Mean--field control and Riccati equations

  • We present a control approach for large systems of interacting agents based on the Riccati equation. If the agent dynamics enjoys a strong symmetry the arising high dimensional Riccati equation is simplified and the resulting coupled system allows for a formal mean--field limit. The steady--states of the kinetic equation of Boltzmann and Fokker Planck type can be studied analytically. In case of linear dynamics and quadratic objective function the presented approach is optimal and is compared to the model predictive control approach introduced in [2].

    Citation: Michael Herty, Lorenzo Pareschi, Sonja Steffensen. Mean--field control and Riccati equations[J]. Networks and Heterogeneous Media, 2015, 10(3): 699-715. doi: 10.3934/nhm.2015.10.699

    Related Papers:

    [1] Edil D. Molina, Paul Bosch, José M. Sigarreta, Eva Tourís . On the variable inverse sum deg index. Mathematical Biosciences and Engineering, 2023, 20(5): 8800-8813. doi: 10.3934/mbe.2023387
    [2] Mert Sinan Oz, Roberto Cruz, Juan Rada . Computation method of the Hosoya index of primitive coronoid systems. Mathematical Biosciences and Engineering, 2022, 19(10): 9842-9852. doi: 10.3934/mbe.2022458
    [3] William E. Fitzgibbon . The work of Glenn F. Webb. Mathematical Biosciences and Engineering, 2015, 12(4): v-xvi. doi: 10.3934/mbe.2015.12.4v
    [4] V. R. Kulli, J. A. Méndez-Bermúdez, José M. Rodríguez, José M. Sigarreta . Revan Sombor indices: Analytical and statistical study. Mathematical Biosciences and Engineering, 2023, 20(2): 1801-1819. doi: 10.3934/mbe.2023082
    [5] Xinmei Liu, Xinfeng Liang, Xianya Geng . Expected Value of Multiplicative Degree-Kirchhoff Index in Random Polygonal Chains. Mathematical Biosciences and Engineering, 2023, 20(1): 707-719. doi: 10.3934/mbe.2023032
    [6] Saylé C. Sigarreta, Saylí M. Sigarreta, Hugo Cruz-Suárez . On degree–based topological indices of random polyomino chains. Mathematical Biosciences and Engineering, 2022, 19(9): 8760-8773. doi: 10.3934/mbe.2022406
    [7] Muhammad Akram, Adeel Farooq, Maria Shabir, Mohammed M. Ali Al-Shamiri, Mohammed M. Khalaf . Group decision-making analysis with complex spherical fuzzy $ N $-soft sets. Mathematical Biosciences and Engineering, 2022, 19(5): 4991-5030. doi: 10.3934/mbe.2022234
    [8] Martin J. Blaser . Studying microbiology with Glenn F. Webb. Mathematical Biosciences and Engineering, 2015, 12(4): xvii-xxii. doi: 10.3934/mbe.2015.12.4xvii
    [9] Wanlin Zhu, Minglei Fang, Xianya Geng . Enumeration of the Gutman and Schultz indices in the random polygonal chains. Mathematical Biosciences and Engineering, 2022, 19(11): 10826-10845. doi: 10.3934/mbe.2022506
    [10] Abdul Qadeer Khan, Azhar Zafar Kiyani, Imtiaz Ahmad . Bifurcations and hybrid control in a 3×3 discrete-time predator-prey model. Mathematical Biosciences and Engineering, 2020, 17(6): 6963-6992. doi: 10.3934/mbe.2020360
  • We present a control approach for large systems of interacting agents based on the Riccati equation. If the agent dynamics enjoys a strong symmetry the arising high dimensional Riccati equation is simplified and the resulting coupled system allows for a formal mean--field limit. The steady--states of the kinetic equation of Boltzmann and Fokker Planck type can be studied analytically. In case of linear dynamics and quadratic objective function the presented approach is optimal and is compared to the model predictive control approach introduced in [2].


    In mathematics chemistry and biology, a chemical compound can be represented by a molecular graph by converting atoms to vertices and bonds to edges. One of the primary mission of QSAR/QSPR research is to accurately convert molecular graphs into numerical values. Graph theoretic invariants of molecular graphs are called molecular descriptors which can be utilized to simulate the structural information of molecules, in order to make worthwhile physical and chemical properties of these molecules can be acquired by single numerical values. Such kinds of molecular descriptors are also referred to as topological indices.

    In the chemical literature, various topological indices relying only on vertex degrees of the molecular graphs can be utilized in QSPR/QSAR investigation on account of them can be obtained directly from the molecular architecture, and can be rapidly calculated for generous molecules (see [1,2]), and we call them VDB (vertex–degree–based) topological indices. To be more precise, for designated nonnegative real numbers $ \{\psi_{ij}\} $ $ (1\leq i\leq j\leq n-1) $, a VDB topological index of a an $ n $-order (molecular) graph $ G $ is expressed as

    $ TI(G)=1ijn1mijψij,
    $
    (1.1)

    where $ m_{ij} $ is the amount of edges connecting an $ i $-vertex and a $ j $-vertex of $ G $. A great deal of well–known VDB topological indices can be obtained by different $ \psi_{ij} $ in expression (1.1). We list some VDB topological indices in Table 1.

    Table 1.  Some well-known VDB topological indices.
    $ \psi_{ij} $ name
    $ i+j $ First Zagreb index
    $ \frac{1}{\sqrt{ij}} $ Randić index
    $ \frac{2\sqrt{ij}}{i+j} $ GA index
    $ \sqrt{\frac{i+j-2}{ij}} $ ABC index
    $ \frac{1}{\sqrt{i+j}} $ Sum–connectivity index
    $ \frac{(ij)^{3}}{(i+j-2)^{3}} $ AZI index
    $ \frac{2}{i+j} $ Harmonic index
    $ |i-j| $ Albertson index
    $ \sqrt{i^{2}+j^{2}} $ Sombor index
    $ \frac{ij}{i+j} $ ISI index

     | Show Table
    DownLoad: CSV

    The first Zagreb index [3] is the very first VDB topological index, as powerful molecular structure-descriptors [2], Zagreb indices can describe the peculiarities of the degree of branching in molecular carbon-atom skeleton. Thereafter, many VDB topological indices have been put forward to simulate physical, chemical, biological, and other attributes of molecules [4,5,6,7]. In 2021, Gutman [8] introduced a new VDB topological index named as the Sombor index which has a linear correlation with the entropy and the enthalpy of vaporization of octanes [9]. Das et al., give sharp bounds for Sombor index of graphs by means of some useful graph parameters and they reveal the relationships between the Sombor index and Zagreb indices of graphs [10]. Recently, Steiner Gutman index was introduced by Mao and Das [11] which incorporate Steiner distance of a connected graph $ G $. Nordhaus-Gaddum-type results for the Steiner Gutman index of graphs were given in [12]. In 2022, Shang study the Sombor index and degree-related properties of simplicial networks [13]. For more details of VDB topological indices, one can see [3,14,15,16,17,18,19,20,21,22,23,24,25,26] and the books [27,28,29].

    Fluoranthene is a eminent conjugated hydrocarbon which abound in coal tar [30]. A fluoranthene–type benzenoid system (f-benzenoid for short) is formed from two benzenoid units joined by a pentagon [31,32]. The ordinary structure modality of a f-benzenoid $ F $ is shown in Figure 1, where segments $ X $ and $ Y $ are two benzenoid systems. Each f-benzenoid possesses exactly one pentagon [32]. More and more attention is paid to f-benzenoids after the flash vacuum pyrolysis experiments of these nonalternant polycyclic aromatic hydrocarbons [33].

    Figure 1.  The ordinary structure modality of a f-benzenoid (F) and its construction from two benzenoid systems $ X $ and $ Y $.

    In the whole article, the terminology and notation are chiefly derived from [34,35,36,37,38,39,40,41]. A vertex of degree $ k $ is called a $ k $-vertex, and an edge linking a $ k $-vertex and a $ j $-vertex is designated as a $ (k, j) $-edge. Let $ n_{k} $ be the number of $ k $-vertices and let $ m_{kj} $ be the number of $ (k, j) $-edges in the molecular graph $ G $. A benzenoid system without internal vertices is said to be catacondensed. Analogously, a f-benzenoid $ F $ containing a unique internal vertex is referred to as catacatacondensed. We use $ h $-hexagon benzenoid system (or $ h $-hexagon f-benzenoid) to represent a benzenoid system (or f-benzenoid) containing $ h $ hexagons.

    Let $ L_{h} $ represent the $ h $-hexagon linear chain (as shown in Figure 2(a)). An f-benzenoid $ FL_{h} $ $ (h\geq3) $ obtaining from pieces $ X = L_{2} $ and $ Y = L_{h-2} $ is named as f-linear chain (as shown in Figure 2(b)).

    Figure 2.  Linear chain and f-linear chain.

    A fissure (resp. bay, cove, fjord and lagoon) of a f-benzenoid $ F $ is a path of degree sequences $ (2, 3, 2) $ (resp. $ (2, 3, 3, 2) $, $ (2, 3, 3, 3, 2) $, $ (2, 3, 3, 3, 3, 2) $ and $ (2, 3, 3, 3, 3, 3, 2) $) on the perimeter of $ F $ (see Figure 3). Fissures, bays, coves, fjords and lagoons are said to be different kinds of inlets and their number are signified by $ f $, $ B $, $ C $, $ F_{j} $ and $ L $, respectively [32,37]. Inlets determine many electronic and topological properties of f-benzenoids. Then, it can be found that $ f+2B+3C+4F_{J}+5L $ is the number of $ 3 $-vertices on the perimeter of $ F $. It is noted that lagoons cannot occur in the theory of benzenoid systems. For convenience, let $ r = f+B+C+F_{j}+L $ to represent the total number of inlets and $ b = B+2C+3F_{j}+4L $ is referred to as the quantity of bay regions, In addition, $ b $ is exactly the quantity of $ (3, 3) $-edges on the perimeter of $ F $. It is obvious that $ b\geq2 $ for any f-benzenoid $ F $.

    Figure 3.  Structural features occurring on the perimeter of f-benzenoids.

    It is noted that any f-benzenoid $ F $ contains merely either $ 2 $-vertex or $ 3 $-vertex. The vertices not on the perimeter are said to be internal, and we use $ n_{i} $ to represent their number.

    Lemma 1.1. [32] Let $ F $ be an $ n $-order, $ h $-hexagon f-benzenoid with $ m $ edges and $ n_{i} $ internal vertices. Then

    $ {\rm (i)} $ $ n = 4h+5-n_{i} $;

    $ {\rm (ii)} $ $ m = 5h+5-n_{i} $.

    Lemma 1.2. [32] Let $ F $ be an $ n $-order and $ h $-hexagon f-benzenoid with $ r $ inlets, Then

    $ {\rm (i)} $ $ m_{22} = n-2h-r $;

    $ {\rm (ii)} $ $ m_{23} = 2r $;

    $ {\rm (iii)} $ $ m_{33} = 3h-r $.

    From the perspective of mathematics and chemistry, finding the extremal values of some useful $ TI $ for significant classes of graphs is very interesting [14,19,23,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56].

    As a matter of convenience, we use $ \Gamma_{m} $ to represent the collection of f-benzenoids containing exactly $ m $ edges. In [45], we derived extremal values for $ TI $ among all f-benzenoids with given order. It is noted that structure of f-benzenoids with given order is different from that of f-benzenoids with given number of edges. And we found that the technique for studying $ TI $ among all f-benzenoids with given order can not be used directly to investigate $ TI $ for all f-benzenoids with fixed number of edges. For this reason, we concentrate on the research of extremal values for $ TI $ among all f-benzenoids with given size.

    The main idea of this work is to construct f-benzenoids owning maximal $ r $ and minimal $ h $ at the same time in $ \Gamma_{m} $ depending on the number $ m $ is congruent to $ 0, 1, 2, 3 $ or $ 4 $ modulo $ 5 $. By making use of this technique, we obtain the extremum of $ TI $ over $ \Gamma_{m} $ and characterize their corresponding graphs on the basis of $ m $ is congruent to $ 0, 1, 2, 3 $ or $ 4 $ modulo $ 5 $. Afterwards the extremums of some well-known $ TI $ over $ \Gamma_{m} $ can be got by use of the previous results.

    The structure of this paper is as below. We first determine the maximal $ r $ in the set $ \Gamma_{m} $ in Section 2. By utilizing these results, we find the extremum of several famed $ TI $ over $ \Gamma_{m} $ in Section 3.

    We will find the f-benzenoids with maximal $ r $ in $ \Gamma_{m} $ in this section. Figure 4 illustrates three f-benzenoids pertaining to $ \Gamma_{42} $.

    Figure 4.  Some f-benzenoids in $ \Gamma_{42} $.

    At first, we try to obtain the maximum and minimum number of hexagons in any $ F\in\Gamma_{m} $.

    The spiral benzenoid system [57] $ T_{h} $ is a benzenoid system whose structure is in a "spiral" manner as illustrated in Figure 5. $ T_{h} $ has maximal $ n_{i} $ in all $ h $-hexagon benzenoid systems.

    Figure 5.  The spiral benzenoid system $ T_{h} $ with maximal number of internal vertices.

    As a matter of convenience, let $ SH_{h} $ $ (h\geq3) $ represent the collection of f-benzenoids formed by two spiral benzenoids $ X $ and $ Y $. Particularly, a f-spiral benzenoid is a f-benzenoid $ F^{\ast}\in SH_{h} $ in which $ X = T_{h-1} $ and $ Y = T_{1} $ (as shown in Figure 6). It is easy to see that that

    $ n_{i}(F^{\ast}) = 2h-\lceil\sqrt{12(h-1)-3}\rceil. $
    Figure 6.  f-benzenoid $ F'\in SH_{h} $ whose two pieces $ X $ and $ Y $ are both spiral benzenoid systems, and f-spiral benzenoid $ F^{\ast}\in SH_{h} $ with two pieces $ X = T_{h-1} $ and $ Y = T_{1} $.

    In [40], we proved that for every $ F'\in SH_{h} $ $ (h\geq3) $, the inequality

    $ ni(F)ni(F)
    $
    (2.1)

    holds, and the following graph operations were introduced.

    Operation 1. For any $ h $-hexagon f-benzenoid $ F $ having two segments $ X $ and $ Y $, let $ h_{1} = h(X) $ and $ h_{2} = h(Y) $. By substituting spiral benzenoid systems $ T_{h_{1}} $ and $ T_{h_{2}} $ for $ X $ and $ Y $, severally, another f-benzenoid $ F'\in SH_{h} $ can be obtained (as shown in Figure 7).

    Figure 7.  f-benzenoid $ F'\in SH_{h} $ is obtained from $ F $ by applying Operation 1 to it.

    For any $ h $-hexagon f-benzenoid $ F $, when $ h = 3 $, it is easily checked that

    $ ni(F)=1=2×312(31)3.
    $
    (2.2)

    When $ h\geq4 $, let $ h_{1} = h(X) $ and $ h_{2} = h(Y) $. Another $ F'\in SH_{h} $ (as shown in Figure 7) in which $ X = T_{h_{1}} $ and $ Y = T_{h_{2}} $ can be acquired by applying Operation 1 to $ F $. It is apparently that $ n_{i}(X)\leq n_{i}(T_{h_{1}}) $, $ n_{i}(Y)\leq n_{i}(T_{h_{2}}) $, therefore

    $ ni(F)=ni(X)+ni(Y)+1ni(Th1)+ni(Th2)+1=ni(F).
    $
    (2.3)

    So, the following Lemma can be deduced by Eqs (2.1) and (2.3).

    Lemma 2.1. [41] Let $ F $ be an $ h(h\geq3) $-hexagon f-benzenoid. Then

    $ ni(F)2h12(h1)3,
    $
    (2.4)

    and the equality is established when $ F $ is $ F^{\ast} $.

    For any $ F\in \Gamma_{m} $, $ h(F) $ over $ \Gamma_{m} $ is variable. Sharp bounds for $ h(F) $ in $ \Gamma_{m} $ is given below.

    Theorem 2.1. For any f-benzenoid $ F\in \Gamma_{m} $,

    $ 15(m4)h(F)m113(2m+4m31),
    $
    (2.5)

    where $ \lceil x\rceil $ is the smallest integer larger or equal to $ x $.

    Proof. On one hand, from Lemma 1.1 $ {\rm(ii)} $ we know that $ m = 5h(F)+5-n_{i}(F) $. Combining the fact that $ n_{i}(F)\geq1 $ for any $ F\in \Gamma_{m} $, we get

    $ h(F)\geq \left\lceil\frac{1}{5}(m-4)\right\rceil. $

    On the other hand, by Lemma 2.1 we know that $ n_{i}(F)\leq n_{i}(F^{\ast}) $. Consequently, from $ m = 5h(F)+5-n_{i}(F) $ we have

    $ m-3h(F)-5\geq\left\lceil\sqrt{12(h(F)-1)-3}\right\rceil\geq\sqrt{12(h(F)-1)-3}. $

    Hence,

    $ (3h(F)+(3-m))^{2}\geq4m-31. $

    Due to the fact that $ 3h(F)+(3-m) < 0 $, we deduce

    $ 3h(F)+(3-m)\leq-\sqrt{4m-31}, $

    i.e., $ h(F)\leq m-1-\left\lceil\frac{1}{3} \left(2m+\sqrt{4m-31}\right)\right\rceil. $

    Remark 1. Theorem 2.1 implies that f-spiral benzenoid $ F^{\ast} $ has the maximal number of hexagons over $ \Gamma_{m} $.

    For the sake of obtaining the extremum $ TI $ among all f-benzenoids in $ \Gamma_{m} $, we need to find the f-benzenoids $ F\in\Gamma_{m} $ possessing maximal $ r $.

    Recall that convex benzenoid systems (CBS for brevity) are a particular sort of benzenoid systems lack of bay regions [14]. Let $ \mathcal{HS}_{h} $ be the collection of benzenoid systems containing $ h $ hexagons.

    Lemma 2.2. [42] Let $ H\in \mathcal{HS}_{h} $. Under the below cases, $ H $ is definitely not a CBS:

    $ \rm (i) $ If $ h\geq4 $ and $ n_{i} = 1 $;

    $ \rm (ii) $ If $ h\geq5 $ and $ n_{i} = 2 $;

    $ \rm (iii) $ If $ h\geq6 $ and $ n_{i} = 3 $.

    Lemma 2.3. [52] Let $ H\in \mathcal{HS}_{h} $ such that $ n_{i}(H) = 4 $. Then $ H $ is bound to embody a subbenzenoid system given in Figure 8, there does not exist hexagons which are adjacent to fissures.

    Figure 8.  Benzenoid systems with 1, 2, 3 and 4 internal vertices, respectively.

    Lemma 2.4. Let $ S\in \mathcal{HS}_{h} $. If $ h\geq7 $ and $ n_{i}(S) = 4 $, then $ S $ is not a CBS.

    Proof. Let $ S $ be an $ h $ $ (h\geq7) $-hexagon benzenoid system, $ n_{i}(S) = 4 $, then by Lemma 2.3 $ S $ must contain one of the benzenoid systems of the form given in Figure 7. The proof is carried out in two cases.

    Case 1. If these four internal vertices form a path $ P_{4} $ or a $ K_{1,3} $, then $ S $ contains one of benzenoid systems $ (d) $–$ (f) $ in Figure 7 as its subbenzenoid systems. It is noted that $ h\geq7 $, by Lemma 2.2, it must not exist hexagons contiguous to the fissures, so, $ S $ has at least one hexagon contiguous to a $ (2, 2) $-edge, by means of such hexagons, it is succeeded in converting one of the fissures into a cove, bay or fjord. Hence, $ b(S)\geq1 $.

    Case 2. If these four internal vertices are not adjacent then $ S $ has possibility subbenzenoid systems as follows.

    1) There exist one type $ (a) $ and one type $ (c) $ benzenoid systems in $ S $;

    2) There exist two type $ (b) $ benzenoid systems in $ S $;

    3) There exist two type $ (a) $ and one type $ (b) $ benzenoid systems in $ S $.

    4) There exist four type $ (a) $ benzenoid systems in $ S $

    By Lemma 2.2, neither hexagons may be adjacent to the fissures in any of the cases indicated above. Since $ h\geq7 $, $ S $ has at least one hexagon contiguous to a $ (2, 2) $-edge, by means of such hexagons, it is succeeded in making one of the fissures become a cove, bay or fjord. Therefore, $ b(S)\geq1 $.

    The proof is completed.

    Lemma 2.5. [45] Let $ F $ be an $ h $-hexagon f-benzenoid. Then

    $ \rm 1) $ If $ n_{i} = 1 $, then $ r(F)\leq r(FL_{h}) = 2h-3\ (h\geq3) $;

    $ \rm 2) $ If $ n_{i} = 2 $, then $ r(F)\leq r(G_{h}) = 2h-4\ (h\geq4) $;

    $ \rm 3) $ If $ n_{i} = 3 $, then $ r(F)\leq r(R_{h}) = 2h-5\ (h\geq5) $;

    $ \rm 4) $ If $ n_{i} = 4 $, then $ r(F)\leq r(Z_{h}) = 2h-6\ (h\geq6) $.

    Next we find the f-benzenoids with maximal $ r $ in $ \Gamma_{m} $ with a fixed $ n_{i} $. Recall that $ M_{h} $, $ N_{h} $ and $ Q_{h} $ (see Figure 9) are benzenoid systems, and $ G_{h} $ (see Figure 10), $ R_{h} $ (see Figure 11), $ Z_{h} $ (see Figure 12) are f-benzenoids.

    Figure 9.  Three types of benzenoid systems.
    Figure 10.  f-benzenoids $ G_{4} $, and $ G_{h} $ $ (h\geq5) $.
    Figure 11.  f-benzenoids $ R_{5} $, and $ R_{h} $ $ (h\geq6) $.
    Figure 12.  f-benzenoids $ Z_{6} $, and $ Z_{h} $ $ (h\geq7) $.
    Figure 13.  f-benzenoids $ U_{7} $, and $ U_{h} $ $ (h\geq8) $.

    Lemma 2.6. [41] Let $ F $ be an $ h $-hexagon f-benzenoid. Then

    $ r(F)\leq r(FL_{h}) = 2h-3. $

    Lemma 2.7. [32] For any $ h $-hexagon f-benzenoid including $ n_{i} $ internal vertices and $ b $ bay regions, the number of $ (2, 2) $-edge and $ (2, 3) $-edge are $ m_{22} = b+5, m_{23} = 4h-2n_{i}-2b $, respectively.

    From Lemmas 1.2 $ {\rm(ii)} $ and 8, we get

    $ r=2hnib
    $
    (2.6)

    Furthermore, by Lemma 1.1 $ {\rm(ii)} $ and Eq (2.6), we deduce

    $ r=m3h5b
    $
    (2.7)

    Theorem 2.2. Let $ F $ be an $ h $-hexagon f-benzenoid. If $ n_{i} = 5 $, then $ r(F)\leq r(U_{h}) = 2h-7 $ $ (h\geq7) $.

    Proof. Let $ h _{1} = h(X) $ and $ h_{2} = h(Y) $, $ X $ and $ Y $ are two segments of $ F $. If $ n_{i} = 5 $, by the structure of f-benzenoid, equality $ n_{i}(X)+n_{i}(Y) = 4 $ holds, so, we have the following five cases.

    Case 1. $ n_{i}(X) = 1 $, $ n_{i}(Y) = 3 $, i.e., there exist one internal vertex and three internal vertices in $ X $ and $ Y $, respectively.

    Subcase 1.1. If $ h_{1} = 3 $, then $ X = M_{3} $.

    Subcase 1.1.1. If $ h_{2} = 5 $, i.e., $ Y = Q_{5} $, then $ F $ is the f-benzenoid $ D_{1} $, $ D_{2} $ or $ D_{3} $ (see Figure 14). It is clear that $ r(F) = r(D_{1}) = 8\leq 2h-7 $, $ r(F) = r(D_{2}) = 7\leq 2h-7 $ or $ r(F) = r(D_{3}) = 8\leq 2h-7 $.

    Figure 14.  f-benzenoids $ D_{1} $, $ D_{2} $, $ D_{3} $, $ D_{4} $ and $ D_{5} $.

    Subcase 1.1.2. If $ h_{2}\geq6 $, by Lemma 2.2 and the hypothesis that $ n_{i}(Y) = 3 $, $ Y $ is not a CBS, so $ b(Y)\geq1 $. Furthermore, $ b(F)\geq3 $, combining Eq (2.6) we obtain $ r = 2h-n_{i}-b\leq2h-8 < 2h-7 $.

    Subcase 1.2. If $ h_{1}\geq4 $, according to Lemma 2.2, $ X $ is definitely not a CBS, i.e., $ b(X)\geq1 $.

    Subcase 1.2.1. If $ h_{2} = 5 $, i.e., $ Y = Q_{5} $. It is clear that $ b(F)\geq4 $, then Eq (2.6) deduces $ r\leq2h-9 < 2h-7 $.

    Subcase 1.2.2. If $ h_{2}\geq6 $, $ Y $ is definitely not not a CBS according to Lemma 2.2, so, $ b(Y)\geq1 $. It is clear that $ b(F)\geq5 $, consequently from Eq (2.6) we obtain $ r\leq2h-10 < 2h-7 $.

    Case 2. $ n_{i}(X) = 3 $ and $ n_{i}(Y) = 1 $.

    Subcase 2.1. If $ h_{1} = 5 $, then $ X = Q_{5} $.

    Subcase 2.1.1. If $ h_{2} = 3 $, i.e., $ Y = M_{3} $, then $ F $ is the f-benzenoid $ D_{4} $, $ D_{5} $, $ D_{6} $ (see Figure 14), or $ D_{7} $ (as shown in Figure 15). $ r(F) = r(D_{4}) = 8\leq 2h-7 $, $ r(F) = r(D_{5}) = 7\leq 2h-7 $, $ r(F) = r(D_{6}) = 8\leq 2h-7 $, $ r(F) = r(D_{7}) = 7\leq 2h-7 $.

    Figure 15.  f-benzenoids $ D_{7} $, $ D_{8} $ and $ D_{9} $.

    Subcase 2.1.2. If $ h_{2}\geq4 $, $ Y $ is surely not a CBS in light of Lemma 2.2, i.e., $ b(X)\geq1 $. Hence, we have $ b(F)\geq4 $, it follows from Eq (2.6) that $ r\leq2h-9 < 2h-7 $.

    Subcase 2.2. If $ h_{1}\geq6 $, by Lemma 2.2, $ X $ is definitely not a CBS, hence $ b(X)\geq1 $.

    Subcase 2.2.1. If $ h_{2} = 3 $, i.e., $ Y = M_{3} $. We have $ b(F)\geq4 $, and Eq (2.6) infers that $ r\leq2h-9 < 2h-7 $.

    Subcase 2.2.2. f $ h_{2}\geq4 $, by Lemma 2.2, $ Y $ is certainly not a CBS, i.e., $ b(X)\geq1 $. Hence we have $ b(F)\geq5 $, by Eq (2.6), $ r\leq2h-10 < 2h-7 $.

    Case 3. $ n_{i}(X) = 2 $, $ n_{i}(Y) = 2 $, i.e., $ X $ and $ Y $ both have two internal vertices.

    Subcase 3.1. If $ h_{1} = 4 $, we note that $ n_{i}(X) = 2 $, so $ X $ must be the benzenoid system $ (b) $ in Figure 9.

    Subcase 3.1.1. If $ h_{2} = 4 $, $ Y $ is surely the benzenoid system $ (b) $ in Figure 9 according to the hypothesis $ n_{i}(Y) = 2 $, therefore, $ F $ is $ D_{8} $ or $ D_{9} $ (as shown in Figure 15). We get $ r(F) = r(D_{8}) = 8 < 2h-7 $ or $ r(F) = r(D_{9}) = 7 < 2h-7 $.

    Subcase 3.1.2. If $ h_{2}\geq5 $, by Lemma 2.2 and that $ n_{i}(Y) = 2 $, $ Y $ is not a CBS, so we know that $ b(X)\geq1 $. Then $ b(F)\geq4 $, by Eq (2.6) and the fact that $ n_{i} = 5 $, $ r\leq2h-9 < 2h-7 $.

    Subcase 3.2. If $ h_{2} = 4 $, we note that $ n_{i}(Y) = 2 $, so $ Y $ must be the benzenoid system $ (b) $ in Figure 8.

    Subcase 3.2.1. If $ h_{1} = 4 $, $ X $ must also be the benzenoid system $ (b) $ in Figure 9. Hence, $ F $ is $ D_{8} $ or $ D_{9} $ (as shown in Figure 15). $ r(F) = r(D_{8}) = 8\leq 2h-7 $ or $ r(F) = r(D_{9}) = 7\leq 2h-7 $.

    Subcase 3.2.2. If $ h_{1}\geq5 $, by Lemma 2.2 and $ n_{i}(X) = 2 $, $ X $ is definitely not a CBS, i.e., $ b(X)\geq1 $. Hence, $ b(F)\geq4 $, by Eq (2.6) and the fact that $ n_{i} = 5 $, we have $ r\leq2h-9 < 2h-7 $.

    Subcase 3.3. If $ h_{1}\geq5 $, $ h_{2}\geq5 $, it is noted that $ n_{i}(X) = n_{i}(Y) = 2 $, neither $ X $ nor $ Y $ are definitely CBS according to Lemma 2.2. So, both $ b(X) $ and $ b(Y) $ are greater than $ 1 $. Hence, $ b(F)\geq5 $, on the basis of Eq (2.6) we get $ r\leq2h-10 < 2h-7 $.

    Case 4. $ n_{i}(X) = 4 $ and $ n_{i}(Y) = 0 $, i.e., $ X $ contains four internal vertices, $ Y $ is a catacondensed benzenoid system.

    Subcase 4.1. If $ h_{1} = 6 $, then $ X $ is the benzenoid system $ (d) $, $ (e) $ or $ (f) $ in Figure 9.

    Subcase 4.1.1. If $ h_{2} = 1 $, $ F $ is the f-benzenoid $ D_{10} $, $ D_{11} $, $ D_{12} $ (see Figure 16), $ D_{13} $ (see Figure 17) or $ U_{7} $ (see Figure 12). $ r(F) = r(D_{10}) = 6\leq 2h-7 $, $ r(F) = r(D_{11}) = 6\leq 2h-7 $, $ r(F) = r(D_{12}) = 6\leq 2h-7 $, $ r(F) = r(D_{13}) = 6\leq 2h-7 $ or $ r(F) = r(U_{7}) = 7 = 2h-7 $.

    Figure 16.  f-benzenoids $ D_{10} $, $ D_{11} $ and $ D_{12} $.
    Figure 17.  f-benzenoids $ D_{13} $, $ D_{14} $, $ D_{15} $, $ D_{16} $, $ D_{17} $, $ D_{18} $, $ D_{19} $, $ D_{20} $ and $ D_{21} $.

    Subcase 4.1.2. If $ h_{2}\geq2 $, we have $ b(F)\geq2 $, by Eq (2.6), $ r\leq2h-7 $.

    Subcase 4.2. If $ h_{1}\geq7 $, in the light of Lemma 2.4, $ X $ is definitely not a CBS, hence $ b(Y)\geq1 $. In this situation $ b(F)\geq3 $, we get the inequality $ r\leq2h-8 < 2h-7 $ according to Eq (2.6).

    Case 5. $ n_{i}(X) = 0 $ and $ n_{i}(Y) = 4 $, i.e., $ X $ is a catacondensed benzenoid system, $ Y $ has four internal vertices.

    Subcase 5.1. If $ h_{2} = 6 $, then $ Y $ is the benzenoid system $ (d) $, $ (e) $ or $ (f) $ in Figure 8.

    Subcase 5.1.1. If $ h_{1} = 2 $, $ X $ must be the linear chain $ L_{2} $. In this event, $ F $ is $ D_{14} $, $ D_{15} $, $ D_{16} $, $ D_{17} $, $ D_{18} $, $ D_{19} $, $ D_{20} $ or $ D_{21} $ (see Figure 17). By further checking, we gain that $ r(F) = r(D_{14}) = 7\leq 2h-7 $, $ r(F) = r(D_{15}) = 8\leq 2h-7 $, $ r(F) = r(D_{16}) = 8\leq 2h-7 $, $ r(F) = r(D_{17}) = 7\leq 2h-7 $, $ r(F) = r(D_{18}) = 7\leq 2h-7 $, $ r(F) = r(D_{19}) = 8\leq 2h-7 $, $ r(F) = r(D_{20}) = 6\leq 2h-7 $ or $ r(F) = r(D_{21}) = 6\leq 2h-7 $.

    Subcase 5.1.2. If $ h_{1}\geq3 $, bearing in mind that $ X $ is a catacondensed benzenoid system and $ Y $ is the benzenoid system $ (d) $, $ (e) $ or $ (f) $ in Figure 8, then $ F $ must have f-benzenoid $ D_{14} $, $ D_{15} $, $ D_{16} $, $ D_{17} $, $ D_{18} $, $ D_{19} $, $ D_{20} $ or $ D_{21} $ (see Figure 17) as its subgraph.

    Subcase 5.1.2.1. If $ D_{14} $ is a subgraph in $ F $, it is obvious that $ D_{14} $ has two coves. Since $ X $ is a catacondensed benzenoid system and $ h_{1}\geq3 $, $ F $ has at least one hexagon contiguous to a $ (2, 2) $-edge of $ X $, and such hexagons can convert one fissure into a bay, or convert one cove into a fjord, or convert one fjord into a lagoon. In this instance $ b(F)\geq4 $. Consequently, $ r\leq2h-9 < 2h-7 $ can be got according to Eq (2.6).

    Subcase 5.1.2.2. If $ D_{15} $, $ D_{16} $ or $ D_{19} $ is a subpart f-benzenoid in $ F $, it is obvious each one of $ D_{15} $, $ D_{16} $ and $ D_{19} $ has a bay and a cove. Since $ X $ is a catacondensed benzenoid system and $ h_{1}\geq3 $, $ F $ contains at least one hexagon adjoining a $ (2, 2) $-edge of $ X $, and such hexagons will make one fissure become a bay, or make one cove become a fjord, or make one fjord become a lagoon. Consequently, $ b(F)\geq4 $, by Eq (2.6) it follows that $ r\leq2h-9 < 2h-7 $.

    Subcase 5.1.2.3. If $ D_{17} $ is a subpart f-benzenoid in $ F $, it is obvious that $ D_{17} $ has a fjord and a bay. Since $ X $ is a catacondensed benzenoid system and $ h_{1}\geq3 $, $ F $ has at least one hexagon adjoining a $ (2, 2) $-edge of $ X $, and such hexagons will convert one fissure into a bay, or convert one cove into a fjord, or convert one fjord into a lagoon. Consequently, $ b(F)\geq4 $, by Eq (2.6) it follows that $ r\leq2h-9 < 2h-7 $.

    Subcase 5.1.2.4. If $ D_{18} $ is a subpart f-benzenoid in $ F $, it is obvious that $ D_{18} $ has a fjord and two bays. Since $ X $ is a catacondensed benzenoid system and $ h_{1}\geq3 $, there exists has at least one hexagon adjoining a $ (2, 2) $-edge of $ X $ in $ F $, and these hexagons will convert one of the fissures into a bay, or convert one cove into a fjord, or convert one fjord into a lagoon. Consequently, $ b(F)\geq4 $, in light of Eq (2.6), $ r\leq2h-9 < 2h-7 $.

    Subcase 5.1.2.5. If $ D_{20} $ or $ D_{21} $ is a subpart f-benzenoid in $ F $, it is obvious that both $ D_{20} $ and $ D_{21} $ have a bay and two fjords. Since $ X $ is a catacondensed benzenoid system and $ h_{1}\geq3 $, $ F $ contains at least one hexagon adjoining a $ (2, 2) $-edge of $ X $, and such hexagons will make one fissure become a bay, or make one cove become a fjord, or make one fjord become a lagoon. Consequently, $ b(F)\geq4 $, according to Eq (2.6), $ r\leq2h-9 < 2h-7 $.

    Subcase 5.2. If $ h_{2}\geq7 $, by Lemma 2.4 and the fact that $ n_{i}(Y) = 4 $, $ Y $ is certainly not a CBS, i.e., $ b(Y)\geq1 $.

    Subcase 5.2.1. If $ h_{1} = 2 $, i.e., $ X = L_{2} $. From the structure of f-benzenoid, $ F $ is formed from $ X $ and $ Y $ joined by a pentagon, it is easily seen that there are at least one bay or one cove arisen in the process of construction of $ F $. It is clear that $ b(F)\geq2 $, by Eq (2.6) we have $ r\leq2h-7 $.

    Subcase 5.2.2. If $ h_{1}\geq3 $, we know that $ F $ is formed by joining from $ X $ and $ Y $ through a pentagon, in this construction process of $ F $, it is easily seen that there are at least one bay or one cove arisen. Then $ b(F)\geq2 $, by Eq (2.6), $ r\leq2h-7 $.

    The proof is completed.

    We recall that $ FL_{h} $ is the f-linear chain with $ h $ hexagons [40]. Extremal f-benzenoids with maximal $ r $ in $ \Gamma_{m} $ were determined in the following theorem.

    Theorem 2.3. Let $ F\in\Gamma_{m} $. Then

    $ \rm 1) $ If $ m\equiv0(mod5) $, then $ r(F)\leq \frac{2m-35}{5} = r(U_{\frac{m}{5}}) $;

    $ \rm 2) $ If $ m\equiv1(mod5) $, then $ r(F)\leq \frac{2m-32}{5} = r(Z_{\frac{m-1}{5}}) $;

    $ \rm 3) $ If $ m\equiv2(mod5) $, then $ r(F)\leq \frac{2m-29}{5} = r(R_{\frac{m-2}{5}}) $;

    $ \rm 4) $ If $ m\equiv3(mod5) $, then $ r(F)\leq \frac{2m-26}{5} = r(G_{\frac{m-3}{5}}) $;

    $ \rm 5) $ If $ m\equiv4(mod5) $, then $ r(F)\leq \frac{2m-23}{5} = r(FL_{\frac{m-4}{5}}) $.

    Proof. We know by Eq (2.5) that

    $ \left\lceil\frac{1}{5}(m-4)\right\rceil \leq h(F)\leq m-1-\left\lceil\frac{1}{3}\left(2m+\sqrt{4m-31}\right)\right\rceil. $

    1) If $ m\equiv0(mod5) $, then $ \left\lceil\frac{1}{5}(m-4)\right\rceil = \frac{m}{5} $. If $ h = \frac{m}{5} $, then by Lemma 1.1 $ {\rm(ii)} $

    $ m = 5h(F)+5-n_{i}(F) = m+5-n_{i}(F), $

    it means that $ n_{i}(F) = 5 $. Furthermore, Theorem 2.2 infers that $ r(F)\leq r(U_{\frac{m}{5}}) $ and we are done. So assume now that $ h(F)\geq\frac{m}{5}+1 $, then by equality (2.7) and the fact that $ b(F)\geq2 $

    $ r(F) = m-5-3h(F)-b(F)\leq m-5-3(\frac{m}{5}+1)-b(F) $
    $ \leq \frac{2m}{5}-10 = \frac{2m-50}{5}\leq \frac{2m-35}{5} = r(U_{\frac{m}{5}}). $

    2) If $ m\equiv1(mod5) $, then $ \left\lceil\frac{1}{5}(m-4)\right\rceil = \frac{m-1}{5} $. If $ h(F) = \frac{m-1}{5} $, then by Lemma 1.1 $ {\rm(ii)} $

    $ m = 5h(F)+5-n_{i}(F) = m+4-n_{i}(F), $

    thus $ n_{i}(F) = 4 $. Then $ r(F)\leq r(Z_{\frac{m-1}{5}}) $ by part 4 of Lemma 2.5. Otherwise $ h(F)\geq\frac{m-1}{5}+1 $, then by equality (2.7) and the obvious fact that $ b(F)\geq2 $

    $ r(F) = m-5-3h(F)-b(F)\leq m-5-3(\frac{m-1}{5}+1)-b(F) $
    $ \leq \frac{2m+3}{5}-10 = \frac{2m-47}{5}\leq \frac{2m-32}{5} = r(Z_{\frac{m-1}{5}}). $

    3) If $ m\equiv2(mod5) $, then $ \left\lceil\frac{1}{5}(m-4)\right\rceil = \frac{m-2}{5} $. If $ h(F) = \frac{m-2}{5} $, then by Lemma 1.1 $ {\rm(ii)} $

    $ m = 5h(F)+5-n_{i}(F) = m+3-n_{i}(F), $

    and so $ n_{i}(F) = 3 $. Then $ r(F)\leq r(R_{\frac{m-2}{5}}) $ by part 3 of Lemma 2.5. So assume now that $ h(F)\geq\frac{m-2}{5}+1 $, then by Eq (2.7) and the fact that $ b(F)\geq2 $

    $ r(F) = m-5-3h(F)-b(F)\leq m-5-3(\frac{m-2}{5}+1)-b(F) $
    $ \leq \frac{2m+6}{5}-10 = \frac{2m-44}{5}\leq \frac{2m-29}{5} = r(R_{\frac{m-2}{5}}). $

    4) If $ m\equiv3(mod5) $, then $ \left\lceil\frac{1}{5}(m-4)\right\rceil = \frac{m-3}{5} $. If $ h(F) = \frac{m-3}{5} $, then by Lemma 1.1 $ {\rm(ii)} $

    $ m = 5h(F)+5-n_{i}(F) = m+2-n_{i}(F), $

    thus $ n_{i}(F) = 2 $. By Lemma 2.5, $ r(F)\leq r(G_{\frac{m-3}{5}}) $ and we are done. If $ h(F)\geq\frac{m-3}{5}+1 $, then by equality (2.7) and the fact that $ b(F)\geq2 $

    $ r(F) = m-5-3h(F)-b(F)\leq m-5-3(\frac{m-3}{5}+1)-b(F) $
    $ \leq \frac{2m+9}{5}-10 = \frac{2m-41}{5}\leq \frac{2m-26}{5} = r(G_{\frac{m-3}{5}}). $

    5) If $ m\equiv4(mod5) $, then $ \left\lceil\frac{1}{5}(m-4)\right\rceil = \frac{m-4}{5} $. Since $ h\geq\frac{m-4}{5} $ and $ b(F)\geq2 $, then by Eq (2.7), we have

    $ r(F) = m-5-3h(F)-b(F)\leq m-5-\frac{3m-12}{5}-b(F) $
    $ \leq \frac{2m+12}{5}-7 = \frac{2m-23}{5} = r(FL_{\frac{m-4}{5}}). $

    In this part, we attempt to find the extremal values of $ TI $ over $ \Gamma_{m} $.

    It is noted that a f-benzenoid $ F $ contains only $ 2 $-vertex and $ 3 $-vertex. Hence, equation (1.1) reduces to

    $ TI(F)=m22ψ22+m23ψ23+m33ψ33,
    $
    (3.1)

    In the light of Lemmas 1.1 and 1.2,

    $ TI(F)=ψ22m+3(ψ33ψ22)h+(2ψ23ψ22ψ33)r,
    $
    (3.2)

    If $ U, V\in\Gamma_{m} $ then clearly

    $ TI(U)TI(V)=3(ψ33ψ22)(h(U)h(V))          +(2ψ23ψ22ψ33)(r(U)r(V)).
    $
    (3.3)

    For convenience, we set $ s = \psi_{33}-\psi_{22} $, $ q = 2\psi_{23}-\psi_{22}-\psi_{33} $.

    Theorem 3.1. For any $ F\in\Gamma_{m} $, we have the following results.

    $ {\rm a.} $ If $ s\leq0 $ and $ q\geq0 $,

    $ TI(F)\leq\left\{TI(Um5),if m0(mod 5)TI(Zm15),if m1(mod 5)TI(Rm25),if m2(mod 5)TI(Gm35),if m3(mod 5)TI(FLm45),if m4(mod 5)
    \right. $

    $ {\rm b.} $ If $ s\geq0 $ and $ q\leq0 $,

    $ TI(F)\geq\left\{TI(Um5),if m0(mod 5)TI(Zm15),if m1(mod 5)TI(Rm25),if m2(mod 5)TI(Gm35),if m3(mod 5)TI(FLm45),if m4(mod 5)
    \right. $

    Proof. Let $ F\in\Gamma_{m} $. By Eq (2.5)

    $ h(F)\geq \left\lceil\frac{1}{5}(m-4)\right\rceil = \left\{h(Um5),if m0(mod 5)h(Zm15),if m1(mod 5)h(Rm25),if m2(mod 5)h(Gm35),if m3(mod 5)h(FLm45),if m4(mod 5)
    \right. $

    i.e., f-benzenoids $ U_{\frac{m}{5}} $, $ Z_{\frac{m-1}{5}} $, $ R_{\frac{m-2}{5}} $, $ G_{\frac{m-3}{5}} $ and $ FL_{\frac{m-4}{5}} $ have minimal $ h $ over the set $ \Gamma_{m} $. Meanwhile, by Theorem 2.3, we have

    $ r(F)\leq \left\{r(Um5),if m0(mod 5)r(Zm15),if m1(mod 5)r(Rm25),if m2(mod 5)r(Gm35),if m3(mod 5)r(FLm45),if m4(mod 5)
    \right. $

    i.e., these five f-benzenoids have maximal number of inlets over $ \Gamma_{m} $. Hence, for any f-benzenoids $ F\in\Gamma_{m} $ and $ V\in\{U_{\frac{m}{5}}, Z_{\frac{m-1}{5}}, R_{\frac{m-2}{5}}, G_{\frac{m-3}{5}}, FL_{\frac{m-4}{5}}\} $, $ h(F)-h(V)\geq0 $ and $ r(F)-r(V)\leq0 $ hold simultaneously, from Eq (2.7), we have

    $ TI(F)-TI(V) = 3s(h(F)-h(V))+q(r(F)-r(V)). $

    If $ s\leq0 $ and $ q\geq0 $, then $ TI(F)-TI(V)\leq0 $, i.e., $ V $ reaches the maximum value of $ TI $ over $ \Gamma_{m} $. If $ s\geq0 $ and $ q\leq0 $, then $ TI(F)-TI(V)\geq0 $, i.e., $ V $ reaches the minimum value of $ TI $ over $ \Gamma_{m} $. Furthermore, which $ V\in\{U_{\frac{m}{5}}, Z_{\frac{m-1}{5}}, R_{\frac{m-2}{5}}, G_{\frac{m-3}{5}}, FL_{\frac{m-4}{5}}\} $ is the extremal graph depending on $ m $ is congruent to $ 0, 1, 2, 3 $ or $ 4 $ modulo $ 5 $.

    Example 1. Values of $ s $ and $ q $ for several famous $ TI $ are listed in Table 2:

    Table 2.  Values of $ s $ and $ q $ for six famous $ TI $.
    $ ij $ $ \frac{1}{\sqrt{ij}} $ $ \frac{2\sqrt{ij}}{i+j} $ $ \frac{1}{\sqrt{i+j}} $ $ \frac{(ij)^{3}}{(i+j-2)^{3}} $ $ \sqrt{\frac{i+j-2}{ij}} $
    $ q $ -1 -0.0168 -0.0404 -0.0138 -3.390 0.040
    $ s $ 5 -0.1667 0 -0.091 3.390 -0.040

     | Show Table
    DownLoad: CSV

    Therefore, the minimum extreme value of $ TI $ for the second Zagreb index, GA index and the AZI index can be determined in the light of Theorems 2.3 and 3.1, and we can obtain the maximum extreme value of $ TI $ for the ABC index.

    If f-benzenoid $ F\in\Gamma_{m} $, then from the Eqs (2.3) and (2.6) and Lemma 1.1$ ({\rm ii}) $ we have

    $ TI(F)=(2ψ23ψ33)m+6(ψ33ψ23)h(2ψ23ψ22ψ33)b               5(2ψ23ψ22ψ33).
    $
    (3.4)

    Consequently, for f-benzenoids $ U, V\in\Gamma_{m} $

    $ TI(U)TI(V)=6(ψ33ψ23)(h(U)h(V))          +(2ψ23+ψ22+ψ33)(b(U)b(V)).
    $
    (3.5)

    Set $ u = 6(\psi_{33}-\psi_{23}) $ and keep in mind that $ q = 2\psi_{23}-\psi_{22}-\psi_{33} $. Then

    $ TI(U)TI(V)=u(h(U)h(V))q(b(U)b(V)).
    $
    (3.6)

    It is noted that Eq (3.6) can be decided only by $ h $, $ b $ and the signs of $ u $ and $ q $. For any $ F\in\Gamma_{m} $, We know that

    $ h(F)\leq m-1-\left\lceil\frac{1}{3}\left(2m+\sqrt{4m-31}\right)\right\rceil, $

    and the equality can be achieved precisely when $ F $ is the f-spiral benzenoid $ F^{\ast} $ [41].

    In [41], we proved that $ n_{i}(F^{\ast}) = 2h-\left\lceil\sqrt{12(h-1)-3}\right\rceil $. But, $ b(F^{\ast})\neq2 $ may occur. It is noticeable if $ X $ in $ F^{\ast} $ is a CBS, $ F^{\ast} $ is a f-benzenoid satisfying that $ b(F^{\ast}) = 2 $ or $ 3 $. For the sake of simplicity, Let $ \mathbb{N} $ be the set of positive integers.

    The CBS, $ W = H(l_{1}, l_{2}, l_{3}, l_{4}, l_{5}, l_{6}) $ (as shown in Figure 18), can be completely determined by the positive integers $ l_{1}, l_{2}, l_{3}, l_{4} $ [14].

    Figure 18.  The general form of a CBS. The parameters $ l_{i}\geq 1, i = 1, 2, \cdots, 6, $ count the number of hexagons on the respective side of CBS.

    The following lemma gave requirements that there exists CBS with maximal $ n_{i} $ [53].

    Lemma 3.1. [53] Let $ h\in\mathbb{N} $. The conditions below are isovalent:

    (a) There is a CBS $ W $ containing $ h $ hexagons and $ 2h+1-\lceil\sqrt{12h-3} \ \rceil $ number of internal vertices.

    (b) There exist $ l_{1}, l_{2}, l_{3}, l_{4}\in\mathbb{N} $ satisfying the following equation

    $ h=l1l3+l1l4+l2l3+l2l4l2l312l1(l1+1)12l4(l4+1)+112h3 =l1+2l2+2l3+l43}
    $
    (3.7)

    If for $ h\in\mathbb{N} $, Eq (3.7) has a solution $ l_{1}, l_{2}, l_{3}, l_{4}\in\mathbb{N} $, then there is a CBS $ W $ meeting the conditions that $ n_{i}(W) = n_{i}(T_{h}) $.

    Now, we concentrate on the research for $ TI $ of f-benzenoids. For a $ h-1\in\mathbb{N} $, supposing that the system below

    $ h1=l1l3+l1l4+l2l3+l2l4l2l312l1(l1+1)12l4(l4+1)+112(h1)3 =l1+2l2+2l3+l43 li{l1,l2,l3,l4,l5,l6}, li=2}
    $
    (3.8)

    has a solution $ \{l_{1}, l_{2}, l_{3}, l_{4}\} $, then a CBS $ W_{h-1} $ containing $ n_{i}(W_{h-1}) = 2(h-1)+1-\left\lceil\sqrt{12(h-1)-3}\right\rceil $ number of internal vertices exists. Note that $ l_{i} = 2 $ in system (3.8), i.e., there exists one fissure on the side of $ l_{i} $ of $ W_{h-1} $, let $ u, w, v $ in Figure 1 represent the three vertices of this fissure. Now, we obtain an f-spiral benzenoid $ F^{\ast}_{1} $ in which $ X = W_{h-1} $ and $ Y = L_{1} $. It is obvious that

    $ ni(F1)=2h12(h1)3
    $
    (3.9)

    and $ b(F^{\ast}_{1}) = 2 $. (as shown in Figure 19)

    Figure 19.  A f-spiral benzenoid $ F^{\ast}_{1} $ whose fragment $ X $ is a convex spiral benzenoid system $ W_{h-1} $.

    Theorem 3.2. Let $ h-1\in\mathbb{N} $ such that the Eq (3.8) has a solution, and $ m = 3h+5+\left\lceil\sqrt{12(h-1)-3}\right\rceil $. Then for any $ F\in\Gamma_{m} $

    1) $ TI(F^{\ast}_{1})\geq TI(F) $, when $ u\geq0 $ and $ q\geq0 $;

    2) $ TI(F^{\ast}_{1})\leq TI(F) $, when $ u\leq0 $ and $ q\leq0 $.

    Proof. From Lemma 1.1 $ ({\rm ii}) $ and Eq (3.9), we have

    $ m(F^{\ast}_{1}) = 5h+5-(2h-\left\lceil\sqrt{12(h-1)-3}\right\rceil) = 3h+5+\left\lceil\sqrt{12(h-1)-3}\right\rceil $

    and so

    $ h = m-1-\left\lceil\frac{1}{3} \left(2m+\sqrt{4m-31}\right)\right\rceil. $

    It is obvious that $ b(F^{\ast}_{1}) = 2 $ and $ b(F)\geq2 $ for any $ F\in\Gamma_{m} $. Hence by Eq (3.6), we have

    $ TI(F)-TI(F^{\ast}_{1}) = u(h(F)-h(F^{\ast}_{1}))-q(b(F)-b(F^{\ast}_{1})) $
    $ = u\left[h(F)-\left(m-1-\left\lceil\frac{1}{3} \left(2m+\sqrt{4m-31}\right)\right\rceil\right)\right]-q[b(F)-2]. $

    And by Eq (2.5)

    $ h(F)\leq m-1-\left\lceil\frac{1}{3}\left(2m+\sqrt{4m-31}\right)\right\rceil. $

    If $ u\geq0 $ and $ q\geq0 $ then $ TI(F)-TI(F^{\ast}_{1})\leq0 $, i.e., $ F^{\ast}_{1} $ achieves maximal $ TI $ in $ \Gamma_{m} $. Similarly, if $ u\leq0 $ and $ q\leq0 $ then $ TI(F)-TI(F^{\ast}_{1})\geq0 $, i.e., $ F^{\ast}_{1} $ obtains minimal $ TI $ in $ \Gamma_{m} $.

    Example 2. The values of $ u $ and $ q $ for some famous $ TI $ are listed in the following Table 3:

    Table 3.  Values of $ u $ and $ q $ for six famous $ TI $.
    $ ij $ $ \frac{1}{\sqrt{ij}} $ $ \frac{2\sqrt{ij}}{i+j} $ $ \frac{1}{\sqrt{i+j}} $ $ \frac{(ij)^{3}}{(i+j-2)^{3}} $ $ \sqrt{\frac{i+j-2}{ij}} $
    $ q $ -1 -0.0168 -0.0404 -0.0138 -3.390 0.040
    $ u $ 18 -0.449 0.121 -0.233 20.344 -0.242

     | Show Table
    DownLoad: CSV

    Hence, by Theorem 3.1 we can deduce the minimal values of the Randć index and the the sum–connectivity index in f-spiral benzenoid $ F^{\ast}_{1} $ for those $ h $ such that Eq (3.8) holds.

    Example 3. Take consideration of the generalized Randć index

    $ R_{\alpha}(G) = \sum\limits_{1\leq i\leq j\leq n-1}m_{ij}(ij)^{\alpha}, $

    where $ \alpha\in\mathbb{R} $. Note that

    $ q = 2(6^{\alpha})-4^{\alpha}-9^{\alpha} = -4^{\alpha}((\frac{3}{2})^{\alpha}-1)^{2}\leq0 $

    for all $ \alpha\in\mathbb{R} $. Moreover, $ s = 9^{\alpha}-4^{\alpha}\geq0 $ if and only if $ \alpha\geq0 $ if and only if $ u = 6(9^{\alpha}-6^{\alpha})\geq0 $. Hence, by Theorem 3.1, the minimal value of $ R_{\alpha}(G) $ is obtained for all $ \alpha\geq0 $, and for any $ \alpha\leq0 $, the minimal value of $ R_{\alpha}(G) $ can be attained by the f-spiral benzenoid $ F^{\ast}_{1} $ for those $ h $ such that Eq (3.8) holds.

    This work investigates extremum $ TI $ over the collection of f-benzenoids having same number of edges. In practical terms, there are many other types of very useful topological indices for instance graph energy [58,59,60,61,62], Wiener index [63], Randić energy [64], Wiener polarity index [65], incidence energy [66], Harary index [67], entropy measures [68,69] and HOMO-LUMO index [70]. So, determining these topological indices for f-benzenoids is going to be extraordinary fascinating.

    It is noted that the current framework is for studying topological indices of deterministic networks. But random networks would be a very promising direction. In [71,72], the distance Estrada index of random graphs was discussed, and the author went deeply into (Laplacian) Estrada index for random interdependent graphs. So, studying VDB topological indices of random and random interdependent graphs is another interesting problem.

    This work was supported by Ningbo Natural Science Foundation (No. 2021J234). The authors are very grateful to anonymous referees and editor for their constructive suggestions and insightful comments, which have considerably improved the presentation of this paper.

    The authors declare there is no conflict of interest.

    [1] G. Albi, Kinetic Approximation, Stability and Control of Collective Behavior in Self-Organized Systems, Ph.D. thesis, Università degli Studi di Ferrara, 2013.
    [2] G. Albi, M. Herty and L. Pareschi, Kinetic description of optimal control problems and applications to opinion consensus, Comm. Math. Sciences, to appear, (2015).
    [3] G. Albi and L. Pareschi, Binary interaction algorithms for the simulation of flocking and swarming dynamics, Multiscale Model. Simul., 11 (2013), 1-29. doi: 10.1137/120868748
    [4] ________, Modeling of self-organized systems interacting with a few individuals: From microscopic to macroscopic dynamics, Appl. Math. Lett., 26 (2013), 397-401.
    [5] G. Albi, L. Pareschi and M. Zanella, Boltzmann-type control of opinion consensus through leaders, Phil. Trans. R. Soc. A, 372 (2014), 20140138, 18pp. doi: 10.1098/rsta.2014.0138
    [6] D. Armbruster and C. Ringhofer, Thermalized kinetic and fluid models for re-entrant supply chains, SIAM J. Multiscale Modeling and Simulation, 3 (2005), 782-800. doi: 10.1137/030601636
    [7] N. Bellomo, G. Ajmone Marsan and A. Tosin, Complex Systems and Society. Modeling and Simulation, Springer Briefs in Mathematics, Springer, New York, 2013. doi: 10.1007/978-1-4614-7242-1
    [8] N. Bellomo and J. Soler, On the mathematical theory of the dynamics of swarms viewed as complex systems, Math. Models Methods Appl. Sci., 22 (2012), 1140006, 29pp. doi: 10.1142/S0218202511400069
    [9] A. Bensoussan, J. Frehse and P. Yam, Mean Field Games and Mean Field Type Control Theory, Springer Briefs in Mathematics, Springer, New York, 2013. doi: 10.1007/978-1-4614-8508-7
    [10] A. Borzí and S. Wongkaew, Modeling and control through leadership of a refined flocking system, Math. Models Methods Appl. Sci., 25 (2015), 255-282. doi: 10.1142/S0218202515500098
    [11] L. Boudin and F. Salvarani, A kinetic approach to the study of opinion formation, ESAIM: Math. Mod. Num. Anal., 43 (2009), 507-522. doi: 10.1051/m2an/2009004
    [12] R. M. Colombo and N. Pogodaev, Confinement strategies in a model for the interaction between individuals and a continuum, SIAM J. Appl. Dyn. Syst., 11 (2012), 741-770. doi: 10.1137/110854321
    [13] _______, On the control of moving sets: Positive and negative confinement results, SIAM J. Control Optim., 51 (2013), 380-401.
    [14] S. Cordier, L. Pareschi and G. Toscani, On a kinetic model for a simple market economy, J. Stat. Phys., 120 (2005), 253-277. doi: 10.1007/s10955-005-5456-0
    [15] I. Couzin, J. Krause, N. Franks and S. Levin, Effective leadership and decision-making in animal groups on the move, Nature, 433 (2005), 513-516. doi: 10.1038/nature03236
    [16] F. Cucker and S. Smale, Emergent behavior in flocks, IEEE Transactions on Automatic Control, 52 (2007), 852-862. doi: 10.1109/TAC.2007.895842
    [17] P. Degond, M. Herty and J.-G. Liu, Flow on sweeping networks, Multiscale Modeling & Simulation, 12 (2014), 538-565. doi: 10.1137/130927061
    [18] P. Degond, J. Liu and C. Ringhofer, Evolution of the distribution of wealth in economic neighborhood by local nash equilibrium closure, preprint, 2013.
    [19] P. Degond, J.-G. Liu, S. Motsch and V. Panferov, Hydrodynamic models of self-organized dynamics: Derivation and existence theory, Methods and Applications of Analysis, 20 (2013), 89-114. doi: 10.4310/MAA.2013.v20.n2.a1
    [20] P. Degond, J.-G. Liu and C. Ringhofer, Large-scale dynamics of mean-field games driven by local nash equilibria, Journal of Nonlinear Science, 24 (2014), 93-115. doi: 10.1007/s00332-013-9185-2
    [21] P. Degond and S. Motsch, Continuum limit of self-driven particles with orientation interaction, Math. Models Methods Appl. Sci., 18 (2008), 1193-1215. doi: 10.1142/S0218202508003005
    [22] B. Düring, P. Markowich, J.-F. Pietschmann and M.-T. Wolfram, Boltzmann and Fokker-Planck equations modelling opinion formation in the presence of strong leaders, Proc. R. Soc. A, 465 (2009), 3687-3708. doi: 10.1098/rspa.2009.0239
    [23] A. Fleig, L. Grüne and R. Guglielmi, Some results on model predictive control for the Fokker-Planck equation, in Proceedings of the 21st International Symposium on Mathematical Theory of Networks and Systems (MTNS 2014), 2014, 1203-1206.
    [24] M. Fornasier, J. Haskovec and G. Toscani, Fluid dynamic description of flocking via Povzner-Boltzmann equation, Physica D, 240 (2011), 21-31. doi: 10.1016/j.physd.2010.08.003
    [25] M. Fornasier, B. Piccoli and F. Rossi, Mean-field sparse optimal control, Phil. Trans. R. Soc. A, 372 (2014), 20130400, 21pp. doi: 10.1098/rsta.2013.0400
    [26] M. Fornasier and F. Solombrino, Mean-field optimal control, ESAIM Control Optim. Calc. Var., 20 (2014), 1123-1152. doi: 10.1051/cocv/2014009
    [27] S. Galam, Y. Gefen and Y. Shapir, Sociophysics: A new approach of sociological collective behavior, J. Math. Sociology, 9 (1982), 1-13.
    [28] J. Gómez-Serrano, C. Graham and J.-Y. Le Boudec, The bounded confidence model of opinion dynamics, Math. Models Methods Appl. Sci., 22 (2012), 1150007, 46pp. doi: 10.1142/S0218202511500072
    [29] S.-Y. Ha and E. Tadmor, From particle to kinetic and hydrodynamic descriptions of flocking, Kinet. Relat. Models, 1 (2008), 415-435. doi: 10.3934/krm.2008.1.415
    [30] R. Hegselmann and U. Krause, Opinion dynamics and bounded confidence, models, analysis and simulation, Journal of Artificial Societies and Social Simulation, 5 (2002), p2.
    [31] M. Herty and L. Pareschi, Fokker-planck asymptotics for traffic flow models, Kinet. Relat. Models, 3 (2010), 165-179. doi: 10.3934/krm.2010.3.165
    [32] M. Herty and C. Ringhofer, Averaged kinetic models for flows on unstructured networks, Kinet. Relat. Models, 4 (2011), 1081-1096. doi: 10.3934/krm.2011.4.1081
    [33] _______, Feedback controls for continuous priority models in supply chain management, Comput. Methods Appl. Math., 11 (2011), 206-213.
    [34] J.-M. Lasry and P.-L. Lions, Mean field games, Japanese Journal of Mathematics, 2 (2007), 229-260. doi: 10.1007/s11537-007-0657-8
    [35] D. Maldarella and L. Pareschi, Kinetic models for socio-economic dynamics of speculative markets, Physica A, 391 (2012), 715-730. doi: 10.1016/j.physa.2011.08.013
    [36] P. Markowich, M. Burger and L. Caffarelli, eds., Partial differential equation models in the socio-economic sciences, Philos. Trans. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci., 372 (2014), 20130406, 8pp. doi: 10.1098/rsta.2013.0406
    [37] D. Q. Mayne and H. Michalska, Receding horizon control of nonlinear systems, IEEE Trans. Automat. Control, 35 (1990), 814-824. doi: 10.1109/9.57020
    [38] S. Motsch and E. Tadmor, Heterophilious dynamics enhances consensus, SIAM Review, 56 (2014), 577-621. doi: 10.1137/120901866
    [39] G. Naldi, L. Pareschi and G. Toscani, eds., Mathematical Modeling of Collective Behavior in Socio-Economic and Life Sciences, Modeling and Simulation in Science, Engineering and Technology, Birkhäuser, Boston, 2010. doi: 10.1007/978-0-8176-4946-3
    [40] L. Pareschi and G. Toscani, Interacting Multi-Agent Systems. Kinetic Equations & Monte Carlo Methods, Oxford University Press, 2013.
    [41] E. D. Sontag, Mathematical Control Theory, Texts in Applied Mathematics, 6, Springer-Verlag, New York, 1998. doi: 10.1007/978-1-4612-0577-7
    [42] G. Toscani, Kinetic models of opinion formation, Commun. Math. Sci., 4 (2006), 481-496. doi: 10.4310/CMS.2006.v4.n3.a1
    [43] T. Vicsek and A. Zafeiris, Collective motion, Physics Reports, 517 (2012), 71-140. doi: 10.1016/j.physrep.2012.03.004
  • This article has been cited by:

    1. Fengwei Li, Qingfang Ye, Extremal graphs with given parameters in respect of general ABS index, 2024, 482, 00963003, 128974, 10.1016/j.amc.2024.128974
    2. Shabana Anwar, Muhammad Kamran Jamil, Amal S. Alali, Mehwish Zegham, Aisha Javed, Extremal values of the first reformulated Zagreb index for molecular trees with application to octane isomers, 2023, 9, 2473-6988, 289, 10.3934/math.2024017
  • Reader Comments
  • © 2015 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(5585) PDF downloads(222) Cited by(26)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog