Opinion formation in voting processes under bounded confidence

  • Received: 01 October 2018 Revised: 01 March 2019
  • 90B10, 91D30, 91B12, 34D20, 15B51

  • In recent years, opinion dynamics has received an increasing attention and various models have been introduced and evaluated mainly by simulation. In this study, we introduce a model inspired by the so-called "bounded confidence" approach where voters engaged in an electoral decision with two options are influenced by individuals sharing an opinion similar to their own. This model allows one to capture salient features of the evolution of opinions and results in final clusters of voters. We provide a detailed study of the model, including a complete taxonomy of the equilibrium points and an analysis of their stability. The model highlights that the final electoral outcome depends on the level of interaction in the society, besides the initial opinion of each individual, so that a strongly interconnected society can reverse the electoral outcome as compared to a society with looser exchange.

    Citation: Sergei Yu. Pilyugin, M. C. Campi. Opinion formation in voting processes under bounded confidence[J]. Networks and Heterogeneous Media, 2019, 14(3): 617-632. doi: 10.3934/nhm.2019024

    Related Papers:

    [1] Wei Liu, Qinghua Zuo, Chen Xu . Finite-time and global Mittag-Leffler stability of fractional-order neural networks with S-type distributed delays. AIMS Mathematics, 2024, 9(4): 8339-8352. doi: 10.3934/math.2024405
    [2] Li Wan, Qinghua Zhou, Hongbo Fu, Qunjiao Zhang . Exponential stability of Hopfield neural networks of neutral type with multiple time-varying delays. AIMS Mathematics, 2021, 6(8): 8030-8043. doi: 10.3934/math.2021466
    [3] Snezhana Hristova, Antonia Dobreva . Existence, continuous dependence and finite time stability for Riemann-Liouville fractional differential equations with a constant delay. AIMS Mathematics, 2020, 5(4): 3809-3824. doi: 10.3934/math.2020247
    [4] Jin Gao, Lihua Dai . Anti-periodic synchronization of quaternion-valued high-order Hopfield neural networks with delays. AIMS Mathematics, 2022, 7(8): 14051-14075. doi: 10.3934/math.2022775
    [5] Yucai Ding, Hui Liu . A new fixed-time stability criterion for fractional-order systems. AIMS Mathematics, 2022, 7(4): 6173-6181. doi: 10.3934/math.2022343
    [6] Qinghua Zhou, Li Wan, Hongbo Fu, Qunjiao Zhang . Exponential stability of stochastic Hopfield neural network with mixed multiple delays. AIMS Mathematics, 2021, 6(4): 4142-4155. doi: 10.3934/math.2021245
    [7] Muhammad Tariq, Sotiris K. Ntouyas, Hijaz Ahmad, Asif Ali Shaikh, Bandar Almohsen, Evren Hincal . A comprehensive review of Grüss-type fractional integral inequality. AIMS Mathematics, 2024, 9(1): 2244-2281. doi: 10.3934/math.2024112
    [8] Saad Ihsan Butt, Artion Kashuri, Muhammad Umar, Adnan Aslam, Wei Gao . Hermite-Jensen-Mercer type inequalities via Ψ-Riemann-Liouville k-fractional integrals. AIMS Mathematics, 2020, 5(5): 5193-5220. doi: 10.3934/math.2020334
    [9] Iman Ben Othmane, Lamine Nisse, Thabet Abdeljawad . On Cauchy-type problems with weighted R-L fractional derivatives of a function with respect to another function and comparison theorems. AIMS Mathematics, 2024, 9(6): 14106-14129. doi: 10.3934/math.2024686
    [10] Rabah Khaldi, Assia Guezane-Lakoud . On a generalized Lyapunov inequality for a mixed fractional boundary value problem. AIMS Mathematics, 2019, 4(3): 506-515. doi: 10.3934/math.2019.3.506
  • In recent years, opinion dynamics has received an increasing attention and various models have been introduced and evaluated mainly by simulation. In this study, we introduce a model inspired by the so-called "bounded confidence" approach where voters engaged in an electoral decision with two options are influenced by individuals sharing an opinion similar to their own. This model allows one to capture salient features of the evolution of opinions and results in final clusters of voters. We provide a detailed study of the model, including a complete taxonomy of the equilibrium points and an analysis of their stability. The model highlights that the final electoral outcome depends on the level of interaction in the society, besides the initial opinion of each individual, so that a strongly interconnected society can reverse the electoral outcome as compared to a society with looser exchange.



    Let $ W $ be a set and $ H:W\longrightarrow W $ be a mapping. A point $ w\in W $ is called a fixed point of $ H\ $if $ w = Hw. $ Fixed point theory plays a fundamental role in functional analysis (see [15]). Shoaib [17] introduced the concept of $ \alpha $-dominated mapping and obtained some fixed point results (see also [1,2]). George et al. [11] introduced a new space and called it rectangular $ b $-metric space ($ r.b.m. $ space). The triangle inequality in the $ b $-metric space was replaced by rectangle inequality. Useful results on $ r.b.m. $ spaces can be seen in ([5,6,8,9,10]). Ćirić introduced new types of contraction and proved some metrical fixed point results (see [4]). In this article, we introduce Ćirić type rational contractions for $ \alpha $ -dominated mappings in $ r.b.m. $ spaces and proved some metrical fixed point results. New interesting results in metric spaces, rectangular metric spaces and $ b $-metric spaces can be obtained as applications of our results.

    Definition 1.1. [11] Let $ U $ be a nonempty set. A function $ d_{lb}:U\times U\rightarrow \lbrack 0, \infty) $ is said to be a rectangular $ b $-metric if there exists $ b\geq 1 $ such that

    (ⅰ) $ d_{lb}(\theta, \nu) = d_{lb}(\nu, \theta $);

    (ⅱ) $ d_{lb}(\theta, \nu) = 0 $ if and only if $ \theta = \nu; $

    (ⅲ) $ d_{lb}(\theta, \nu)\leq b[d_{lb}(\theta, q)+d_{lb}(q, l)+d_{lb}(l, \nu)] $ for all $ \theta, \nu \in U $ and all distinct points $ q, l\in U \backslash \{ \theta, \nu \}. $

    The pair $ (U, d_{lb}) $ is said a rectangular $ b $-metric space (in short, $ r.b.m. $ space) with coefficient $ b $.

    Definition 1.2. [11] Let $ (U, d_{lb}) $ be an $ r.b.m. $ space with coefficient $ b $.

    (ⅰ) A sequence $ \{ \theta_{n}\} $ in $ (U, d_{lb}) $ is said to be Cauchy sequence if for each $ \varepsilon > 0 $, there corresponds $ n_{0}\in \mathbb{N} $ such that for all $ n, m\geq n_{0} $ we have $ d_{lb}(\theta_{m}, \theta_{n}) < \varepsilon $ or $ \lim_{n, m\rightarrow +\infty }d_{lb}(\theta_{n}, \theta_{m}) = 0. $

    (ⅱ) A sequence $ \{ \theta_{n}\} $ is rectangular $ b $-convergent (for short, ($ d_{lb} $)-converges) to $ \theta $ if $ \lim_{n\rightarrow +\infty }d_{lb}(\theta _{n}, \theta) = 0. $ In this case $ \theta $ is called a ($ d_{lb} $)-limit of $ \{ \theta _{n}\}. $

    (ⅲ) $ (U, d_{lb}) $ is complete if every Cauchy sequence in $ U \; d_{lb} $-converges to a point $ \theta \in U $.

    Let $ \varpi _{b} $, where $ b\geq 1, $ denote the family of all nondecreasing functions $ \delta _{b}:[0, +\infty)\rightarrow \lbrack 0, +\infty) $ such that $ \sum_{k = 1}^{+\infty }b^{k}\delta _{b}^{k}(t) < +\infty $ and $ b\delta _{b}(t) < t $ for all $ t > 0, $ where $ \delta _{b}^{k} $ is the $ k^{th} $ iterate of $ \delta _{b}. $ Also $ b^{n+1}\delta _{b}^{n+1}(t) = b^{n}b\delta _{b}(\delta _{b}^{n}(t)) < b^{n}\delta _{b}^{n}(t). $

    Example 1.3. [11] Let $ U \; = \mathbb{N}. $ Define $ d_{lb}:U\times U\rightarrow \mathbb{R} ^{+}\cup \{0\} $ such that $ d_{lb}(u, v) = d_{lb}(v, u) $ for all $ u, v\in U $ and $ \alpha > 0 $

    $ dlb(u,v)={0, if u=v;10α, if u=1, v=2;α, if u{1,2} and v{3};2α, if u{1,2,3} and v{4};3α, if u or v{1,2,3,4} and uv.
    $

    Then $ (U, d_{lb}) $ is an $ r.b.m. $ space with $ b = 2 > 1. $ Note that

    $ d(1,4)+d(4,3)+d(3,2)=5α<10α=d(1,2).
    $

    Thus $ d_{lb} $ is not a rectangular metric.

    Definition 1.4. [17] Let $ (U, d_{lb}) $ be an $ r.b.m. $ space with coefficient $ b $. Let $ S:U\rightarrow U $ be a mapping and $ \alpha :U\times U\rightarrow \lbrack 0, +\infty) $. If $ A\subseteq U, $ we say that the $ S $ is $ \alpha $-dominated on $ A, $ whenever $ \alpha (i, Si)\geq 1 $ for all $ i\in A. $ If $ A = U $, we say that $ S $ is $ \alpha $-dominated$. $

    For $ \theta, \nu \in U $, $ a > 0, $ we define $ D_{lb}(\theta, \nu) $ as

    $ Dlb(θ,ν)=max{dlb(θ,ν),dlb(θ,Sθ).dlb(ν,Sν)a+dlb(θ,ν),dlb(θ,Sθ),dlb(ν,Sν)}.
    $

    Now, we present our main result.

    Theorem 2.1. Let $ (U, d_{lb}) $ be a complete $ r.b.m. $ space with coefficient $ b $, $ \alpha :U\times U\rightarrow \lbrack 0, \infty), \; S:U\rightarrow U $, $ \{ \theta_{n}\} $ be a Picard sequence and $ S $ be a $ \alpha $-dominated mapping on $ \{ \theta_{n}\}. $ Suppose that, for some $ \delta _{b}\in \varpi _{b} $, we have

    $ dlb(Sθ,Sν)δb(Dlb(θ,ν)),
    $
    (2.1)

    for all $ \theta, \nu \in \{ \theta_{n}\} $ with $ \alpha (\theta, \nu)\geq 1 $. Then $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }\in U. $ Also, if (2.1) holds for $ \theta^{\ast } $ and $ \alpha (\theta_{n}, \theta^{\ast })\geq 1 $ for all $ n\in \mathbb{N} \cup \{0\} $, then $ S $ has a fixed point $ \theta^{\ast } $ in $ U. $

    Proof. Let $ \theta_{0}\in U $ be arbitrary. Define the sequence $ \{ \theta _{n}\} $ by $ \theta_{n+1} = S \theta_{n} $ for all $ n\in \mathbb{N} \cup \{0\}. $ We shall show that $ \{ \theta_{n}\} $ is a Cauchy sequence. If $ \theta _{n} = \theta_{n+1}, $ for some $ n\in \mathbb{N}, $ then $ \theta_{n} $ is a fixed point of $ S $. So, suppose that any two consecutive terms of the sequence are not equal. Since $ S:U\rightarrow U $ be an $ \alpha $-dominated mapping on $ \{ \theta_{n}\} $, $ \alpha (\theta _{n}, S \theta_{n})\geq 1 $ for all $ n\in \mathbb{N} \cup \{0\} $ and then $ \alpha (\theta_{n}, \theta_{n+1})\geq 1 $ for all $ n\in \mathbb{N} \cup \{0\}. $ Now by using inequality (2.1), we obtain

    $ dlb(θn+1,θn+2)=dlb(Sθn,Sθn+1)δb(Dlb(θn,θn+1))δb(max{dlb(θn,θn+1),dlb(θn,θn+1).dlb(θn+1,θn+2)a+dlb(θn,θn+1),dlb(θn,θn+1),dlb(θn+1,θn+2)})δb(max{dlb(θn,θn+1),dlb(θn+1,θn+2)}).
    $

    If $ \max \{d_{lb}(\theta_{n}, \theta_{n+1}), d_{lb}(\theta_{n+1}, \theta _{n+2})\} = d_{lb}(\theta_{n+1}, \theta_{n+2}), $ then

    $ dlb(θn+1,θn+2)δb(dlb(θn+1,θn+2))bδb(dlb(θn+1,θn+2)).
    $

    This is the contradiction to the fact that $ b\delta _{b}(t) < t $ for all $ t > 0. $ So

    $ max{dlb(θn,θn+1),dlb(θn+1,θn+2)}=dlb(θn,θn+1).
    $

    Hence, we obtain

    $ dlb(θn+1,θn+2)δb(dlb(θn,θn+1))δ2b(dlb(θn1,θn))
    $

    Continuing in this way, we obtain

    $ dlb(θn+1,θn+2)δn+1b(dlb(θ0,θ1)).
    $
    (2.2)

    Suppose for some $ n, m\in \mathbb{N} $ with $ m > n, $ we have $ \theta_{n} = \theta_{m} $. Then by (2.2)

    $ dlb(θn,θn+1)=dlb(θn,Sθn)=dlb(θm,Sθm)=dlb(θm,θm+1)δmnb(dlb(θn,θn+1))<bδb(dlb(θn,θn+1))
    $

    As $ d_{lb}(\theta_{n}, \theta_{n+1}) > 0, $ so this is not true, because $ b\delta _{b}(t) < t $ for all $ t > 0. $ Therefore, $ \theta_{n}\neq \theta_{m} $ for any $ n, m\in \mathbb{N} $. Since $ \sum_{k = 1}^{+\infty }b^{k}\delta _{b}^{k}(t) < +\infty, $ for some $ \nu \in \mathbb{N}, $ the series $ \sum_{k = 1}^{+\infty }b^{k}\delta _{b}^{k}(\delta _{b}^{\nu -1}(d_{lb}(\theta_{0}, \theta_{1}))) $ converges. As $ b\delta _{b}(t) < t, $ so

    $ bn+1δn+1b(δν1b(dlb(θ0,θ1)))<bnδnb(δν1b(dlb(θ0,θ1))), for all nN.
    $

    Fix $ \varepsilon > 0. $ Then $ \frac{\varepsilon }{2} = \varepsilon ^{\prime } > 0. $ For $ \varepsilon ^{\prime }, $ there exists $ \nu (\varepsilon ^{\prime })\in \mathbb{N} $ such that

    $ bδb(δν(ε)1b(dlb(θ0,θ1)))+b2δ2b(δν(ε)1b(dlb(θ0,θ1)))+<ε
    $
    (2.3)

    Now, we suppose that any two terms of the sequence $ \{ \theta_{n}\} $ are not equal. Let $ n, m\in \mathbb{N} $ with $ m > n > \nu (\varepsilon ^{\prime }). $ Now, if $ m > n+2 $,

    $ dlb(θn,θm)b[dlb(θn,θn+1)+dlb(θn+1,θn+2)+dlb(θn+2,θm)]b[dlb(θn,θn+1)+dlb(θn+1,θn+2)]+b2[dlb(θn+2,θn+3)+dlb(θn+3,θn+4)+dlb(θn+4,θm)]b[δnb(dlb(θ0,θ1))+δn+1b(dlb(θ0,θ1))]+b2[δn+2b(dlb(θ0,θ1))+δn+3b(dlb(θ0,θ1))]+b3[δn+4b(dlb(θ0,θ1))+δn+5b(dlb(θ0,θ1))]+bδnb(dlb(θ0,θ1))+b2δn+1b(dlb(θ0,θ1))+b3δn+2b(dlb(θ0,θ1))+=bδb(δn1b(dlb(θ0,θ1)))+b2δ2b(δn1b(dlb(θ0,θ1)))+.
    $

    By using (2.3), we have

    $ dlb(θn,θm)<bδb(δν(ε)1b(dlb(θ0,θ1)))+b2δ2b(δν(ε)1b(dlb(θ0,θ1)))+<ε<ε.
    $

    Now, if $ m = n+2 $, then we obtain

    $ dlb(θn,θn+2)b[dlb(θn,θn+1)+dlb(θn+1,θn+3)+dlb(θn+3,θn+2)]b[dlb(θn,θn+1)+b[dlb(θn+1,θn+2)+dlb(θn+2,θn+4)+dlb(θn+4,θn+3)]+dlb(θn+3,θn+2)]bdlb(θn,θn+1)+b2dlb(θn+1,θn+2)+bdlb(θn+2,θn+3)+b2dlb(θn+3,θn+4)+b3[dlb(θn+2,θn+3)+dlb(θn+3,θn+5)+dlb(θn+5,θn+4)]bdlb(θn,θn+1)+b2dlb(θn+1,θn+2)+(b+b3)dlb(θn+2,θn+3)+b2dlb(θn+3,θn+4)+b3dlb(θn+5,θn+4)+b4[dlb(θn+3,θn+4)+dlb(θn+4,θn+6)+dlb(θn+6,θn+5)]bdlb(θn,θn+1)+b2dlb(θn+1,θn+2)+(b+b3)dlb(θn+2,θn+3)+(b2+b4)dlb(θn+3,θn+4)+b3dlb(θn+5,θn+4)+b4dlb(θn+6,θn+5)+b5[dlb(θn+4,θn+5)+dlb(θn+5,θn+7)+dlb(θn+7,θn+6)]bdlb(θn,θn+1)+b2dlb(θn+1,θn+2)+(b+b3)dlb(θn+2,θn+3)+(b2+b4)dlb(θn+3,θn+4)+(b3+b5)dlb(θn+4,θn+5)+<2[bdlb(θn,θn+1)+b2dlb(θn+1,θn+2)+b3dlb(θn+2,θn+3)+b4dlb(θn+3,θn+4)+b5dlb(θn+4,θn+5)+]2[bδnb(dlb(θ0,θ1))+b2δn+1b(dlb(θ0,θ1))+b3δn+2b(dlb(θ0,θ1))+]<2[bδb(δν(ε)1b(dlb(θ0,θ1)))+b2δ2b(δν(ε)1b(dlb(θ0,θ1)))+]<2ε=ε.
    $

    It follows that

    $ limn,m+dlb(θn,θm)=0.
    $
    (2.4)

    Thus $ \{ \theta_{n}\} $ is a Cauchy sequence in $ (U, d_{lb}) $. As $ (U, d_{lb}) $ is complete, so there exists $ \theta^{\ast } $ in $ U $ such that $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }, $ that is,

    $ limn+dlb(θn,θ)=0.
    $
    (2.5)

    Now, suppose that $ d_{lb}(\theta^{\ast }, S \theta^{\ast }) > 0 $. Then

    $ dlb(θ,Sθ)b[dlb(θ,θn)+dlb(θn,θn+1)+dlb(θn+1,Sθ)b[dlb(θ,θn+1)+dlb(θn,θn+1)+dlb(Sθn,Sθ).
    $

    Since $ \alpha (\theta_{n}, \theta^{\ast })\geq 1, $ we obtain

    $ dlb(θ,Sθ)bdlb(θ,θn+1)+bdlb(θn,θn+1)+bδb(max{dlb(θn,θ),dlb(θ,Sθ).dlb(θn,θn+1)a+dlb(θn,θ), dlb(θn,θn+1) dlb(θ,Sθ)}).
    $

    Letting $ n\rightarrow +\infty $, and using the inequalities (2.4) and (2.5), we obtain $ d_{lb}(\theta^{\ast }, S \theta^{\ast })\leq b\delta _{b}(d_{lb}(\theta^{\ast }, S \theta^{\ast })). $ This is not true, because $ b\delta _{b}(t) < t $ for all $ t > 0 $ and hence $ d_{lb}(\theta^{\ast }, S \theta^{\ast }) = 0 $ or $ \theta^{\ast } = S \theta^{\ast } $. Hence $ S $ has a fixed point $ \theta^{\ast } $ in $ U. $

    Remark 2.2. By taking fourteen different proper subsets of $ D_{lb}(\theta, \nu), $ we can obtainvnew results as corollaries of our result in a complete $ r.b.m. $ space with coefficient $ b. $

    We have the following result without using $ \alpha $-dominated mapping.

    Theorem 2.3. Let $ (U, d_{lb}) $ be a complete $ r.b.m. $ space with coefficient $ b, \; S:U\rightarrow U $, $ \{ \theta_{n}\} $ be a Picard sequence. Suppose that, for some $ \delta _{b}\in \varpi _{b} $, we have

    $ dlb(Sθ,Sν)δb(Dlb(θ,ν))
    $
    (2.6)

    for all $ \theta, \nu \in \{ \theta_{n}\} $. Then $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }\in U. $ Also, if (2.6) holds for $ \theta^{\ast } $, then $ S $ has a fixed point $ \theta^{\ast } $ in $ U. $

    We have the following result by taking $ \delta _{b}\left(t\right) = ct $, $ t\in \mathbb{R} ^{+} $ with $ 0 < c < \frac{1}{b} $ without using $ \alpha $-dominated mapping.

    Theorem 2.4. Let $ (U, d_{lb}) $ be a complete $ r.b.m. $ space with coefficient $ b $, $ S:U\rightarrow U $, $ \{ \theta_{n}\} $ be a Picard sequence. Suppose that, for some $ 0 < c < \frac{1}{b} $, we have

    $ dlb(Sθ,Sν)c(Dlb(θ,ν))
    $
    (2.7)

    for all $ \theta, \nu \in \{ \theta_{n}\} $. Then $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }\in U. $ Also, if (2.7) holds for $ \theta^{\ast } $, then $ S $ has a fixed point $ \theta^{\ast } $ in $ U. $

    Ran and Reurings [16] gave an extension to the results in fixed point theory and obtained results in partially ordered metric spaces. Arshad et al. [3] introduced $ \preceq $-dominated mappings and established some results in an ordered complete dislocated metric space. We apply our result to obtain results in ordered complete $ r.b.m. $ space.

    Definition 2.5. $ (U, \preceq, d_{lb}) $ is said to be an ordered complete $ r.b.m. $ space with coefficient $ b $ if

    (ⅰ) $ (U, \preceq) $ is a partially ordered set.

    (ⅱ) $ (U, d_{lb}) $ is an $ r.b.m. $ space.

    Definition 2.6. [3] Let $ U $ be a nonempty set, $ \preceq $ is a partial order on $ \theta. $ A mapping $ S:U\rightarrow U $ is said to be $ \preceq $-dominated on $ A $ if $ a\preceq Sa $ for each $ a\in A\subseteq \theta. $ If $ A = U, $ then $ S:U\rightarrow U $ is said to be $ \preceq $-dominated.

    We have the following result for $ \preceq $-dominated mappings in an ordered complete $ r.b.m. $ space with coefficient $ b $.

    Theorem 2.7. Let $ (U, \preceq, d_{lb}) $ be an ordered complete $ r.b.m. $ space with coefficient $ b $, $ S:U\rightarrow U, \; \{ \theta_{n}\} $ be a Picard sequence and $ S $ be a $ \preceq $-dominated mapping on $ \{ \theta _{n}\}. $ Suppose that, for some $ \delta _{b}\in \varpi _{b} $, we have

    $ dlb(Sθ,Sν)δb(Dlb(θ,ν)),
    $
    (2.8)

    for all $ \theta, \nu \in \{ \theta_{n}\} $ with $ \theta \preceq \nu $. Then $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }\in U. $ Also, if (2.8) holds for $ \theta^{\ast } $ and $ \theta_{n}\preceq \theta^{\ast } $ for all $ n\in \mathbb{N} \cup \{0\} $. Then $ S $ has a fixed point $ \theta^{\ast } $ in $ U $.

    Proof. Let $ \alpha :U\times U\rightarrow \lbrack 0, +\infty) $ be a mapping defined by $ \alpha (\theta, \nu) = 1 $ for all $ \theta, \nu \in U $ with $ \theta \preceq \nu $ and $ \alpha (\theta, \nu) = \frac{4}{11} $ for all other elements $ \theta, \nu \in U. $ As $ S $ is the dominated mappings on $ \{ \theta_{n}\}, $ so $ \theta \preceq S \theta $ for all $ \theta \in \{ \theta_{n}\}. $ This implies that $ \alpha (\theta, S \theta) = 1 $ for all $ \theta \in \{ \theta_{n}\}. $ So $ S:U\rightarrow U $ is the $ \alpha $-dominated mapping on $ \{ \theta_{n}\}. $ Moreover, inequality (2.8) can be written as

    $ dlb(Sθ,Sν)δb(Dlb(θ,ν))
    $

    for all elements $ \theta, \nu $ in $ \{ \theta_{n}\} $ with $ \alpha (\theta, \nu)\geq 1. $ Then, as in Theorem 2.1, $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }\in U. $ Now, $ \theta_{n}\preceq \theta^{\ast } $ implies $ \alpha (\theta_{n}, \theta^{\ast })\geq 1. $ So all the conditions of Theorem 2.1 are satisfied. Hence, by Theorem 2.1, $ S $ has a fixed point $ \theta^{\ast } $ in $ U $.

    Now, we present an example of our main result. Note that the results of George et al. [11] and all other results in rectangular $ b $-metric space are not applicable to ensure the existence of the fixed point of the mapping given in the following example.

    Example 2.8. Let $ U = A\cup B, $ where $ A = \{\frac{1}{n}:n\in \{2, 3, 4, 5\}\} $ and $ B = [1, \infty ]. $ Define $ d_{l}:U\times U\rightarrow \lbrack 0, \infty) $ such that $ d_{l}(\theta, \nu) = d_{l}(\nu, \theta) $ for $ \theta, \nu \in U $ and

    $ {dl(12,13)=dl(14,15)=0.03dl(12,15)=dl(13,14)=0.02dl(12,14)=dl(15,13)=0.6dl(θ,ν)=|θν|2    otherwise
    $

    be a complete $ r.b.m. $ space with coefficient $ b \; = 4 > 1 $ but$ \ (U, d_{l}) $ is neither a metric space nor a rectangular metric space. Take $ \delta _{b}(t) = \frac{t}{10}, $ then $ b\delta _{b}(t) < t. $ Let $ S:U\rightarrow U $ be defined as

    $ Sθ={15        ifθA13        ifθ=19θ100+85 otherwise.
    $

    Let $ \theta_{0} = 1. $ Then the Picard sequence $ \{ \theta_{n}\} $ is $ \{1, \frac{1}{ 3}, \frac{1}{5}, \frac{1}{5}, \frac{1}{5}, \cdots \} $. Define

    $ α(θ,ν)={85        ifθ,ν{θn}47            otherwise.
    $

    Then $ S\ $is an $ \alpha $-dominated mapping on $ \{ \theta_{n}\}. $ Now, $ S $ satisfies all the conditions of Theorem 2.1. Here $ \frac{1}{5} $ is the fixed point in $ U. $

    Jachymski [13] proved the contraction principle for mappings on a metric space with a graph. Let $ (U, d) $ be a metric space and $ \bigtriangleup $ represents the diagonal of the cartesian product $ U\times U. $ Suppose that $ G $ be a directed graph having the vertices set $ V(G) $ along with $ U, $ and the set $ E(G) $ denoted the edges of $ U_{{}} $ included all loops, i.e., $ E(G)\supseteq \bigtriangleup. $ If $ G $ has no parallel edges, then we can unify $ G $ with pair $ (V(G), E(G)) $. If $ l $ and $ m $ are the vertices in a graph $ G, $ then a path in $ G $ from $ l $ to $ m $ of length $ N \; (N\in \mathbb{N}) $ is a sequence $ \{ \theta_{i}\}_{i = o}^{N} $ of $ N+1 $ vertices such that $ l_{o} = l, \; l_{N} = m $ and $ (l_{n-1}, l_{n})\in E(G) $ where $ i = 1, 2, \cdots N $ (see for detail [7,8,12,14,18,19]). Recently, Younis et al. [20] introduced the notion of graphical rectangular $ b $-metric spaces (see also [5,6,21]). Now, we present our result in this direction.

    Definition 3.1. Let $ \theta $ be a nonempty set and $ G = (V(G), E(G)) $ be a graph such that $ V(G) = U $ and $ A\subseteq U $. A mapping $ S:U\rightarrow U $ is said to be graph dominated on $ A $ if $ (\theta, S \theta)\in E(G) $ for all $ \theta \in A $.

    Theorem 3.2. Let $ (U, d_{lb}) $ be a complete rectangular $ b $ -metric space endowed with a graph $ G $, $ \{ \theta_{n}\} $ be a Picard sequence and $ S:U\rightarrow U $ be a graph dominated mapping on $ \{ \theta_{n}\} $. Suppose that the following hold:

    (i) there exists $ \delta _{b}\in \varpi _{b} $ such that

    $ dlb(Sθ,Sν)δb(Dlb(θ,ν)),
    $
    (3.1)

    for all $ \theta, \nu \in \{ \theta_{n}\} $ and $ (\theta_n, \nu)\in E(G) $. Then $ (\theta_{n}, \theta_{n+1})\in E(G) $ and $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }. $ Also, if (3.1) holds for $ \theta^{\ast } $ and $ (\theta_{n}, \theta^{\ast })\in E(G) $ for all $ n\in \mathbb{N} \cup \{0\} $, then $ S $ has a fixed point $ \theta^{\ast } $ in $ U. $

    Proof. Define $ \alpha :U\times U\rightarrow \lbrack 0, +\infty) $ by

    $ α(θ,ν)={1, ifθ,νU, (θ,ν)E(G)14,                  otherwise.
    $

    Since $ S $ is a graph dominated on $ \{ \theta_{n}\}, $ for $ \theta \in \{ \theta_{n}\}, \; (\theta, S \theta)\in E(G) $. This implies that $ \alpha (\theta, S \theta) = 1 $ for all $ \theta \in \{ \theta_{n}\}. $ So $ S:U\rightarrow U $ is an $ \alpha $-dominated mapping on $ \{ \theta_{n}\}. $ Moreover, inequality (3.1) can be written as

    $ dlb(Sθ,Sν)δb(Dlb(θ,ν)),
    $

    for all elements $ \theta, \nu $ in $ \{ \theta_{n}\} $ with $ \alpha (\theta, \nu)\geq 1. $ Then, by Theorem 2.1, $ \{ \theta_{n}\} $ converges to $ \theta^{\ast }\in U. $ Now, $ (\theta_{n}, \theta^{\ast })\in E(G) $ implies that $ \alpha (\theta_{n}, \theta^{\ast })\geq 1. $ So all the conditions of Theorem 2.1 are satisfied. Hence, by Theorem 2.1, $ S $ has a fixed point $ \theta^{\ast } $ in $ U $.

    The authors would like to thank the Editor, the Associate Editor and the anonymous referees for sparing their valuable time for reviewing this article. The thoughtful comments of reviewers are very useful to improve and modify this article.

    The authors declare that they have no competing interests.



    [1] Cluster formation in opinion dynamics: A qualitative analysis. Z. Angew. Math. Phys. (2010) 61: 583-602.
    [2] Asymptotic analysis of continuous opinion dynamics models under bounded confidence. Commun. Pure Appl. Anal. (2013) 12: 1487-1499.
    [3] Continuous and discontinuous opinion dynamics with bounded confidence. Nonlinear Anal. Real World Appl. (2012) 13: 1239-1251.
    [4]

    S. Chatterjee, Reaching a consensus: Some limit theorems, Proc. Int. Statist. Inst., 159–164.

    [5] Toward consensus: Some convergence theorems on repeated averaging. J. Appl. Prob. (1977) 14: 89-97.
    [6] Approaching consensus can be delicate when positions harden. Stochastic Proc. and Appl. (1986) 22: 315-322.
    [7] Mixing beliefs among interacting agents. Advances in Complex Systems (2000) 3: 87-98.
    [8] Consensus formation under bounded confidence. Nonlinear Analysis (2001) 47: 4615-4621.
    [9] The Krause – Hegselmann consensus model with discrete opinions. Internat. J. Modern Phys. C (2004) 15: 1021-1029.
    [10] A formal theory of social power. Psychological Review (1956) 63: 181-194.
    [11] Reaching a consensus. J. Amer. Statist. Assoc. (1974) 69: 118-121.
    [12]

    F. Harary, A criterion for unanimity in French's theory of social power, in Studies in Social Power (ed. D. Cartwright), Institute for Social Research, Ann Arbor, 1959.

    [13] The Hegselmann – Krause dynamics for equally spaced agents. J. Difference Equ. Appl. (2016) 22: 1621-1645.
    [14]

    R. Hegselmann and A. Flache, Understanding complex social dynamics: A plea for cellular automata based modelling, Journal of Artificial Societies and Social Simulation, 1 (1998).

    [15] Opinion dynamics and bounded confidence: Models, analysis and simulation. Journal of Artificial Societies and Social Simulation (2002) 5: 1-33.
    [16] Opinion dynamics under the influence of radical groups, charismatic leaders, and other constant signals: A simple unifying model. Networks and Heterogeneous Media (2015) 10: 477-509.
    [17] Clustering and asymptotic behavior in opinion formation. J. Differential Equations (2014) 257: 4165-4187.
    [18]

    U. Krause, A discrete nonlinear and non - autonomous model of consensus formation, in Communications in Difference Equations (ed. S. Elaydi et al.), Gordon and Breach Publ., Amsterdam, 2000.

    [19]

    U. Krause, Soziale dynamiken mit vielen interakteuren. eine problemskizze, in Modellierung und Simulation von Dynamiken mit vielen interagierenden Akteuren (eds. U. Krause and M. Stockler), Universitat Bremen, 1997.

    [20] Compromise, consensus, and the iteration of means. Elem. Math. (2009) 64: 1-8.
    [21]

    U. Krause, Positive Dynamical Systems in Discrete Time. Theory, Models, and Applications, De Gruyter, Berlin, 2015.

    [22] On the Hegselmann – Krause conjecture in opinion dynamics. Journal of Difference Equations and Applications (2011) 17: 859-876.
    [23] Social consensus and rational agnoiology. Synthese (1975) 31: 141-160.
    [24]

    K. Lehrer and C. Wagner, Rational Consensus in Science and Society, D. Reidel Publ. Co., Dordrecht-Boston, Mass., 1981.

    [25] A stabilization theorem for dynamics of continuous opinions. Physica A (2005) 355: 217-223.
    [26] Continuous opinion dynamics under bounded confidence: A survey. International Journal of Modern Physics C (2007) 18: 1819-1838.
    [27]

    W. Ren and Y. Cao, Distributed Coordination of Multi-agent Networks. Emergent Problems, Models, and Issues, Springer, 2011.

    [28] Consensus through respect: A model of rational group decision-making. Philosophical Studies (1978) 34: 335-349.
    [29] The Hegselmann – Krause dynamics for the continuous-agent model and a regular opinion function do not always lead to consensus. IEEE Trans. Automat. Control (2015) 60: 2416-2421.
    [30]

    G. Weisbuch, G. Deffuant and G. Nadal, Interacting agents and continuous opinion dynamics, in Heterogenous Agents, Interactions and Economic Performance (eds. R. Cowan and N. Jonard), Lecture Notes in Economics and Mathematical Systems, 521, Springer, Berlin, 2003.

    [31] On the control through leadership of the Hegselmann – Krause opinion formation model. Mathematical Models and Methods in Applied Sciences (2015) 25: 565-585.
  • This article has been cited by:

    1. Haydar Akca, Chaouki Aouiti, Farid Touati, Changjin Xu, Finite-time passivity of neutral-type complex-valued neural networks with time-varying delays, 2024, 21, 1551-0018, 6097, 10.3934/mbe.2024268
  • Reader Comments
  • © 2019 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(3600) PDF downloads(348) Cited by(17)

Figures and Tables

Figures(5)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog