Perturbation analysis of the effective conductivity of a periodic composite

  • We consider the effective conductivity λeff of a periodic two-phase composite obtained by introducing into an infinite homogeneous matrix a periodic set of inclusions of a different material. Then we study the behavior of λeff upon perturbation of the shape of the inclusions, of the periodicity structure, and of the conductivity of each material.

    Citation: Paolo Luzzini, Paolo Musolino. Perturbation analysis of the effective conductivity of a periodic composite[J]. Networks and Heterogeneous Media, 2020, 15(4): 581-603. doi: 10.3934/nhm.2020015

    Related Papers:

    [1] Na Pang . Nonlinear neural networks adaptive control for a class of fractional-order tuberculosis model. Mathematical Biosciences and Engineering, 2023, 20(6): 10464-10478. doi: 10.3934/mbe.2023461
    [2] Shanjing Ren . Global stability in a tuberculosis model of imperfect treatment with age-dependent latency and relapse. Mathematical Biosciences and Engineering, 2017, 14(5&6): 1337-1360. doi: 10.3934/mbe.2017069
    [3] Georgi Kapitanov . A double age-structured model of the co-infection of tuberculosis and HIV. Mathematical Biosciences and Engineering, 2015, 12(1): 23-40. doi: 10.3934/mbe.2015.12.23
    [4] Silvia Martorano Raimundo, Hyun Mo Yang, Ezio Venturino . Theoretical assessment of the relative incidences of sensitive andresistant tuberculosis epidemic in presence of drug treatment. Mathematical Biosciences and Engineering, 2014, 11(4): 971-993. doi: 10.3934/mbe.2014.11.971
    [5] Abba B. Gumel, Baojun Song . Existence of multiple-stable equilibria for a multi-drug-resistant model of mycobacterium tuberculosis. Mathematical Biosciences and Engineering, 2008, 5(3): 437-455. doi: 10.3934/mbe.2008.5.437
    [6] C. Connell McCluskey . Global stability for an SEI epidemiological model with continuous age-structure in the exposed and infectious classes. Mathematical Biosciences and Engineering, 2012, 9(4): 819-841. doi: 10.3934/mbe.2012.9.819
    [7] Tao-Li Kang, Hai-Feng Huo, Hong Xiang . Dynamics and optimal control of tuberculosis model with the combined effects of vaccination, treatment and contaminated environments. Mathematical Biosciences and Engineering, 2024, 21(4): 5308-5334. doi: 10.3934/mbe.2024234
    [8] Juan Pablo Aparicio, Carlos Castillo-Chávez . Mathematical modelling of tuberculosis epidemics. Mathematical Biosciences and Engineering, 2009, 6(2): 209-237. doi: 10.3934/mbe.2009.6.209
    [9] Xi-Chao Duan, Xue-Zhi Li, Maia Martcheva . Dynamics of an age-structured heroin transmission model with vaccination and treatment. Mathematical Biosciences and Engineering, 2019, 16(1): 397-420. doi: 10.3934/mbe.2019019
    [10] Surabhi Pandey, Ezio Venturino . A TB model: Is disease eradication possible in India?. Mathematical Biosciences and Engineering, 2018, 15(1): 233-254. doi: 10.3934/mbe.2018010
  • We consider the effective conductivity λeff of a periodic two-phase composite obtained by introducing into an infinite homogeneous matrix a periodic set of inclusions of a different material. Then we study the behavior of λeff upon perturbation of the shape of the inclusions, of the periodicity structure, and of the conductivity of each material.



    The graphs considered throughout this paper are finite and connected. The graph-theoretical terminology and notation that are used in this study without explaining here can be found in the books [1,2].

    A graph invariant is a function f defined on the set of all graphs with the condition that f(G1)=f(G2) whenever G1 and G2 are isomorphic. The real-valued graph invariants are commonly known as topological indices in mathematical chemistry, particularly in chemical graph theory [20].

    The Sombor index and the reduced Sombor index abbreviated as SO and SOred, respectively, are the topological indices introduced recently by mathematical chemist Ivan Gutman in his seminal paper [9]. The Sombor index SO has attained attention from many scientific groups all over the world in a very short time, which resulted in many publications; for example, see the review papers [8,15] and associated papers listed therein. The chemical applicability of the indices SO and SOred is also well documented. Redžepović [17] examined the discriminating and predictive ability of the indices SO and SOred on a large class of isomers and found that both of these indices have good discriminating and predictive potential. Deng et al. [7] compared the predictive ability of SO on octane isomers with that of similar kind of existing topological indices and showed that SO has a higher accuracy in predicting physico-chemical properties of the the aforementioned chemical compounds; Liu et al. [13] conducted a similar comparative study for SOred and concluded that SOred outperforms in several cases. Also, it was demonstrated in [14] that boiling points of benzenoid hydrocarbons are highly correlated with SO and SOred. By considering these chemical applications of SO and SOred, it make sense to study further these indices, particularly for molecular graphs (these are graphs of maximum degree at most 4).

    In chemical graph theory, molecular trees play an important role because a certain class of chemical compounds can be viewed by using the concept of molecular trees. Thereby, in the present paper, we study the topological indices SO and SOred for molecular trees under certain constraints. Deng et al. [7] determined the trees possessing the maximum and minimum values of SO and SOred among all molecular trees of a given order; see also [5,18] where the same problem regarding SO was solved independently. Let us consider the following problem.

    Problem 1. Characterize the trees possessing the maximum and minimum values of SO and SOred among all molecular trees of a given order and with a fixed number of (i) branching vertices (ii) vertices of degree 2.

    The minimal part of Problem 1 concerning SO has already been solved in [3] where several other interesting extremal problems were also resolved. The main objective of this study is to give a solution to the maximal part of Problem 1. Detail about the mathematical study of the Sombor index for general trees can be found in [5,6,10,12,19,21,22].

    Let V(G) and E(G) denote the set of vertices and edges, respectively, in the graph G. For the vertex vV(G), the degree of v is denoted by dG(v) (or simply by dv if only one graph is under consideration). A vertex uV(G) is said to be a pendent vertex or a branching vertex if du=1 or du3, respectively. The set NG(u) consists of the vertices of the graph G that are adjacent to the vertex u. Let ni(G) denotes the count of vertices having degree i in the graph G. Denote by xi,j(G) the cardinality of the set consisting of the edges connecting the vertex of degree i with the vertex of degree j in the graph G. A graph of order n is also known as an n-vertex graph.

    For a graph G, the Sombor index and reduced Sombor index abbreviated as SO and SOred, respectively, are defined [9] as

    SO(G)=uvE(G)d2u+d2vandSOred(G)=uvE(G)(du1)2+(dv1)2.

    We start this section with the following elementary result, noted in [16].

    Lemma 1. The function f defined as

    f(x,y)=x2+y2(xc)2+y2,

    with 1c<x and y>0, is strictly decreasing in y and strictly increasing in x.

    For r3, each of the vertices u2,u3,,ur1 of the path P:u1u2ur in a graph is called internal vertex of P. Denote by Pi,j(C) (or simply Pi,j) the path from a branching vertex of degree i to a branching vertex of degree j in a molecular tree C such that all the internal vertices (if exist) of Pi,j have degree 2.

    Lemma 2. If the molecular tree C contains a path P4,4:v1v2vs and an edge uw such that du=1 and dw=3 with the condition that v1 lies on the wvs path, then for the molecular tree C obtained from C by deleting the edges v1v2, uw and adding new edges v2w, v1u, the inequalities SO(C)<SO(C) and SOred(C)<SOred(C) hold, where C and C have the same degree sequence.

    Proof. Clearly, the trees C and C have the same degree sequence. Also, we note that

    SO(C)=SO(C)+1017+4+d2v29+d2v2<SO(C)

    and

    SOred(C)=SOred(C)+23+9+(dv21)24+(dv21)2<SO(C).

    Lemma 3. For a molecular tree C, if x1,3(C)>0 such that C contains the paths P3,4:u1u2ut and P4,3:utut+1us1us, where t2, s3, and t<s, then a molecular tree C can be obtained with x1,3(C)<x1,3(C) such that SO(C)<SO(C) and SOred(C)<SOred(C), where C and C have the same degree sequence.

    Proof. Suppose that uvE(C) such that du=1 and dv=3. We prove the result by considering two possible cases.

    Case I. s4.

    Note that one of the paths P4,3 and P4,3 contains exactly one internal vertex and the other contains no internal vertex. Thus, without loss of generality it can be assumed that t=2. If s=3, then we take C=C{uv,u1u2,u2u3}+{u1u3,uu2,u2v}, and for s=4, we take C=C{u1u2,u3u4,uv}+{u1u4,uu2,u3v}. In either case, we note that both C and C have the same degree sequence, x1,3(C)<x1,3(C), and

    SO(C)=SO(C)+5+103217<SO(C)
    SOred(C)=SOred(C)+13122<SOred(C),

    as required.

    Case II. s>4.

    If t=2 or t=s1, the result is proved in a fully analogous way as in Case I. In what follows, suppose that 2<t<s2. When s=5 then (t=3 and) we take C=C{u2u3,u4u5,uv}+{u2u5,u3u,u4v}, and otherwise (that is, if s6 then) we take C=C{u2u3,ut1ut,us1us,uv}+{u2us,uut,us1u3,ut1v}. In either case, we observe that both C and C have the same degree sequence, x1,3(C)<x1,3(C), and

    SO(C)=SO(C)+10+251317<SO(C),
    SOred(C)=SOred(C)+1015<SOred(C),

    as desired.

    Lemma 4. Let C be a molecular tree with xy,uv,wzE(C), where the vertices x, y, u, v, w, z are chosen in such a way that du=1, dx=3=dy, dv=dw=dz=4, and w lies on one of the three paths vz, xz, yz paths, and that x1,3(C)=0. If NC(z)={w,z1,z2,z3} and C=C{xy,z1z,z2z,z3z,uv}+{xu,uy,z1u,z2u,z3v} then SO(C)<SO(C) and SOred(C)<SOred(C), where C and C have the same degree sequence.

    Proof. It can be easily observed that C and C have the same degree sequence. Also, after elementary calculations, we arrive at SO(C)=SO(C)+7210<SO(C) and SOred(C)=SOred(C)+52213<0.

    Lemma 5. Let C be a molecular tree with uv,zw,wzE(C), where the vertices u,v,w,z,zV(C) are chosen in such a way that dw=2 and min{du,dv,dz,dz}3 provided that du+dv>dz+dz. If C=C{uv,zw,wz}+{zz,uw,wv}, then SO(C)<SO(C) and SOred(C)<SOred(C), where C and C have the same degree sequence.

    Proof. We observe that both the trees C and C have the same degree sequence and

    SO(C)=SO(C)+d2u+d2v+4+d2z+4+d2z4+d2v4+d2ud2z+d2z=SO(C)+I1

    and

    SOred(C)=SOred(C)+(du1)2+(dv1)2+1+(dz1)2+1+(dz1)21+(du1)21+(dv1)2(dz1)2+(dz1)2=SOred(C)+I2,

    where

    I1=d2u+d2v+4+d2z+4+d2z4+d2v4+d2ud2z+d2z

    and

    I2=(du1)2+(dv1)2+1+(dz1)2+1+(dz1)21+(du1)2 1+(dv1)2(dz1)2+(dz1)2.

    To complete the proof, it is enough to show that I1<0 and I2<0. Recall that du+dv{7,8} and hence we consider two cases accordingly.

    ⅰ) If du+dv=8 then du=4=dv and hence dz+dz{6,7}, and therefore we get I1<0 and I2<0

    ⅱ) If du+dv=7 then du=3 and dv=4 (or du=4 and dv=3) and hence dz=3=dz and thence we have I1<0 and I2<0.

    Lemma 6. Let C be a molecular tree with uv,vw,xyE(C), where the vertices u,v,w,x,y are chosen in such a way that du=2=dv, dw2, and min{dx,dy}3. If C=C{uv,vw,xy}+{uw,xv,vy} then SO(C)<SO(C) and SOred(C)<SOred(C), where C and C have the same degree sequence.

    Proof. Clearly, the degree sequences of C and C is the same. Since dx,dy{3,4}, we have

    SO(C)=SO(C)+22+d2x+d2y4+d2x4+d2y<SO(C)

    and

    SOred(C)=SOred(C)+2+(dx1)2+(dy1)2 1+(dx1)21+(dy1)2<SOred(C).

    For n4, denote by Cn,nb the collection of all n-vertex molecular trees with nb branching vertices, where nb12(n2). As the path Pn is the unique graph in Cn,0, where SO(Pn)=22(n3)+25 and SOred(Pn)=2(n3)+2, and C1 (depicted in Figure 1) is the unique graph in C4,1 (whose (reduced) Sombor index value is given in Table 1), in rest of the investigation we assume that n>4 and nb1. We also define the sub-classes Cp and Cq of Cn,nb as follows:

    Cp={CCn,nb:n2(C)=0,n3(C)0}, (4.1)
    Cq={CCn,nb:n3(C)=0,n2(C)0}. (4.2)
    Figure 1.  The molecular trees C1,C2,,C6.
    Table 1.  The (reduced) Sombor indices of the trees C1,C2,,C6.
    SO(Ci) SOred(Ci)
    C1 310 6
    C2 417 12
    C3 22(n5)+210+13+5 2(n5)+5+5
    C4 22(n6)+3(17+5) 2(n6)+10+10
    C5 410+32 22+8
    C6 210+317+5 13+13

     | Show Table
    DownLoad: CSV

    Lemma 7. If Cb is a molecular tree with the maximum Sombor index (respectively, reduced Sombor index) over the class Cn,nb, then either CbCp or CbCq.

    Proof. Contrarily, assume that CbCn,nb(CpCq). There must be vertices u and v in Cb with du=3 and dv=2. Let Nu(Cb)={u1,u2,u3} and Nv(Cb)={v1,v2}, where v2 and u3 lie on the uv path. If uvE(C) then we take v2=u and u3=v, and also it possible that u3=v2. If C is the tree deduced from C by dropping the edges uu1, uu2 and adding two new edges vu1, vu2 in C, then CCn,nb and (by keeping in mind Lemma 1, we have)

    SO(Cb)=SO(C)+3i=19+d2ui+2j=14+d2vj1+d2u32i=116+d2ui2j=116+d2vj<SO(C)+9+d2u31+d2u3+2j=1(4+d2vj16+d2vj)SO(C)+2j=1(4+d2vj16+d2vj)+135SO(C)+35+1382<SO(C),

    which is contradicting our assumption concerning the choice of Cb.

    Similarly, we have

    SOred(Cb)=SOred(C)+3i=14+(dui1)2+2j=11+(dvj1)2(du31)22i=19+(dui1)22j=19+(dvj1)2<SOred(C)+4+(du31)2(du31)2+2j=11+(dvj1)22j=19+(dvj1)2SOred(C)+2j=11+(dvj1)22j=19+(dvj1)2+51SOred(C)+5+210621<SOred(C),

    a contradiction.

    Lemma 8. For nb>1, if Cb is the tree with the maximum Sombor index (respectively reduced Sombor index) in the class Cn,nb, then x1,2(Cb)=0.

    Proof. Contrariwise, assume that there is a path P:v1v2vt in Cb, where t3 such that dv1=1, dvt>2 and dvj=2 for all 2jt1. Since n2(Cb)1, by Lemma 7 we must have dvt=4. Also, at least one of the neighbors of vt different from vt1 is non-pendent because nb>1. Take wNCb(vt){vt1} such that dw2. As n3(Cb)=0 by Lemma 7, we have dw{2,4}. If C=Cb{v1v2,vt1v1,uw}+{v1vt,uv2,vt1w}, then we have CCn,nb and

    SO(Cb)=SO(C)+517+16+d2w4+d2w<SO(C),

    and

    SOred(Cb)=SOred(C)+13+9+(dw1)21+(dw1)2<SOred(C).

    a contradiction.

    For a non-trivial molecular tree C of order n, the following identities hold:

    n=n1(C)+n2(C)+n3(C)+n4(C), (4.3)
    n1(C)+2n2(C)+3n3(C)+4n4(C)=2(n1), (4.4)
    nb=n3(C)+n4(C). (4.5)

    The results in the following lemma directly follows from Eqs (4.3)–(4.5).

    Lemma 9. For a molecular tree CCn,nb, the following statements hold:

    (i) If CCp, then n1(C)=nnb, n3(C)=3nbn+2 and n4(C)=n2nb2.

    (ii) If CCq, then n1(C)=2nb+2, n2(C)=n3nb2 and n4(C)=nb.

    (iii) CCpCq if and only if nb=n23.

    Lemma 10. For a molecular tree C with the maximum Sombor index (respectively reduced Sombor index) over the class Cn,nb, n2(C)1 if and only if 0nb<n23 or n3nb+3.

    Proof. If n2(C)1, then n3(C)=0 by Lemma 7 and hence CCq; now, by using the result n2(C)=n3nb2 of Lemma 9 it is deduced that nb<n23 or n3nb+3.

    Conversely, suppose that 0nb<n23 or n3nb+3 with nb0. Contrarily, assume that n2(C)=0. From Eqs (4.3)–(4.5), it follows that 2nb+n4(C)=n2, which together with the assumption n3nb+3 implies that n4(C)nb+1, which is a contradiction.

    Note that C2 and C3 with n=5 (see Figure 1) are the only molecular trees in C5,1, and it can be easily observed that SO(C2)>SO(C3) (respectively SOred(C2)>SOred(C3)); see Table 1. For n5, we define the the following sub-classes of Cn,nb:

    B0={CCq: nb=1 and x1,2(C)=1},

    B1={CCq: 1<nb<n14 and x1,2(C)=0=x4,4(C)},

    B2={CCq: n14nb<n23 and x1,2(C)=0=x2,2(C)},

    B3={CCp: n23<nb3n78 and x1,3(C)=0=x3,3(C)},

    B4={CCp: 3n78<nb2n65 and x1,3(C)=0=x4,4(C) and x3,30},

    B5={CCp: 2n65<nbn22 and x4,4=0 and x3,3=n3(C)1},

    where Cp and Cq are defined in (4.1) and (4.2). For a molecular tree C, we have

    1j4;jixi,j(C)+2xi,i(C)=iniwherei=1,2,3,4. (4.6)

    Theorem 1. If n6 and CCn,1, then

    i) SO(C)22(n6)+35+317,

    ii) SOred(C)2(n6)+10+10.

    The equalities occur if and only if CB0.

    Proof. Let n>5 and C1Cn,1 be a molecular tree with the maximum Sombor index (respectively reduced Sombor index). Let v be the unique branching vertex of C1.

    Claim 1. dv=4.

    Contrariwise, suppose that dv=3. The constraint n>5 ensures that there is a vertex u of degree 2 in C1 which is adjacent to a pendent vertex, say w. Let C be the new tree obtained from C1 by deleting the edge uw and adding the new edge vw. Certainly, we have CCn,1.

    Note that

    SO(C1)=SO(C)+zNC1(v)9+d2z+ξzNC1(v)16+d2z17<SO(C),

    and

    SOred(C1)=SOred(C)+zNC1(v)4+(dz1)2+ηzNC1(v)9+(dz1)23.<SOred(C),

    which leads to the contradiction to our assumption concerning C1, where

    ξ={5 if uvV(C1),8 if uvV(C1),andη={1 if uvV(C1),2 if uvV(C1).

    Claim 2. The vertex v has exactly one non pendent neighbor.

    Since n6, the vertex v has at least one non pendent neighbor. Contrarily assume that P1:v1v2vrv and P2:w1w2wsv are two paths in C1 with dv1=1=dw1 and dvi=2=dwj for 2ir and 2js. If C=C1{v2v1,vrv}+{v1v,vrw1}, then we have CCn,1 and

    SO(C1)=SO(C)+4+d2v1+d2v+58<SO(C),

    and

    SOred(C1)=SOred(C)+1+(dv1)2(dv1)2+12<SOred(C),

    a contradiction (where dv=4 by Claim 1).

    Now, the desired result follows from Claims 1 and 2.

    Theorem 2. If CCn,nb and 1<nb<n14, then

    (i) SO(C)22(n1)+217(nb+1)+45(nb1)82nb,

    (ii) SOred(C)2(n4nb1)+210(nb1)+6(nb+1),

    and the equalities occur if and only if CB1.

    Proof. Denote by Cb the molecular tree with the maximum Sombor index (respectively reduced Sombor index) in the class Cn,nb, for 1<nb<n14. Since nb<n14<n23, Lemma 10 ensures that n2(Cb)>0, which together with Lemma 7 implies that CbCq, and hence by Lemma 9(b) one has n1(Cb)=2nb+2, n2(Cb)=n3nb2 and n4(Cb)=nb. Because of the constraint nb>1, Lemma 8 guaranties that

    x1,2(Cb)=0, (4.7)

    plugging it into (4.6) for i=1, we get

    x1,4(Cb)=n1(Cb)=2nb+2. (4.8)

    Since nb<n14 or 4nb<n1 which gives n4(Cb)1=nb1<n3nb2=n2(Cb) and therefore

    n4(Cb)n2(Cb). (4.9)

    We claim that x2,2(Cb)0. Contrarily, assume that x2,2(Cb)=0. Then, (4.6) with i=2 gives

    x2,4(Cb)=2n2(Cb). (4.10)

    Equations (4.3) and (4.4) implies that

    n1(Cb)2n4(Cb)=2 (4.11)

    Also, (4.6) with i=4 yields

    2x4,4(Cb)=4n4(Cb)x1,4(Cb)x2,4(Cb) (4.12)

    Using (4.8)–(4.11) in (4.12), we have

    2x4,4(Cb)=4n4(Cb)n1(Cb)2n2(Cb)4n4(Cb)n1(Cb)2n4(Cb)=2,

    a contradiction. Hence, the claim x2,2(Cb)0 is true.

    We also claim that

    x4,4(Cb)=0. (4.13)

    Suppose to the contrary that x4,4(Cb)0. Take xyE(Cb) such that dx=dy=4. Since x2,2(Cb)0, take uv,vwE(Cb) such that du=dv=2 and dw2. If C is the tree obtained by applying the transformation mentioned in the statement of Lemma 6, then Lemma 6 guaranties that SO(Cb)<SO(C) and SOred(Cb)<SOred(C), which is a contradiction to the definition of Cb. Therefore, x4,4(Cb)=0.

    Now, by using Eqs (4.6)–(4.8) and (4.13), we get x2,4(Cb)=2nb2 and x2,2(Cb)=n4nb1. Hence, SO(Cb)=22(n1)+217(nb+1)+45(nb1)82nb and SOred(Cb)=2(n4nb1)+210(nb1)+6(nb+1), which completes the proof.

    Theorem 3. If CCn,nb such that n14nb<n23, then

    i) SO(C)45(n3nb2)42(n4nb1)+217(nb+1),

    ii) SOred(C)210(n3nb2)32(n4nb1)+6(nb+1),

    and the equalities occur if and only if CB2.

    Proof. Denote by Cb the molecular tree with the maximum Sombor index (respectively reduced Sombor index) from the class Cn,nb for n14nb<n23. By Lemma 10, the inequality n2(Cb)>0 holds for nb<n23 and ultimately we have CbCq or n3(Cb)=0 because of Lemma 7. The equations n1(Cb)=2nb+2, n2(Cb)=n3nb2 and n4(Cb)=nb follow from (b) part of Lemma 9, and (4.7) and (4.8) hold by Lemma 8. Also, the inequality n2(Cb)n4(Cb)1 is obtained from n14nb. By using the method used in the proof of Theorem 2, we get x2,2(Cb)=0, x2,4(Cb)=2n2(Cb)=2n6nb4 and x4,4(Cb)=nb1n2(Cb)=4nbn+1. Hence, CbB2 or SO(Cb)=45(n3nb2)42(n4nb1)+217(nb+1) and SOred(Cb)=210(n3nb2)32(n4nb1)+6(nb+1), which completes the proof.

    Theorem 4. Let CCn,nb be the molecular tree such that nb=n23, then

    i) SO(C)=2173(n+1)+423(n5),

    ii) SOred(C)=2(n+1)+2(n5).

    Proof. The part (c) of Lemma 9 ensures that CCpCq, which implies that n3(C)=0=n2(C). Thus, x1,4(C)=nnb=23(n+1) and x4,4(C)=nb1=13(n5).

    Theorem 5. If CCn,nb such that n23<nb3n78, then

    i) SO(C)17(nnb)+5(9nb3n+6)+42(3n8nb7),

    ii) SOred(C)3(nnb)+13(9nb3n+6)+32(3n8nb7),

    and the equalities occur if and only if CB3.

    Proof. Denote by Cb the molecular tree with the maximum Sombor index (respectively reduced Sombor index) in the class Cn,nb, for n23<nb3n78. By Lemma 10, we have n2(Cb)=0 (as n23<nb), which implies that CbCp. Also, Lemma 9(a) guaranties that n1(Cb)=nnb, n3(Cb)=3nbn+2 and n4(Cb)=n2nb2. Now, the inequality nb3n78 can be written as 6nb2n+5<n2nb2, which leads us to the fact that 2n3(Cb)+1n4(Cb). Using Lemmas 2–4 and keeping in mind the fact n4(Cb)2n3(Cb)+1, we have

    x1,3(Cb)=0 (4.14)

    and

    x1,4(Cb)=nnb. (4.15)

    Now, using Lemmas 2–4 and Eqs (4.6), (4.14) and (4.15), we have x3,3(Cb)=0, x3,4(Cb)=3n3(Cb)=9nb3n+6, x4,4(Cb)=n4(Cb)2n3(Cb)1=3n8nb7. Hence, SO(Cb)=17(nnb)+5(9nb3n+6)+42(3n8nb7), and SOred(Cb)=3(nnb)+13(9nb3n+6)+32(3n8nb7). This completes the proof.

    Theorem 6. If CCn,nb such that 3n78<nb2n65, then

    i) SO(C)17(nnb)+32(8nb3n+7)+5(3n7nb8),

    ii) SOred(C)3(nnb)+22(8nb3n+7)+13(3n7nb8),

    and the equalities occur if and only if CB4.

    Proof. Denote by Cb the molecular tree with the maximum Sombor index (respectively reduced Sombor index) from the class Cn,nb for 3n78<nb2n65. By using Lemma 10 it is easy to check that n2(Cb)=0 as n23<3n78<nb, which implies that CbCp and further (a) part of Lemma 9 concludes that n1(Cb)=nnb, n3(Cb)=3nbn+2 and n4(Cb)=n2nb2. Note that nb2n65 can be easily written as 3nbn+4n2nb2, which leads us to the fact n3(Cb)+2n4(Cb).

    From Lemmas 2–4 it is clear that we have to place the vertices as described in the proof of Theorem 5. Keeping in mind the fact 2n3(Cb)+1>n4(Cb)n3(Cb)+2 and Eqs (4.6), (4.14) and (4.15), we have x3,3(Cb)=2n3(Cb)+1n4(Cb)=8nb3n+7, x3,4(Cb)=n3(Cb)+2+2(n4(Cb)(n3(Cb)+2))=3n7nb8 and x4,4(Cb)=0. Hence, SO(Cb)=17(nnb)+32(8nb3n+7)+5(3n7nb8) and SOred(Cb)=3(nnb)+22(8nb3n+7)+13(3n7nb8), which completes the proof.

    Theorem 7. If CCn,nb such that 2n65<nbn22, then

    i) SO(C)(5+317)(n2nb2)10(2n5nb6)32(n3nb1),

    ii) SOred(C)(5n8nb6)+13(n2nb2)22(n3nb1),

    and the equalities occur if and only if CbB5.

    Proof. Denote by Cb the molecular tree with the maximum Sombor index (respectively reduced Sombor index) from the class Cn,nb for 2n65<nbn22. By using Lemma 10 it is easy to check that n2(Cb)=0 as n23<2n65<nb, which implies that CbCp and further (a) part of Lemma 9 concludes that n1(Cb)=nnb, n3(C)=3nbn+2 and n4(Cb)=n2nb2. Note that nb>2n65 can be easily written as 3nbn+4>n2nb2, which leads us to the fact n3(Cb)+2>n4(Cb). From Lemmas 2–4 it is clear that we have to place the vertices of degree 4 between the pendent vertices and the vertices of degree 3. The fact n3(Cb)+2>n4(Cb) gives the result

    x1,4(Cb)=3n4(Cb)=3n6nb6, (4.16)

    and

    x1,3(Cb)=5nb2n+6. (4.17)

    Now, using (4.6), (4.16) and (4.17), we have x3,3(Cb)=n3(Cb)1=3nbn+1, x3,4(Cb)=n4(Cb)=n2nb2 and x4,4(Cb)=0. Hence, SO(Cb)=(5+317)(n2nb2)10(2n5nb6)32(n3nb1) and SOred(Cb)=(5n8nb6)+13(n2nb2)22(n3nb1), which completes the proof.

    Denote by Cn,q the class of all n-vertex molecular trees, where q is the number of vertices of degree 2. Next, we are going to obtain the upper bounds for the molecular trees with respect to Sombor index and reduced Sombor index from the collection of molecular trees Cn,q for 0qn2. It is obvious that the path graph Pn is the unique graph for q=n2, and there does not exist any graph corresponding to the value q=n3 in the collection Cn,q. Hence, we proceed with the assumption that 0qn4.

    Lemma 11. If the molecular tree CCn,q is the tree with the maximum Sombor index (respectively reduced Somber index), then n3(C)2.

    Proof. Suppose contrarily that n3(C)>2, and there are vertices u,w and z of degree 3 such that w is located on the unique uz path. Consider Nz(C)={z1,z2,z3} with the assumption that the vertex z3 lies on the uz path in C (z3 may coincide with w). Now, a tree C can be obtained from the collection Cn,q such as C=C{zz1,zz2}+{uz1,wz2}, which gives the following result:

    SO(C)=SO(C)+xNC(u)d2x+9+yNC(w)d2y+9+3i=1d2zi+9xNC(u)d2x+16yNC(w)d2y+162i=1d2zi+16d2z3+1<SO(C)+xNC(u)d2x+9+3i=1d2zi+9xNC(u)d2x+162i=1d2zi+16d2z3+1SO(C)+5(5)202+102<SO(C),

    a contradiction for the chosen C.

    Similarly

    SOred(C)=SOred(C)+xNC(u)(dx1)2+4+yNC(w)(dy1)2+4+3i=1(dzi1)2+4xNC(u)(dx1)2+9yNC(w)(dy1)2+92i=1(dzi1)2+9(dz31)2<SOred(C)+513152+2<SOred(C),

    which also leads to a contradiction.

    Lemma 12. If CCn,q, then

    i) n3(C)=0 if and only if nq20 (mod 3), n4(C)=nq23 and n1(C)=23(nq+1),

    ii) n3(C)=1 if and only if nq10 (mod 3), n4(C)=nq43 and n1(C)=23(nq1)+1,

    iii) n3(C)=2 if and only if nq0 (mod 3), n4(C)=nq63 and n1(C)=23(nq).

    Proof. The following equation can be drawn by using Eqs (4.3) and (4.4):

    n1(C)=n3(C)+2n4(C)+2. (4.18)

    Now, using Eqs (4.3) and (4.18), we get

    nq22n3(C)=3n4(C) (4.19)

    or

    nq22n3(C)0(mod3). (4.20)

    By solving the Eqs (4.4) and (4.18) for the values of n1(C) and n4(C), we get

    n1(C)=2n2q+2n3(C)3, (4.21)

    and

    n4(C)=nq22n3(C)3. (4.22)

    The required results are directly followed by Eqs (4.20)–(4.22).

    Lemma 13. If CCn,q is the tree with the maximum Sombor index (respectively reduced Somber index) with n4qn5, then x1,2(C)1.

    Proof. Note that there is a unique branching vertex (say) u in C with the fact that du=3 if q=n4, and du=4 if q=n5. Let us assume that, for l,m3, there are paths u1u2ulu and u1u2umu in C such that dui=2=duj for all 2il and 2jm, and du1=1=du1. If a tree C in the class Cn,q is chosen as C=C{ul1ul}+{ul1u1}, then

    SO(C)=SO(C)+4+d2u1+d2u+522,

    and

    SOred(C)=SOred(C)+1+(du1)2(du1)2+12,

    by using the fact du=3 or du=4 in the above mentioned results it can easily be checked that SO(C)<SO(C) and SOred(C)<SOred(C), which is a contradiction.

    Lemma 14. If CCn,q is the tree with the maximum Sombor index (respectively reduced Somber index) such that qn6, then x1,2(C)=0.

    Proof. The results can be proved by using the same process as done in Lemma 8.

    Note that C4,0 and C5,0 contain unique trees C1 and C2, respectively, with the Sombor and reduced Sombor index values given in Table 1. Further more, by using Lemma 13 it can be observed that C5 and C6 are the molecular trees with maximum (reduced) Somber index values (given in Table 1) among the graphs in C6,0 and C7,0, respectively. Among all the molecular trees Cn,q, where q1, If we consider C3={CCn,q:x1,3(C)=2,x1,2(C)=1,x2,3(C)=1,x2,2(C)=n5} and C4={CCn,q:x1,4(C)=3,x1,2(C)=1,x2,4(C)=1,x2,2(C)=n6} (given in Figure 1), the following result is observed:

    Theorem 8. For the molecular tree CCn,q, where n5, the following results hold:

    a) If CCn,n4C3, then SO(C)<SO(C3).

    b) If CCn,n5C4, then SO(C)<SO(C4).

    Proof. Using Lemma 13, it can be concluded that C3 among the class Cn,n4 and C4 among Cn,n5, respectively, contain the maximum (reduced) Sombor index value (given in Table 1), which completes the proof.

    Consider the following subsets of Cn,q:

    Q0={CCn,q: q<n5 and n3(C)=0 such that x1,4(C)=23(nq+1) and x2,2(C)=0 whenever x4,4(C)0},

    Q1={CCn,q: q<n5 and n3(C)=1 such that x1,2(C)=0 and x1,3(C)0 implies that there does not exist P4,4 in C moreover if x2,j(C)0 for 2j3, then x4,4(C)=0},

    Q2={CCn,q: n3(C)=2 such that x1,2(C)=0, x1,3(C)0 P4,4=0 and P3,30, and whenever x1,3(C)=0 along with P3,30, then P4,4=0, furthermore x2,i(C)0 xj,k=0, where 2i3 and 3j,k4}.

    Theorem 9. Let CCn,q for q<n5 such that n3(C)=0. Then

    SO(C){2173(nq+1)+423(n4q5)+45qifq<n54,2173(nq+1)+453(nq5)223(n4q5)ifqn54.

    And

    SOred(C){2(n4q5)+2(nq+1)+210qifq<n54,2(nq+1)+2103(nq5)23(n4q5)ifqn54.

    Equalities hold if and only if CQ0.

    Proof. Denote by Cq0 the molecular tree having maximum Sombor index (or reduced Sombor index) among the collection Cn,q such that n3(Cq0)=0. By using Lemma 12 it follows that nq20 (mod 3), n4(Cq0)=nq23 and n1(Cq0)=23(nq+1). The vertices of degree 2 are to be placed according to the conditions proved in Lemmas 6 and 14, that is all the pendent vertices are to be attached to the vertices of degree 4 i.e., Lemma 14 concludes that

    x1,2(Cq0)=0, (4.23)

    which implies that

    x1,4(Cq0)=23(nq+1). (4.24)

    Further more, the vertices of degree 2 are to be placed between the vertices of degree 4 in such a way that if there is an edge connecting the vertices of degree 4, then no two vertices of degree 2 are adjacent. The above discussion leads us to the fact that Cq0Q0. Now the following two cases arise here:

    Case I. n2(Cq0)<n4(Cq0)1 or q<n54.

    In this case, Lemma 6 implies that

    x2,2(Cq0)=0. (4.25)

    By using (4.6), (4.23)–(4.25), we have x2,4(Cq0)=2q and x4,4(Cq0)=n4q53. Hence, SO(Cq0)=2173(nq+1)+423(n4q5)+45q and SOred(Cq0)=2(n4q5)+2(nq+1)+210q.

    Case II. n2(Cq0)n4(Cq0)1 or qn54.

    In this case, by Lemma 6, we have

    x4,4(Cq0)=0. (4.26)

    Using (4.6), (4.23)–(4.26), we get x2,2(Cq0)=4qn+53 and x2,4(Cq0)=2(n4(Cq0)1)=23(nq5). Hence, SO(Cq0)=2173(nq+1)+453(nq5)223(n4q5) and SOred(Cq0)=2(nq+1)+2103(nq5)23(n4q5), which completes the proof.

    Theorem 10. Let CCn,q such that n3(C)=1. Then

    SO(C){210+317+5ifq=0andn=7,22(q1)+25+210+317+13ifq=n71,617+10(10+1)ifq=0andn=10,617+5+25+13+10ifq=1andn=11,22(q2)+617+45+213+10ifq2andnq=10,(255)q+917+15ifq2andnq=13,(22)(q3)+917+3(25+13)ifq>2andnq=13,173(2n2q+1)+45q+423(n4q13)+15ifq<n134,173(2n2q+1)+253(n+2q13)+133(4qn+13)+53(n4q4)ifn134q<n43,173(2n2q+1)+223(4qn+4)+253(2n2q15)+313ifqn43.

    And

    SOred(C){13(13+1)ifq=0andn=7,2(q1)+10+5+13ifq=n71,20+213ifq=0andn=10,20+13+10+5ifq=1andn=11,2(q2)+20+210+25ifq2andnq=10,13(3q)+27+10qifq2andnq=13,2(q3)+27+3(5+10)ifq>2andnq=13,(2n2q+1)+210q+2(n4q13)+313ifq<n134,(2n2q+1)+103(n+2q13)+53(4qn+13)+133(n4q4)ifn134q<n43,(2n2q+1)+23(4qn+4)+103(2n2q15)+313ifqn43.

    Equalities hold if and only if CQ1.

    Proof. If Cq1 denotes the molecular tree having maximum Sombor index (or reduced Sombor index) among the collection Cn,q, where n3(Cq1)=1, then by using Lemma 12 we have nq10 (mod 3), n1(Cq1)=2n2q+13, n4(Cq1)=nq43. Note that Lemmas 2, 14 and 5 show that Cq1Q1. This implies that

    x1,2(Cq1)=0, (4.27)

    and the vertices of degree 4 are to be placed in the three neighbors of the vertex of degree 3 in such a way that if a pendent vertex is present in Cq1 which is adjacent to the vertex of degree 3, then there must not exists a P4,4 path in Cq1. The following cases are possible:

    Case 1. n4(Cq1)=1

    Subcase 1.1. q=0.

    In this case n=7, x2,j(Cq1)=0 for all 1j4, x3,4(Cq1)=1, x1,3(Cq1)=2 and x1,4(Cq1)=3. The graph here is C6 given in Figure 1 with (reduced) Sombor index value given in Table 1.

    Subcase 1.2. q>0.

    This holds for n8, by keeping in mind the Lemmas 5 and 6, and using the results in Eqs (4.6) and (4.27), we have x3,4(Cq1)=0, x2,3(Cq1)=1=x2,4(Cq1), x1,3(Cq1)=2, x1,4(Cq1)=3 and x2,2(Cq1)=q1. Hence, SO(Cq1)=22(q1)+25+210+317+13 and SOred=2(q1)+10+5+13.

    Case 2. n4(Cq1)=2.

    Subcase 2.1. q=0 In this case we have n=10, where Lemma 2 shows that

    x3,4(Cq1)=2. (4.28)

    By using Eqs (4.6), (4.27) and (4.28), we have x1,3(Cq1)=1 and x1,4(Cq1)=6, which gives SO(Cq1)=617+10(10+1) and SOred(Cq1)=20+213.

    Subcase 2.2. q=1

    Here n=11, By using Eqs (4.6), (4.27), we have x1,3(Cq1)=1, x1,4(Cq1)=6 and x3,4(Cq1)=1=x2,3(Cq1)=x2,4(Cq1), which gives SO(Cq1)=617+5+25+13+10 and SOred(Cq1)=20+13+10+5.

    Subcase 2.3. q2

    This holds for n12, and by keeping in mind the Lemmas 5, 6 and the results in Eqs (4.6) and (4.27), we have x1,3(Cq1)=1, x1,4(Cq1)=6, x2,4(Cq1)=2=x2,3(Cq1) and x2,2(Cq1)=q2. Hence, SO(Cq1)=22(q2)+617+45+213+10 and SOred(Cq1)=2(q2)+20+210+25.

    Case 3. n4(Cq1)=3

    Lemma 2 gives

    x1,3(Cq1)=0. (4.29)

    Subcase 3.1. q2

    Again Lemmas 5, 6 and the Eqs (4.6), (4.27) and (4.29) imply that, x1,4(Cq1)=9, x2,3(Cq1)=x2,4(Cq1)=q, x2,2(Cq1)=0=x4,4(Cq1) and x3,4(Cq1)=3q. Hence, SO(Cq1)=(255)q+917+15 and SOred(Cq1)=13(3q)+27+10q.

    Subcase 3.2. q>2

    Here x1,4(Cq1)=9, x2,3(Cq1)=x2,4(Cq1)=3, x2,2(Cq1)=q3 and x3,4(Cq1)=0 are obtained by using Lemmas 5, 6 and the Eqs (4.6), (4.27) and (4.29). Hence, SO(Cq1)=(22)(q3)+917+3(25+13) and SOred(Cq1)=2(q3)+27+3(5+10).

    Case 4. n4(Cq1)>3

    Subcase 4.1. q<n134

    It is easy to check that q<n134 implies that q<n4(Cq1)3 or q<x4,4(Cq1). Lemmas 5, 6 and the Eqs (4.6), (4.27) and (4.29) follow the results x2,2(Cq1)=0=x2,3(Cq1), x2,4(Cq1)=q, x3,4(Cq1)=3, x1,4(Cq1)=2n2q+13 and x4,4(Cq1)=n4q133. Hence, SO(Cq1)=173(2n2q+1)+45q+423(n4q13)+15 and SOred(Cq1)=(2n2q+1)+210q+2(n4q13)+313.

    Subcase 4.2. n134q<n43

    It gives n4(Cq1)3q<n4, which implies that

    x4,4(Cq1)=0, (4.30)

    by keeping in mind the Lemmas 5, 6 and using the Eqs (4.27), (4.29) and (4.30) in (4.6), we have x2,2(Cq1)=0, x2,3(Cq1)=4qn+133, x2,4(Cq1)=n+2q133, x1,4(Cq1)=2n2q+13 and x3,4(Cq1)=n4q43. Hence, SO(Cq1)=173(2n2q+1)+253(n+2q13)+133(4qn+13)+53(n4q4) and SOred(Cq1)=(2n2q+1)+103(n+2q13)+53(4qn+13)+133(n4q4).

    Subcase 4.3. qn43

    This imply that qn4(C), and Lemmas 5 and 6 conclude that

    x3,4(Cq1)=0, (4.31)

    and using the Eqs (4.27), (4.29)–(4.31) in (4.6), we have x1,4(Cq1)=2n2q+13, x2,2(Cq1)=4qn+43, x2,3(Cq1)=3 and x2,4(Cq1)=2n2q153. Hence, SO(Cq1)=173(2n2q+1)+223(4qn+4)+253(2n2q15)+313 and SOred(Cq1)=(2n2q+1)+23(4qn+4)+103(2n2q15)+313.

    Theorem 11. Let CCn,q, where n3(C)=2. Then

    SO(C){410+32ifnq=6andq=0,22(q1)+410+213ifnq=6andq>0,10(18n+q)3+(317+5)(nq6)3+32if9nq15andq=0,103(18n+q)+17(nq6)+(13+255)q+32+15if9nq15and1qn64,103(18n+q)+(317+25)3(nq6)+133(nq)+223(4qn+3)if9nq15andq>n64,2173(nq)+(13+25)q+5(4q)+32ifnq=18and1q4,2173(nq)+613+85+22(q5)ifnq=18andq>4,2173(nq)+(13+25)q+5(6q)ifnq=21and0q6,2173(nq)+6(13+25)+22(q6)ifnq=21andq>6,2173(nq)+45q+42(n4q21)3+30ifnq>21andqn214,2173(nq)+13(4qn+21)3+25(n+2q21)3+5(n4q3)3ifnq>21andn214<qn34,2173(nq)+22(4qn+3)3+45(nq12)3+613ifnq>21andq>n34.

    And

    SOred(C){8+22ifnq=6andq=0,2(q1)+8+25ifnq=6andq>0,2(18n+q)3+(9+13)(nq6)3+22if9nq15andq=0,2(18n+q)3+3(nq6)3+(5+1013)q+22+313if9nq15and1qn64,2(18n+q)3+(9+10)(nq6)3+53(nq)+23(4qn+3)if9nq15andq>n64,2(nq)+(5+10)q+13(4q)+22ifnq=18and1q4,2(nq)+65+410+2(q5)ifnq=18andq>4,2(nq)+(5+10)q+13(6q)ifnq=21and0q6,2(nq)+6(5+10)+2(q6)ifnq=21andq>6,2(nq)+210q+2(n4q21)+613ifnq>21andqn214,2(nq)+210q+5(4qn+21)3+10(n+2q21)3+13(n4q3)3ifnq>21andn214<qn34,2(nq)+65+2(4qn+3)3+210(nq12)3ifnq>21andq>n34.

    Equalities hold if and only if CQ2.

    Proof. Let Cq2 denotes the molecular tree with maximum (reduced) Sombor index value among the class Cn,q, where n3(Cq2)=2. By Lemma 12, nq0 (mod 3), n1(Cq2)=2n2q3 and n4(Cq2)=nq63. By Lemma 14 it holds that

    x1,2(Cq2)=0, (4.32)

    and Lemmas 2–6 provide the consequence Cq2Q2. We consider the following cases:

    Case 1. n4(Cq2)=0 or nq=6

    Subcase 1.1. q=0

    Here, using Eq (4.6), we have x1,3(Cq2)=4 and x3,3(Cq2)=1, which give SO(Cq2)=410+32 and SOred(Cq2)=8+22.

    Subcase 1.2. q>0

    By using (4.32) in (4.6), it is easy to get x1,3(Cq2)=4, x2,2(Cq2)=q1, x2,3(Cq2)=2 and x3,3(Cq2)=0, which give SO(Cq2)=22(q1)+410+213 and SOred(Cq2)=2(q1)+8+25.

    Case 2. 1n4(Cq2)3 or 9nq15

    By using Lemmas 2–4 and Eq (4.32), we have

    x1,3(Cq2)=4n4(Cq2)=18n+q3, (4.33)
    x1,4(Cq2)=n1(Cq2)x1,3(Cq2)=nq6. (4.34)

    and

    x4,4(Cq2)=0. (4.35)

    Subcase 2.1. q=0

    Using Eqs (4.32)–(4.35) in (4.6), we have x3,3(Cq2)=1 and x3,4(Cq2)=nq63, which give SO(Cq2)=103(18n+q)+(17+53)(nq6)+32 and SOred(Cq2)=2(18n+q)3+(9+13)(nq6)3+22.

    Subcase 2.2. 1qn64

    This case holds for qn4(Cq2), so by using Lemmas 5, 6 and the Eqs (4.32)–(4.35) in (4.6), we have x3,3(Cq2)=1, x3,4(Cq2)=3q, x2,3(Cq2)=q=x2,4(Cq2) and x2,2(Cq2)=0, which give SO(Cq2)=103(18n+q)+17(nq6)+(13+255)q+32+15 and SOred(Cq2)=2(18n+q)3+3(nq6)3+(5+1013)q+22+313.

    Subcase 2.3. q>n64

    This holds for q>n4(Cq2), and again by keeping in mind the results used in Subcase 2.2, we may easily get x3,3(Cq2)=0=x3,4(Cq2), x2,3(Cq2)=nq3, x2,4(Cq2)=nq63 and x2,2(Cq2)=4qn+33. Hence, SO(Cq2)=103(18n+q)+(317+25)3(nq6)+133(nq)+223(4qn+3) and SOred(Cq2)=2(18n+q)3+(9+10)(nq6)3+53(nq)+23(4qn+3).

    Case 3. nq=18 or n4(Cq2)=4

    In this case, Lemmas 2–4 and Eq. (4.32) show that (4.35) holds and

    x1,3(Cq2)=0, (4.36)

    also

    x1,4(Cq2)=2n2q3. (4.37)

    Consider the following possibilities:

    Subcase 3.1. 0q4

    Keeping in mind Lemmas 5 and 6, also using the Eqs (4.6), (4.32) and (4.35)–(4.37), we have x2,2(Cq2)=0, x2,3(Cq2)=q=x2,4(Cq2), x3,3(Cq2)=1 and x3,4(Cq2)=4q. Hence, SO(Cq2)=2173(nq)+(13+25)q+5(4q)+32 and SOred(Cq2)=2(nq)+(5+10)q+13(4q)+22.

    Subcase 3.2. q>4

    Taking into account the facts used in Subcase 3.1, we may check that x2,2(Cq2)=q5, x2,3(Cq2)=6, x2,4(Cq2)=4 and x3,3(Cq2)=0=x3,4(Cq2). Hence, SO(Cq2)=2173(nq)+613+85+22(q5) and SOred(Cq2)=2(nq)+65+410+2(q5).

    Case 4. nq>18 or n4(Cq2)>4

    Lemmas 2–4 imply that (4.36) and (4.37) hold and

    x3,3(Cq2)=0. (4.38)

    Subcase 4.1. nq=21 and 0q6

    Note that (4.35) holds in this case, further more Lemmas 5, 6 and the Eqs (4.6), (4.32), (4.35) and (4.36)–(4.38), we have x2,2(Cq2)=0, x2,3(Cq2)=q=x2,4(Cq2) and x3,4(Cq2)=6q. Hence, SO(Cq2)=2173(nq)+(13+25)q+5(6q) and SOred(Cq2)=2(nq)+(5+10)q+13(6q).

    Subcase 4.2. nq=21 and q>6

    Taking into account the facts used in Subcase 3.1, we have x2,2(Cq2)=q6, x2,3(Cq2)=6=x2,4(Cq2) and x3,4(Cq2)=0. Hence, SO(Cq2)=2173(nq)+6(13+25)+22(q6) and SOred(Cq2)=2(nq)+6(5+10)+2(q6).

    Subcase 4.3. nq>21 and qn214

    This case implies that qn4(Cq2)5 for which x2,3(Cq2)=0 by Lemma 5. Now using Lemma 6 and Eqs (4.6), (4.32) and (4.36)–(4.38), we have x2,2(Cq2)=0, x2,4(Cq2)=2q, x4,4(Cq2)=n4q213 and x3,4(Cq2)=6. Hence, SO(Cq2)=2173(nq)+45q+42(n4q21)3+30 and SOred(Cq2)=2(nq)+210q+2(n4q21)+613.

    Subcase 4.4. nq>21 and n214<qn34

    (4.35) holds in this case and also using (4.6), (4.32) and (4.36)–(4.38), we have x2,2(Cq2)=0, x2,3(Cq2)=4qn+213, x2,4(Cq2)=n+2q213 and x3,4(Cq2)=n4q33. Hence, SO(Cq2)=2173(nq)+13(4qn+21)3+25(n+2q21)3+5(n4q3)3 and SOred(Cq2)=2(nq)+210q+5(4qn+21)3+10(n+2q21)3+13(n4q3)3.

    Subcase 4.5. nq>21 and q>n34

    Again using Lemmas 5, 6 and the Eqs (4.6), (4.32), (4.35) and (4.36)–(4.38), we have x2,2(Cq2)=4qn+33, x2,3(Cq2)=6, x2,4(Cq2)=2(nq12)3 and x3,4(Cq2)=0. Hence, SO(Cq2)=2173(nq)+22(4qn+3)3+45(nq12)3+613 and SOred(Cq2)=2(nq)+65+2(4qn+3)3+210(nq12)3.

    In this paper, an extremal chemical-graph-theoretical problem concerning the Sombor index and the reduced Sombor index is addressed. Particularly, the problem of characterizing trees possessing the maximum values of the aforementioned two indices from the class of all molecular trees of a given order and with a fixed number of (i) branching vertices (ii) vertices of degree 2, is solved in this paper. A solution to the minimal version of this problem regarding the Sombor index was reported in [3]. It is believed that the minimal version of the problem under consideration regarding the reduced Sombor index is not difficult and can be solved by using the technique used in [3]; nevertheless, it still seems to be interesting to find such a solution. Solving Problem 1 for the multiplicative and exponential versions of the Sombor and reduced Sombor indices (for example, see [11]) is another possible direction for a future work concerning the present study.

    The authors have no conflict of interest.



    [1] G. Allaire, Shape Optimization by the Homogenization Method, Applied Mathematical Sciences, 146. Springer-Verlag, New York, 2002. doi: 10.1007/978-1-4684-9286-6
    [2] H. Ammari and H. Kang, Polarization and Moment Tensors. With Applications to Inverse Problems and Effective Medium Theory, Applied Mathematical Sciences, 162, Springer, New York, 2007.
    [3] Polarization tensors and effective properties of anisotropic composite materials. J. Differential Equations (2005) 215: 401-428.
    [4] Effective parameters of elastic composites. Indiana Univ. Math. J. (2006) 55: 903-922.
    [5] Boundary layer techniques for deriving the effective properties of composite materials. Asymptot. Anal. (2005) 41: 119-140.
    [6] A. Bensoussan, J.-L. Lions and G. Papanicolaou, Asymptotic Analysis for Periodic Structures, Studies in Mathematics and its Applications, 5, North-Holland Publishing Co., Amsterdam-New York, 1978.
    [7] Transport properties of densely packed composites. Effect of shapes and spacings of inclusions. Quart. J. Mech. Appl. Math. (2004) 57: 495-528.
    [8] Increase and decrease of the effective conductivity of two phase composites due to polydispersity. J. Stat. Phys. (2005) 118: 481-509.
    [9] A heat conduction problem of 2D unbounded composites with imperfect contact conditions. ZAMM Z. Angew. Math. Mech. (2015) 95: 952-965.
    [10] Effective conductivity of a composite material with stiff imperfect contact conditions. Math. Methods Appl. Sci. (2015) 38: 4638-4649.
    [11] A composite material with inextensible-membrane-type interface. Math. Mech. Solids (2019) 24: 499-510.
    [12] G. P. Cherepanov, Inverse problems of the plane theory of elasticity, J. Appl. Math. Mech., 38 (1974), 915–931 (1975); translated from Prikl. Mat. Meh., 38 (1974), 963–979 (Russian). doi: 10.1016/0021-8928(75)90085-4
    [13] A. Cherkaev, Variational Methods for Structural Optimization, Applied Mathematical Sciences, 140. Springer-Verlag, New York, 2000. doi: 10.1007/978-1-4612-1188-4
    [14] Stokes flow in a singularly perturbed exterior domain. Complex Var. Elliptic Equ. (2013) 58: 231-257.
    [15] Weakly singular and microscopically hypersingular load perturbation for a nonlinear traction boundary value problem: A functional analytic approach. Complex Anal. Oper. Theory (2011) 5: 811-833.
    [16] A singularly perturbed nonideal transmission problem and application to the effective conductivity of a periodic composite. SIAM J Appl Math. (2013) 73: 24-46.
    [17] Series expansion for the effective conductivity of a periodic dilute composite with thermal resistance at the two-phase interface. Asymptot. Anal. (2019) 111: 217-250.
    [18] K. Deimling, Nonlinear Functional Analysis, Springer-Verlag, Berlin, 1985. doi: 10.1007/978-3-662-00547-7
    [19] P. Drygaś, S. Gluzman, V. Mityushev and W. Nawalaniec, Applied Analysis of Composite Media. Analytical and Computational Results for Materials Scientists and Engineers, , Woodhead Publishing Series in Composites Science and Engineering. Woodhead Publishing, 2020.
    [20] Effective conductivity of unidirectional cylinders with interfacial resistance. Quart. J. Mech. Appl. Math. (2009) 62: 235-262.
    [21] A. Friedman, Partial Differential Equations, Holt, Rinehart and Winston, Inc., New York-Montreal, Que.-London, 1969.
    [22] L. V. Gibiansky and A. V. Cherkaev, Microstructures of composites of extremal rigidity and exact bounds on the associated energy density, in Topics in the mathematical modelling of composite materials, 273–317, Progr. Nonlinear Differential Equations Appl., 31, Birkhäuser Boston, Boston, MA, 1997.
    [23] D. Gilbarg and N. S. Trudinger, Elliptic Partial Differential Equations of Second Order, 2nd Edition, Vol. 224 of Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], Springer-Verlag, Berlin, 1983. doi: 10.1007/978-3-642-61798-0
    [24] S. Gluzman, V. Mityushev and W. Nawalaniec, Computational Analysis of Structured Media, Mathematical Analysis and Its Applications. Academic Press, London, 2018.
    [25] Asymptotics of the effective conductivity of composites with closely spaced inclusions of optimal shape. Quart. J. Mech. Appl. Math. (2005) 58: 84-106.
    [26] Effective conductivity of 2D disk-ring composite material. Math. Model. Anal. (2013) 18: 386-394.
    [27] V. V. Jikov, S. M. Kozlov and O. A. Oleĭnik, Homogenization of Differential Operators and Integral Functionals, Springer-Verlag, Berlin, 1994. doi: 10.1007/978-3-642-84659-5
    [28] Improved algorithm for analytical solution of the heat conduction problem in doubly periodic 2D composite materials. Complex Var. Elliptic Equ. (2015) 60: 1-23.
    [29] Geometric aspects of averaging. Russian Math. Surveys (1989) 44: 91-144.
    [30] Quantitative analysis of the particles distributions in reinforced composites. Composite Structures (2017) 182: 412-419.
    [31] Asymptotic behaviour of the conformal representation of a Jordan domain with a small hole in Schauder spaces. Comput. Methods Funct. Theory (2002) 2: 1-27.
    [32] A domain perturbation problem for the Poisson equation. Complex Var. Theory Appl. (2005) 50: 851-867.
    [33] Perturbation problems in potential theory, a functional analytic approach. J. Appl. Funct. Anal. (2007) 2: 197-222.
    [34] M. Lanza de Cristoforis and P. Musolino, A perturbation result for periodic layer potentials of general second order differential operators with constant coefficients, Far East J. Math. Sci. (FJMS), 52 (2011) 75–120.
    [35] Real analytic dependence of simple and double layer potentials upon perturbation of the support and of the density. J. Integral Equations Appl. (2004) 16: 137-174.
    [36] M. Lanza de Cristoforis and L. Rossi, Real analytic dependence of simple and double layer potentials for the Helmholtz equation upon perturbation of the support and of the density, in Analytic methods of analysis and differential equations: AMADE 2006, Camb. Sci. Publ., Cambridge, 2008,193–220.
    [37] H. Lee and J. Lee, Array dependence of effective parameters of dilute periodic elastic composite, in Imaging, multi-scale and high contrast partial differential equations, 59–71, Contemp. Math., 660, Amer. Math. Soc., Providence, RI, 2016.
    [38] Exact estimates of conductivity of composites formed by two isotropically conducting media taken in prescribed proportion. Proc. Roy. Soc. Edinburgh Sect. A (1984) 99: 71-87.
    [39] Shape analysis of the longitudinal flow along a periodic array of cylinders. J. Math. Anal. Appl. (2019) 477: 1369-1395.
    [40] P. Luzzini, P. Musolino and R. Pukhtaievych, Real analyticity of periodic layer potentials upon perturbation of the periodicity parameters and of the support, in Proceedings of the 12th ISAAC congress (Aveiro, 2019). Research Perspectives. Birkhuser, accepted.
    [41] G. W. Milton, The Theory of Composites, Cambridge Monographs on Applied and Computational Mathematics, Cambridge University Press, Cambridge, 2002. doi: 10.1017/CBO9780511613357
    [42] Schwarz's alternating method in a matrix form and its applications to composites. Appl. Math. Comput. (2019) 356: 144-156.
    [43] Analytical methods for heat conduction in composites. Math. Model. Anal. (2008) 13: 67-78.
    [44] Optimal distribution of the nonoverlapping conducting disks. Multiscale Model. Simul. (2012) 10: 180-190.
    [45] A singularly perturbed Dirichlet problem for the Laplace operator in a periodically perforated domain. A functional analytic approach. Math. Methods Appl. Sci. (2012) 35: 334-349.
    [46] A. A. Novotny and J. Sokołowski, Topological Derivatives in Shape Optimization, Interaction of Mechanics and Mathematics, Springer, Heidelberg, 2013. doi: 10.1007/978-3-642-35245-4
    [47] Asymptotic behavior of the solution of singularly perturbed transmission problems in a periodic domain. Math. Methods Appl. Sci. (2018) 41: 3392-3413.
    [48] R. Pukhtaievych, Effective conductivity of a periodic dilute composite with perfect contact and its series expansion, Z. Angew. Math. Phys., 69 (2018), Paper no. 83, 22 pp. doi: 10.1007/s00033-018-0976-z
    [49] Dipole matrix for the 2D inclusions close to circular. ZAMM Z. Angew. Math. Mech. (2008) 88: 993-999.
    [50] W. P. Ziemer, Weakly Differentiable Functions. Sobolev Spaces and Functions of Bounded Variation, volume 120 of Graduate Texts in Mathematics. Springer-Verlag, New York, 1989. doi: 10.1007/978-1-4612-1015-3
  • This article has been cited by:

    1. Na Pang, Nonlinear neural networks adaptive control for a class of fractional-order tuberculosis model, 2023, 20, 1551-0018, 10464, 10.3934/mbe.2023461
  • Reader Comments
  • © 2020 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(2230) PDF downloads(348) Cited by(12)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog