
and μ1≥n−k, k≤μ2≤n−k, 2<μ3<k, μn−1<1.
We need to prove that
μn−1(Gn,k)>μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)),for4≤k≤⌈n2⌉−1. |
Since
P(Gn,k+1,x)−P(Gn,k,x)=x(x−4)(n−2k−1),for0<x<1, |
we get P(Gn,k+1,x)−P(Gn,k,x)≤0. Hence P(Gn,k+1,x)≤P(Gn,k,x). So when n is odd and k=⌈n2⌉−1, the equation holds.
Thus we have μn−1(Gn,k)>μn−1(Gn,k+1), that is,
μn−1(Gn,4)>μn−1(Gn,5)>⋯>μn−1(Gn,⌈n2⌉−1)≥μn−1(Gn,⌈n2⌉). | (5.5) |
Hence μn−1(Gn,k)>μn−1(Gn,⌈n2⌉)=μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)).
This completes the proof.
Lemma 5.3. Let G be a bicyclic graph of order n≥8. Then μn−1(G(1,2;2,n−5))>μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)).
Proof. When k=3, we get P(Gn,k,x)=f2(x), that is μn−1(Gn,3)=μn−1(G(1,2;2,n−5)).
By Lemma 5.2, we have P(Gn,k+1,x)≤P(Gn,k,x), and P(Gn,4,x)≤P(Gn,3,x) still hold.
By inequation (5.5), we obtain
μn−1(Gn,3)>μn−1(Gn,4)>⋯>μn−1(Gn,⌈n2⌉−1)≥μn−1(Gn,⌈n2⌉). |
Hence μn−1(G(1,2;2,n−5))>μn−1(Gn,⌈n2⌉)=μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)) for n≥8.
Lemma 5.4. Let G be a bicyclic graph of order n≥8. Then μn−1(G(1,1;3,n−5))−μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))>2n.
Proof. For n=8 and n=9, it can be verified by using Maple.
Let n=8, μn−1(G(1,1;3,n−5))=0.8377 and μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=0.5858. Then μn−1(G(1,1;3,n−5))−μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=0.2519>14, so the conclusion is correct.
Let n=9, μn−1(G(1,1;3,n−5))=0.8169 and μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=0.5344. Then μn−1(G(1,1;3,n−5))−μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=0.2825>29, so the conclusion is correct.
Next we prove when n≥10, the inequality holds.
By Lemma 5.3, we get μn−1(G(1,2;2,n−5))≥μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)), so we can prove μn−1(G(1,1;3,n−5))−μn−1(G(1,2;2,n−5))>2n. Let α=μn−1(G(1,1;3,n−5)), β=μn−1(G(1,2;2,n−5)). Then it is satisfying
h1(x)=x3−(4+n)x2+(5n−2)x−3nandh1(α)=0. |
h2(x)=x3−(3+n)x2+(5n−8)x−2nandh2(β)=0. |
By the implicit function existence theorem and Figure 3, when G≅G(1,1;3,n−5), the relation between the decreases of α and the increase of n, and h1(x) is monotonically increasing on the interval [0,1]. Hence h1(0.81)=−3.713+0.39n>0, h1(0.69)=−2.956−0.26n<0, so 0.69<α<0.81.
Similarly, h2(0.58)=−5.454+0.56n>0, h2(0.43)=−3.915−0.035n<0, so 0.43<β<0.58. Therefore, α−β>0.11>219, that is, when n≥19, hence the conclusion is correct.
When 10≤n≤18, α−β>2n is obvious. The results are shown in Table 1.
n | α | β | α−β | 2n |
10 | 0.8107 | 0.5735 | 0.2372 | 0.200 |
11 | 0.7899 | 0.5566 | 0.2333 | 0.182 |
12 | 0.7804 | 0.5438 | 0.2366 | 0.167 |
13 | 0.7728 | 0.5332 | 0.2396 | 0.154 |
14 | 0.7666 | 0.5248 | 0.2418 | 0.143 |
15 | 0.7612 | 0.5176 | 0.2436 | 0.133 |
16 | 0.7566 | 0.5116 | 0.2450 | 0.125 |
17 | 0.7526 | 0.5064 | 0.2462 | 0.118 |
18 | 0.7491 | 0.5020 | 0.2471 | 0.111 |
So we conclude that when n≥8,
μn−1(G(1,1;3,n−5))−μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))>2n. |
Theorem 5.1. Let G be a connected bicyclic chain graph of order n≥8. Then G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2) attains the maximal Laplacian energy. In particular, when n is odd, LE(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=LE(G(1,2,⌈n2⌉−4;1,1,⌊n2⌋−1)).
Proof. By Lemma 5.1, we can attain the maximal Laplacian energy by comparing μn−1 in equations (5.1), (5.2) and (5.4). It is obvious that LE(G(1,2;2,n−5))<LE(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)). In particular, when n is odd, LE(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))=LE(G(1,2,⌈n2⌉−4;1,1,⌊n2⌋−1)). So
LE(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))−LE(G(1,1;3,n−5))=10+2(n−6)(n+1)n−2μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))−6−2(n−4)(n+1)n+2μn−1(G(1,1;3,n−5))=2(μn−1(G(1,1;3,n−5))−μn−1(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2)))−4n. |
Hence by Lemma 5.4, LE(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))−LE(G(1,1;3,n−5))>0, that is, LE(G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2))>LE(G(1,1;3,n−5)). In conclusion, we get G(1,2,⌈n2⌉−3;1,1,⌊n2⌋−2) has the maximal Laplacian energy among all connected bicyclic chain graphs (n≥8).
In this paper, we introduced the definition of chain graph. We obtain some bounds on ε(G) of the chain graphs. Since the rank of the chain graphs is 2h, we can get some bounds on ε(G) and LE(G) of the chain graphs. We present the upper bound on LE(G) of the chain graphs in terms of vertex cover number. In order to attain the maximal Laplacian energy of bicyclic chain graphs, we compare algebraic connectivity of each kind of bicyclic chain graphs. The problem is still open to discuss what chain graphs give the maximal Laplacian energy for given n and whether it is still related to algebraic connectivity.
This work was supported by National Nature Science Foundation of China (Grant No. 61774137). The authors express their sincere thanks to the anonymous referee for many valuable comments and suggestions.
The authors declare that they have no conflict of interest in this paper.
[1] | K. Simon, Digital 2023: Global Digital Yearbook, We Are Social & Hootsuite: America, 2023. |
[2] |
H. Xu, X. Zhang, H. Gong, The rule of law model of Internet governance, Social Sciences in China, 40 (2019), 135–151. https://doi.org/10.1080/02529203.2019.1639960 doi: 10.1080/02529203.2019.1639960
![]() |
[3] |
B. Abu-Salih, Y. Chan, O. Al-kadi, M. Al-Tawil, P. Wongthongtham, T. Issa, et al., Time-aware domain-based social influence prediction, J. Big Data, 7 (2020), 1−37. https://doi.org/10.1186/s40537-020-0283-3 doi: 10.1186/s40537-020-0283-3
![]() |
[4] |
Y. Hou, F. Meng, J. Wang, Y. Li, Research on two-stage public opinion evolution configuration path based on fuzzy set qualitative comparative analysis, Aslib. J. Inf. Manag., 76 (2024), 677–693. https://doi.org/10.1108/AJIM-10-2022-0464 doi: 10.1108/AJIM-10-2022-0464
![]() |
[5] |
F. Pierri, C. Piccardi, S. Ceri, Topology comparison of Twitter diffusion networks effectively reveals misleading information, Sci. Rep., 10 (2020), 1–9. https://doi.org/10.1038/s41598-020-58166-5 doi: 10.1038/s41598-020-58166-5
![]() |
[6] |
Y. Li, J. Wang, Cross-network propagation model of public opinion information and its control in coupled double-layer online social networks, Aslib. J. Inf. Manag., 74 (2021), 354–376. https://doi.org/10.1108/AJIM-04-2021-0126 doi: 10.1108/AJIM-04-2021-0126
![]() |
[7] |
J. Wang, Y. Li, Research on the propagation and governance of public opinion information under the joint action of internal and external factors, Aslib. J. Inf. Manag., 75 (2023), 193–214. https://doi.org/10.1108/AJIM-02-2022-0065 doi: 10.1108/AJIM-02-2022-0065
![]() |
[8] |
G. Gao, T. Wang, X. Zheng, Y. Chen, X. Xu, A systems dynamics simulation study of network public opinion evolution mechanism, J. Glob. Inf. Manag., 27 (2019), 189–207. https://doi.org/10.4018/JGIM.2019100110 doi: 10.4018/JGIM.2019100110
![]() |
[9] |
H. Schawe, L. Hernandez, Higher order interaction destroy transitions in Deffuant opinion dynamics model, Commun. Phys., 5 (2022), 32. https://doi.org/10.1038/s42005-022-00807-4 doi: 10.1038/s42005-022-00807-4
![]() |
[10] |
H. Chen, X. Zhao, Stochastic evolutionary game model of hot topics propagation for network public opinion, IEEE Trans. Comput. Social Syst, (2023), 1–13. https://doi.org/10.1109/TCSS.2023.3265020 doi: 10.1109/TCSS.2023.3265020
![]() |
[11] |
G. Maji, S. Mandal, S. Sen, A systematic survey on influential spreaders identification in complex networks with a focus on K-shell based techniques, Expert Syst. Appl., 161 (2020), 1–18. https://doi.org/10.1016/j.eswa.2020.113681 doi: 10.1016/j.eswa.2020.113681
![]() |
[12] |
T. Nguyen, T. Phan, M. Nguyen, M. Weidlich, H. Yin, J. Jo, et al., Model-agnostic and diverse explanations for streaming rumor graphs, Knowl. Based Syst., 253 (2022), 109438. https://doi.org/10.1016/j.knosys.2022.109438 doi: 10.1016/j.knosys.2022.109438
![]() |
[13] |
Z. Qiu, X. Yuan, Y. Yin, Research on social governance of network public opinion: An evolutionary game mode, Discrete Dyn. Nat. Soc., (2023), 8530530. https://doi.org/10.1155/2023/8530530 doi: 10.1155/2023/8530530
![]() |
[14] |
L. Cao, G. Wei, J. Su, Public opinion spread risk assessment model on third-party payment rough network, Appl. Soft Comput., 95 (2020), 106532. https://doi.org/10.1016/j.asoc.2020.106532 doi: 10.1016/j.asoc.2020.106532
![]() |
[15] |
O. Williams, L. Lacasa, A. Millian, V. Latora, The shape of memory in temporal networks, Nat. Commun., 13 (2022), 499. https://doi.org/10.1038/s41467-022-28123-z doi: 10.1038/s41467-022-28123-z
![]() |
[16] |
H. Huang, Y. Chen, Y. Ma, Modeling the competitive diffusion of rumor and knowledge and the impacts on epidemic spreading, Appl. Math. Comput., 338 (2021), 125536. https://doi.org/10.1016/j.amc.2020.125536 doi: 10.1016/j.amc.2020.125536
![]() |
[17] |
Y. Lan, Z. Lian, R. Zeng, D. Zhu, Y. Xia, M. Liu, P. Zhang, A statistical model of the impact of online rumors on the information quantity of online public opinion, Physica A, 541 (2020), 123623. https://doi.org/10.1016/j.physa.2019.123623 doi: 10.1016/j.physa.2019.123623
![]() |
[18] | Y. Lan, L. Zhang, W. Wang, L. Zhao, H, Duan, Abnormal perception of network public opinion and empirical research orienting risk monitoring, J. Modern Inform., 42 (2022), 102–108. (in Chinese) |
[19] |
Z. Zhang, C. Liu, X. Zhan, X. Lu, C. Zhang, Y. Zhang, Dynamics of information diffusion and its applications on complex networks, Phys. Rep., 651 (2016), 1–34. https://doi.org/10.1016/j.physrep.2016.07.002 doi: 10.1016/j.physrep.2016.07.002
![]() |
[20] |
H. Yang, C. Gu, M. Tang, S. Cai, Y. Lai, Suppression of epidemic spreading in time-varying multiplex networks, Appl. Math. Model., 75 (2019), 806–818. DOI10.1016/j.apm.2019.07.011 doi: 10.1016/j.apm.2019.07.011
![]() |
[21] |
J. Wang, H. Yu, Y. Li, Research on the co-evolution of temporal networks structure and public opinion propagation, J. Inform. Sci., (2022), 1–15. 10.1177/01655515221121944 doi: 10.1177/01655515221121944
![]() |
[22] |
P. Monin, R. Bookstaber, Information flows and crashes in dynamic social networks, J. Econ. Interact., 16 (2021), 471–495. https://doi.org/10.1007/s11403-020-00310-5 doi: 10.1007/s11403-020-00310-5
![]() |
[23] |
Z. Chen, H. Lan, Dynamics of public opinion: Diverse media and audiences' choices, J. Artif. Soc. S., 24 (2021), 8–19. https://doi.org/10.18564/jasss.4552 doi: 10.18564/jasss.4552
![]() |
[24] |
Y. Wang, L. Han, Q. Qian, J. Xia, J. Li, Personalized recommendation via Multi-dimensional Meta-paths temporal graph probabilistic spreading, Inform. Process. Manag., 59 (2022), 102787. https://doi.org/10.1016/j.ipm.2021.102787 doi: 10.1016/j.ipm.2021.102787
![]() |
[25] | L. Zhang, X. Wang, B. Huang, T. Liu, Research on the topic clustering graph and the transmission path of Micro-blogging users amid COVID-19 based on the LDA Model, J. China Soc. Sci. Tech. Inf., 40 (2021), 234–244. (in Chinese) |
[26] | Y. Li, X. Wang, A. Wangnan, X. Wang, Risk identification and early warning model of social media network public opinion in emergencies, J. China Soc. Sci. Tech. Inf., 41 (2022), 1085–1099. (in Chinese) |
[27] |
B. Anderson, M. Ye, Recent advances in the modelling and analysis of opinion dynamics on influence networks, Int. J. Autom. Comput., 16 (2021), 129–149. https://doi.org/10.1007/s11633-019-1169-8 doi: 10.1007/s11633-019-1169-8
![]() |
[28] |
D. Daley, D. Kendall, Epidemics and Rumors, Nature, 204 (1964), 1118. https://doi.org/10.1038/2041118a0 doi: 10.1038/2041118a0
![]() |
[29] |
D. Zanette, Dynamics of rumor propagation on small-world networks, Phys. Rev. E, 65 (2002), 041908. https://doi.org/10.1103/PhysRevE.65.041908 doi: 10.1103/PhysRevE.65.041908
![]() |
[30] |
W. Wang, M. Tang, H. Stanley, L. Braunstein, Unification of theoretical approaches for epidemic spreading on complex networks, Rep. prog. phys., 80 (2017), 036603. https://doi.org/10.1088/1361-6633/aa5398 doi: 10.1088/1361-6633/aa5398
![]() |
[31] |
W. Wang, Q. Liu, J. Liang, Y. Hu, T. Zhou, Coevolution spreading in complex networks, Phys. Rep., 820 (2019), 1–51. https://doi.org/10.1016/j.physrep.2019.07.001 doi: 10.1016/j.physrep.2019.07.001
![]() |
[32] | H. Zhu, X. Yan, S. Zhang, S. Jin, Modeling of idea diffusion in multiplex networks, Syst. Eng. - Theory & Practice, 40 (2021), 771–780. (in Chinese) |
[33] | H. Zhu, X. Yan, S. Jin, S. Zhang, S. Deng, Research on diffusion model and promotion strategies for creative idea among employees with consideration of multiple channels, Syst. Eng. - Theory & Practice, 43 (2022), 251–265. (in Chinese) |
[34] |
A. Azizi, C. Montalvo, B. Espinoza, Y. Kang, C. Carlos, Epidemics on networks: Reducing disease transmission using health emergency declarations and peer communication, Infect. Dis. Model., 5 (2020), 12–22. https://doi.org/10.1016/j.idm.2019.11.002 doi: 10.1016/j.idm.2019.11.002
![]() |
[35] |
J. Enright, K. Meeks, G. Mertzios, V. Zamaraev, Deleting edges to restrict the size of an epidemic in temporal networks, J. Comput. Syst. Sci., 119 (2021), 60–77. https://doi.org/10.1016/j.jcss.2021.01.007 doi: 10.1016/j.jcss.2021.01.007
![]() |
[36] | M. Wang, Y. Liu, L. Guo, Research on the "Time-Degree-Effect" governance of online public opinion based on evolutionary game, Manag. Rev., 35 (2023), 315–326. (in Chinese) |
[37] |
X. Liu, J. Wang, Y. Li, Research on the co-evolution of competitive public opinion and intervention strategy based on Markov process, J. Inform. Sci., (2023), 01655515221141033. https://doi.org/10.1177/01655515221141033 doi: 10.1177/01655515221141033
![]() |
n | α | β | α−β | 2n |
10 | 0.8107 | 0.5735 | 0.2372 | 0.200 |
11 | 0.7899 | 0.5566 | 0.2333 | 0.182 |
12 | 0.7804 | 0.5438 | 0.2366 | 0.167 |
13 | 0.7728 | 0.5332 | 0.2396 | 0.154 |
14 | 0.7666 | 0.5248 | 0.2418 | 0.143 |
15 | 0.7612 | 0.5176 | 0.2436 | 0.133 |
16 | 0.7566 | 0.5116 | 0.2450 | 0.125 |
17 | 0.7526 | 0.5064 | 0.2462 | 0.118 |
18 | 0.7491 | 0.5020 | 0.2471 | 0.111 |