Research article

Outer space branching search method for solving generalized affine fractional optimization problem

  • Received: 18 July 2022 Revised: 23 September 2022 Accepted: 10 October 2022 Published: 26 October 2022
  • MSC : 90C26, 90C32, 65K05

  • This paper proposes an outer space branching search method, which is used to globally solve the generalized affine fractional optimization problem (GAFOP). First, we will convert the GAFOP into an equivalent problem (EP). Next, we structure the linear relaxation problem (LRP) of the EP by using the linearization technique. By subsequently partitioning the initial outer space rectangle and successively solving a series of LRPs, the proposed algorithm globally converges to the optimum solution of the GAFOP. Finally, comparisons of numerical results are reported to show the superiority and the effectiveness of the presented algorithm.

    Citation: Junqiao Ma, Hongwei Jiao, Jingben Yin, Youlin Shang. Outer space branching search method for solving generalized affine fractional optimization problem[J]. AIMS Mathematics, 2023, 8(1): 1959-1974. doi: 10.3934/math.2023101

    Related Papers:

    [1] Yanyan Cui, Chaojun Wang . Systems of two-dimensional complex partial differential equations for bi-polyanalytic functions. AIMS Mathematics, 2024, 9(9): 25908-25933. doi: 10.3934/math.20241265
    [2] Xiaojun Hu, Qihan Wang, Boyong Long . Bohr-type inequalities for bounded analytic functions of Schwarz functions. AIMS Mathematics, 2021, 6(12): 13608-13621. doi: 10.3934/math.2021791
    [3] Yaser Khatib, Stanford Shateyi . Improvement of inequalities related to powers of the numerical radius. AIMS Mathematics, 2024, 9(7): 19089-19103. doi: 10.3934/math.2024930
    [4] Baishuai Zuo, Chuancun Yin . Stein’s lemma for truncated generalized skew-elliptical random vectors. AIMS Mathematics, 2020, 5(4): 3423-3433. doi: 10.3934/math.2020221
    [5] Maimoona Karim, Aliya Fahmi, Zafar Ullah, Muhammad Awais Tariq Bhatti, Ather Qayyum . On certain Ostrowski type integral inequalities for convex function via AB-fractional integral operator. AIMS Mathematics, 2023, 8(4): 9166-9184. doi: 10.3934/math.2023459
    [6] Jamshed Nasir, Shahid Qaisar, Saad Ihsan Butt, Ather Qayyum . Some Ostrowski type inequalities for mappings whose second derivatives are preinvex function via fractional integral operator. AIMS Mathematics, 2022, 7(3): 3303-3320. doi: 10.3934/math.2022184
    [7] Xue Zhang, Yu Xue . A novel H state observer design method for genetic regulatory networks with time-varying delays. AIMS Mathematics, 2024, 9(2): 3763-3787. doi: 10.3934/math.2024185
    [8] Huiping Jiao, Xiao Zhang, Chao Wei . L-norm minimum distance estimation for stochastic differential equations driven by small fractional Lévy noise. AIMS Mathematics, 2023, 8(1): 2083-2092. doi: 10.3934/math.2023107
    [9] Pshtiwan Othman Mohammed, Dumitru Baleanu, Thabet Abdeljawad, Eman Al-Sarairah, Y. S. Hamed . Monotonicity and extremality analysis of difference operators in Riemann-Liouville family. AIMS Mathematics, 2023, 8(3): 5303-5317. doi: 10.3934/math.2023266
    [10] Yunpeng Zhao, Fei He, Shumin Lu . Several fixed-point theorems for generalized Ćirić-type contraction in Gb-metric spaces. AIMS Mathematics, 2024, 9(8): 22393-22413. doi: 10.3934/math.20241089
  • This paper proposes an outer space branching search method, which is used to globally solve the generalized affine fractional optimization problem (GAFOP). First, we will convert the GAFOP into an equivalent problem (EP). Next, we structure the linear relaxation problem (LRP) of the EP by using the linearization technique. By subsequently partitioning the initial outer space rectangle and successively solving a series of LRPs, the proposed algorithm globally converges to the optimum solution of the GAFOP. Finally, comparisons of numerical results are reported to show the superiority and the effectiveness of the presented algorithm.



    Let D be the unit disk in the complex plane C centered at the origin and denote T=D. The symbol Cn(Ω) stands for the class of all complex-valued n-times continuously differentiable functions from the domain Ω into C. In particular, we let C(Ω)=C0(Ω). A real-valued function uC2(Ω) is called real harmonic if it satisfies the following Laplace's equation:

    Δu(z)=2ux2(z)+2uy2(z)=0,zΩ.

    A complex-valued function ω=u+iv is harmonic if both u and v are real harmonic. We refer the readers to [12] for an excellent discussion on harmonic mappings in the plane.

    Suppose that gC(D). Then, it is well known that the solution of the Poisson's equation Δω=g in D satisfying the boundary condition ω|T=fL1(T) is given by

    ω(z)=P[f](z)G[g](z),zD, (1.1)

    where

    P[f](z)=12π2π0P(z,eiφ)f(eiφ)dφ,G[g](z)=DG(z,w)g(w)dm(w),P(z,eiθ)=1|z|2|zeiθ|2,zD,θR,G(z,w)=12πlog|1z¯wzw|,z,wD,zw, (1.2)

    and m denotes the Lebesgue measure in the plane.

    The hyperbolic metric on D, the Gaussian curvature of which is equal to -4, is given by |dz|/(1|z|2). Then, the hyperbolic distance between two points z1,z2D is defined by

    dhD(z1,z2):=infγ{γ|dz|1|z|2}, (1.3)

    where the infimum is taken over all the rectifiable curves γ which connect z1 and z2 with γD. It is well known that

    dhD(z1,z2)=log|1z1¯z2|+|z1z2||1z1¯z2||z1z2|.

    Also, we can consider the quasihyperbolic metric |dz|/(1|z|) on D. Then, it gives rise to the quasihyperbolic distance between two points z1,z2D, which is defined by

    dqhD(z1,z2):=infγ{γ|dz|1|z|}, (1.4)

    where the infimum is taken over all rectifiable curves γ which connect z1 and z2 with γD. Actually, we can consider the general q-metric |dz|/(1|z|q) on D,q[1,2], which leads to the following q-pseudo distance

    dq(z1,z2):=infγ{γ|dz|1|z|q},q[1,2], (1.5)

    where the infimum is taken over all rectifiable curves γ which connect z1 and z2 with γD. For completeness, we will prove that this q-pseudo distance is a distance in Lemma 7.

    A self-mapping f of D is said to be Lipschitz continuous with respect to the q-distance if there exists a constant L>0, such that the inequality

    dq(f(z1),f(z2))Ldq(z1,z2)

    holds for any z1,z2D.

    The classic Schwarz-Pick lemma plays an important role in complex analysis. It is stated as follows: If f is a holomorphic self-mapping of D, then we have

    |f(z)|1|f(z)|21|z|2,zD, (1.6)

    or

    dhD(f(z1),f(z2))dhD(z1,z2),zD. (1.7)

    The Schwarz-Pick lemmas (1.6) or (1.7) have many generalizations. See the references [1,6,13,30,31,32,34,35,39,46] on this topic. The first work of this paper is to extend (1.7) to the case for the q-distance. It is presented as follows:

    Theorem 1. Suppose that q[1,2]. If f is a holomorphic self-mapping of D satisfying f(0)=0, then the inequality

    dq(f(z1),f(z2))dq(z1,z2) (1.8)

    holds for any z1,z2D.

    Suppose that f satisfies the following the Poisson differential inequality:

    |Δf(z)|a|Df(z)|2+b, (1.9)

    where |Df|=|fz|+|f¯z|. This kind of mapping contains many classical mappings. For example, it includes a holomorphic mapping, a harmonic mapping, the mapping satisfying Poisson's equation Δf=g (where g is continuous on ¯D), the harmonic mapping between Riemannian surfaces satisfying some inequalities and so on. In addition, those mappings that satisfy (1.9) are also closely related to the following partial differential equations

    Δu=Q(ux1,ux2,u,x1,x2);u=(u1(x1,x2),u2(x1,x2),,um(x1,x2)), (1.10)

    where Q=(Q1,Q2,,Qm), Qj are quadratic polynomial on ui/xk, i=1,,m,k=1,2. This equation has a deep connection with the average curvature problem of surfaces in differential geometry, the conjugate isothermal coordinates of surfaces, and the Monge-Ampère equation and so on. Based on this backgrounds, the study of mappings that satisfy the Poisson differential inequality (1.9) has attracted much attention from researchers. See the articles [4,18,20,42,45] and the references therein. Recently, Kalaj [21] extended the results of Bernstein [4] and Heinz [18] to the case for spaces and applied the relevant results to the theory of harmonic quasiconformal mappings. The study of mappings satisfying the Poisson differential inequality that are also quasiconformal or quasiregular have also attracted many authors' interest, see [3,10,22].

    Let f be a quasiconformal self-mapping of the unit disk satisfying the following Poisson differential equation

    |Δf(z)|B|Df(z)|2. (1.11)

    In [23], Kalaj obtained the following result.

    Lemma 1. [23] Suppose that f is a K-quasiconformal self-mapping of D satisfying the Poisson differential equation (1.11), and that f(0)=0. Then there exists a constant C(B,K), which depends only on B and K, such that

    1|z|21|f(z)|2C(B,K). (1.12)

    The inequality (1.12) can be seen as a kind of Schwarz-Pick type inequality for the K-quasiconformal self-mapping of D satisfying the Poisson differential inequality (1.11). Generally, the condition for quasiconformality in Lemma 1 can not be removed. Recently, Zhong et al. [47] showed the following result.

    Theorem 2. [47, Theorem 1.4] For a given q{1}[2,+). Suppose that f:¯D¯D is continuous on ¯D, f|DC2,f|TC2 and f(0)=0, and that it satisfies the following Poisson differential inequality

    |Δf(z)|a|Df(z)|2+b,

    as well as satisfying that |2f(eiφ)φ2|K, where 0<a<1/2,0<b,K<. If

    2q1q+max{a,b}(L2(a,b,K)+1)4<2π,

    then, we have

    1|z|q1|f(z)|q{12π3max{a,b}(L2(a,b,K)+1)4whenq=1,12π(2q1q+1)max{a,b}(L2(a,b,K)+1)4whenq2.

    In the next result of this paper, we will improve Theorem 2 by giving the explicit constant L(a,b,K) and showing that the result is still valid when we consider all of the cases for q>0. Our result is presented as follows.

    Theorem 3. For given q>0, suppose that f:¯D¯D is continuous in ¯D, f|DC2,f|TC2 and f(0)=0, and that it satisfies the Poisson differential inequality |Δf(z)|a|Df(z)|2+b and |2f(eiφ)φ2|K, where 0<a<1/2,0<b,K<. If, in addition, 48ab+640a<9 and max{a,b}(L2(a,b,K)+1)<2/π, then we have

    1|z|q1|f(z)|qmax{1,q}min{1,q}(2πmax{a,b}(L2(a,b,K)+1)), (1.13)

    where

    L(a,b,K)=2max{8+64a3(1a),b+8ab3(1a)}×[1+1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a)]. (1.14)

    As an application of Theorem 3, we provide the following more general form of [47, Corollary 1.5].

    Corollary 1. For a given q>0, suppose that f:¯D¯D is continuous in ¯D, f|DC2,f|TC2 and f(0)=0, and that it satisfies the Poisson differential inequality |Δf(z)|a|Df(z)|2+b and |2f(eiφ)φ2|K, where 0<a<1/2,0<b,K<. If, in addition, 48ab+640a<9 and max{a,b}(L2(a,b,K)+1)<2/π, then we have

    dq(f(z1),f(z2))max{1,q}L(a,b,K)min{1,q}(2πmax{a,b}(L2(a,b,K)+1))dq(z1,fz2). (1.15)

    Proof. According to Lemma 9, we see that there exists a constant L(a,b,K) such that the inequality

    |Df(z)|L(a,b,K)

    holds for any zD. Combining this with Theorem 3, we get

    |Df(z)|1|f(z)|qmax{1,q}L(a,b,K)min{1,q}(2πmax{a,b}(L2(a,b,K)+1))11|z|q.

    By the definition of q-distance, we derive that inequality (1.15) is valid.

    The classic boundary Schwarz lemma for the holomorphic mapping states the following:

    Lemma 2. ([14]) Let f be a holomorphic self-mapping of D satisfying that f(0)=0. If f is differentiable at the point z=1 and satisfies that f(1)=1, then we have the follwoing:

    (i) f(1)1;

    (ii) f(1)=1 if and only if f(z)=z.

    The boundary Schwarz lemma is a basic and important result in complex analysis. Its generalizations and applications have received widespread attention; see [5,7,9,15,16,17,25,26,27,28,29,40,43,44]. In [25], Lemma 2 was generalized to the following form:

    Lemma 3. ([25]) Let f be a holomorphic self-mapping of D satisfying that f(0)=0. If f is differentiable at the point z=αT and satisfies that f(α)=βT. Then we have the following

    (i) ¯βf(α)α1;

    (ii) ¯βf(α)α=1 if and only if f(z)=eiθz, where eiθ=βα1 and θR.

    Recently, the extension of the boundary Schwarz lemma to harmonic mappings has also attracted scholars' attention. In [50], Zhu and Wang established the following harmonic version of the boundary Schwarz lemma.

    Lemma 4. ([50, Theorem 1.2]) Suppose that gC(¯D). If fC2(D)C(T) is a self-mapping of D satisfying the Poisson equation Δf=g and f(0)=0, and if f is differentiable at the point z=1 and satisfies that f(1)=1, then we have

    Re[fx(1)]2π34||g||, (1.16)

    where ||g||:=supzD|g(z)|.

    This result was extended by Mohapatra et al. [37] and Mohapatra [38]. See the high dimensional version of the boundary Schwarz lemma for harmonic mappings in [24,33,36,49]. As a generalization of planar harmonic maps, those self-mappings that satisfy the Poisson differential inequality (1.11) can be naturally considered as analogous to the boundary Schwarz lemma. Unfortunately, those types of mappings generally are not associated with the boundary Schwarz lemma of types (1.16). This is because we have established the following results:

    Theorem 4. There exists a self-mapping f of D, which is differentiable at z=1 and satisfies that f(0)=0 and f(1)=1, in addition to the Poisson differential inequality |Δf(z)|a|Df(z)|2. But it satisfies the equation Re(fx(1))=0.

    According to Lemma 4, we see that there exists a self-mapping of D that satisfies the Poisson differential inequality (1.9), which is associated with a boundary Schwarz lemma of type (1.16). Therefore, we naturally ask the following question: which subset of the family

    F={f:|Δf(z)|a|Df(z)|2+b,f(0)=0,fC2(D)}

    has boundary Schwarz lemma of type (1.16)? Next, we establish a result in this regard.

    Theorem 5. Suppose that f:¯D¯D is continuous in ¯D, f|DC2,f|TC2 and f(0)=0, and that it satisfies the Poisson differential inequality |Δf(z)|a|Df(z)|2+b and |2f(eiφ)φ2|K, where 0<a<1/2,0<b,K<. If, in addition, 48ab+640a<9 and max{a,b}(L2(a,b,K)+1)<8/(3π), then we have

    Re(fx(1))>2π3max{a,b}(L2(a,b,K)+1)4, (1.17)

    where

    L(a,b,K)=2max{8+64a3(1a),b+8ab3(1a)}×[1+1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a)]. (1.18)

    L' Hospital's rule for monotonicity is important to derive some inequalities and it will be used to prove some results.

    Lemma 5. ([2, Theorem 2]) Let <a<b<+ and let f,g:[a,b]R be continuous functions that are differentiable on (a,b) with f(a)=g(a)=0 or f(b)=g(b)=0. Suppose that g(x)0 for each x(a,b). If f/g is increasing (decreasing) on (a,b), then so is f/g.

    The following lemma can be obtained in [48] by letting x=t2.

    Lemma 6. For any q>0, the inequalities

    min{1,q}1xq1xmax{1,q} (2.1)

    hold for any x[0,1).

    Remark 1. In the rest of this paper, we will always use the constant cq, which is defined by

    cq:=min{1,q}max{1,q}. (2.2)

    In Section 1, we defined the following general q-pseudo distance in D:

    dq(z1,z2):=infγ{γ11|z|q|dz|},q[1,2], (2.3)

    where the infimum is taken over all rectifiable curves γ in D connected z1 and z2. Next, we will prove that it is a distance.

    Lemma 7. The function dq(,) defined by (2.3) is a distance.

    Proof. Obviously, for any z1,z2D, we have that dq(z1,z2)0 and dq(z1,z2)=dq(z2,z1). Hence, we only prove that if dq(z1,z2)=0, then z1=z2. This is because, by taking x=|z|2 in Lemma 6, we get

    11|z|2max{1,q/2}11|z|q.

    Hence, by the definitions of hyperbolic distance and q-pseudo distance, we have

    dhD(z1,z2)max{1,q/2}dq(z1,z2).

    Therefore, it must be that dhD(z1,z2)=0 if dq(z1,z2)=0. Since dhD(,) is a distance, we get that z1=z2. This shows that the q-pseudo distance dq(,) defined in D is a distance.

    We also need the following general Schwarz lemma for harmonic self-mappings of the unit disk, which was obtained by Hethcote [19].

    Lemma 8. ([19, Theorem 1]) Suppose that f is a harmonic self-mapping of D. Then we have

    |f(z)1|z|21+|z|2f(0)|4πarctan|z|,zD. (2.4)

    In [18], Heinz obtained a gradient estimate for self-mappings of D satisfying the Poisson differential inequality. The upper bound he obtained was abstract. In what follows, we give an explicitly upper bound based on the proof in [18]. The disadvantage is that the range of a,b becomes smaller.

    Lemma 9. Suppose that f:¯D¯D is continuous in ¯D, f|DC2,f|TC2 and f(0)=0, and that it satisfies the Poisson differential inequality |Δf(z)|a|Df(z)|2+b and |2f(eiφ)φ2|K, where 0<a<1/2,0<b,K<. If 48ab+640a<9, then the inequality

    |Df(z)|L(a,b,K) (2.5)

    holds for any zD, where

    L(a,b,K)=2max{8+64a3(1a),b+8ab3(1a)}×[1+1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a)]. (2.6)

    Proof. First, we calculate c10(a,b,K) in [18, P. 239]. According to [18, Lemma 11] and given

    |ˆF(φ)|8K2π2+2K

    in [18, P. 239] and [18, (2.2.45), P. 238], we get

    |X(w)X(w0)|[1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a)](1r).

    Hence, by [18, (2.2.47), P. 239], we can take

    c10(a,b,K)=1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a). (2.7)

    Next, we will estimate the constant c5(a,b) in [18]. According to the proof of [18, Theorem 2], we can take α=2,β=2(1a) and γ=2b in [18, (2.1.27)]. Therefore, the formula given by [18, (2.1.27)] can be rewritten as

    4AB16aλθ(1λ)2+4abλ(1λ)2+a2(b+16)(1a)(1λ)2ln1θ, (2.8)

    where λ and θ are constants satisfying that 0<λ<θ<1. Now, we take θ=2λ and assume that 0<λ<1/4. Since ln1θ1θ and a21aa (as a<1/2), we have

    4AB8a(1λ)2+4abλ(1λ)2+a(b+16)(1λ)2(12λ). (2.9)

    As the functions φ1(λ)=8a+4abλ+a(b+16)12λ,φ2(λ)=(λ1)2 are monotonically increasing on (0,1/4), we see that the inequality

    4AB8a(1λ)2+4abλ(1λ)2+a(b+16)(1λ)2(12λ)<1 (2.10)

    holds for any 0<λ<1/4 when 48ab+640a<9. Especially, inequality (2.10) holds for any 0<λ1/8. Now, we take λ0=1/8 and θ0=1/4. Then by [18, (2.1.31), P. 227], we have

    M2B(λ0)2max{8+64a3(1a),b+8ab3(1a)}(K+R0).

    Hence, by the inequality given by[18, (2.1.31)], we get

    c5(a,b)=2max{8+64a3(1a),b+8ab3(1a)}. (2.11)

    Combining [18, P. 239] with (2.7) and (2.11), we have

    L(a,b,K)=c10(a,b,K)=2max{8+64a3(1a),b+8ab3(1a)}×[1+1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a)]. (2.12)

    This completes the proof.

    The proof of following estimate can be found in [8].

    Lemma 10. ([8]) Suppose that gC(D). Then, |G[g](z)|(1|z|2)||g||4, where ||g||:=supzD|g(z)|.

    Lemma 11. For any t[0,1), we have

    14πarctant1t2π.

    Proof. Let φ1(t)=14πarctant and φ2(t)=1t,t[0,1). Then, φ1(1)=φ2(1)=0 and the function

    φ1(t)φ2(t)=4π11+t2

    is monotonically decreasing on t[0,1). By L' Hospital's rule for monotonicity [2], we see that the function φ1/φ2 is monotonically decreasing on t[0,1). Hence, by L'Hospital's rule, we get

    14πarctant1tlimt114πarctant1t=limt14π11+t2=2π.

    This finishes the proof.

    The following result was proved by Ruscheweyh [41].

    Lemma 12. ([41]) Suppose that f is a holomorphic self-mapping of D. Then,

    |f(n)(z)|1|f(z)|2n!(1+|z|)n1(1|z|2)n. (2.13)

    In what follows, we generalize Lemma 12 to the following form.

    Proposition 1. For any q[1,2], suppose that f is a holomorphic self-mapping from the unit disk D into itself; then, we have

    |fn(z)|n!(1|z|)n11|f(z)|q(1+|z||f(0)|)q(|z|+|f(0)|)q1|f(0)|2(1+|z||f(0)|)2q,zD, (2.14)

    where n{1,2,}.

    Proof. Let

    φ(t)=1t21tq,t[0,1).

    Since (1t2)(1tq)=2qt2q is monotonically increasing on [0,1), the function φ is also monotonically increasing on [0,1). By Lindelöf's inequality, we have that |f(z)||z|+|f(0)|1+|z||f(0)|,zD. Therefore,

    1|f(z)|21|f(z)|q1(|z|+|f(0)|1+|z||f(0)|)21(|z|+|f(0)|1+|z||f(0)|)q=1|z|2(1+|z||f(0)|)q(|z|+|f(0)|)q1|f(0)|2(1+|z||f(0)|)2q.

    Combining this with Lemma 12, we get

    |f(n)(z)|n!(1|z|)n11|f(z)|21|z|2n!(1|z|)n11|f(z)|q(1+|z||f(0)|)q(|z|+|f(0)|)q1|f(0)|2(1+|z||f(0)|)2q. (2.15)

    This has finished the proof.

    Remark 2. If q=2, then Proposition 1 becomes Lemma 12.

    Proof. By using Proposition 1 for n=1 and f(0)=0, we get

    |f(z)|1|f(z)|q1|z|q. (3.1)

    Namely,

    |f(z)|1|f(z)|q11|z|q.

    According to the definition of q-distance, we have that dq(f(z1),f(z2))dq(z1,z2). This finishes the proof.

    Remark 3. Considering the case for n=1 and f(0)=0 in Proposition 1, we get the following Schwarz-Pick lemma: Suppose that f is a holomorphic self-mapping from unit disk D into itself with f(0)=0; then, we have

    |f(z)|1|f(z)|q1|z|q,zD. (3.2)

    From the proof of Proposition 1, we see that the inequality (3.2) is sharp. The equation holds if and only if |f(z)|=|z|, which is subject to the condition if and only if f(z)=eiθz,θR is true.

    Proof. According to Lemma 9, we see that there exists a constant L(a,b,K), such that the inequality

    |Df(z)|L(a,b,K)

    holds for any zD, where

    L(a,b,K)=2max{8+64a3(1a),b+8ab3(1a)}×[1+1a12a3Kπ3+3aπ2(12a)8K2π2+2K3+3b2(12a)]. (4.1)

    Now, let

    Δf(z)=h(z),zD, (4.2)

    where h(z)=l(z)(|Df(z)|2+1) and ||l||max{a,b}. For example, we can define l as follows:

    l(z):=Δf(z)(|Df(z)|2+1)1,zD.

    Since fC2(D), we see that hC(D). Hence, by Lemma 9, we get

    ||h||max{a,b}(L2(a,b,K)+1).

    By virtue of the solution of Poisson's equation, we have

    f(z)=P[k](z)G[h](z), (4.3)

    where f|T:=k. Since P[k] is a harmonic self-mapping of D, we get the following by Lemma 8

    |P[k](z)1|z|21+|z|2P[k](0)|4πarctan|z|. (4.4)

    Combining Lemmas 6, 8, 10 and 11, we get

    1|f(z)|q1|z|q=1|f(z)|q1|f(z)|1|z|1|z|q1|f(z)|1|z|cq1|f(z)|1|z|=cq1|P[k](z)1|z|21+|z|2P[k](0)+1|z|21+|z|2P[k](0)G[h](z)|1|z|cq1|P[k](z)1|z|21+|z|2P[k](0)|1|z|21+|z|2|P[k](0)||G[h](z)|1|z|cq14πarctan|z|1|z|21+|z|2|P[k](0)|(1|z|2)||h||41|z|=cq(14πarctan|z|1|z|1+|z|1+|z|2|G[h](0)|(1+|z|)||h||4)cq(2π12||h||12||h||)=cq(2π||h||)cq(2πmax{a,b}(L2(a,b,K)+1))>0. (4.5)

    This completes the proof.

    Proof. Consider the function

    f(z)=(3|z|2)z2,zD,

    then, |f(z)|<1,f(eiθ)=eiθ,fz(z)=32|z|22,f¯z(z)=z22,Δf(z)=4fz¯z(z)=4z and |fz(z)|+|f¯z(z)|=3|z|22. Therefore, f is a self-mapping of D satisfying that |Δf(z)|4(Df(z))2 and f(0)=0. But Re(fx(1))=Re(fz(1)+f¯z(1))=0.

    Proof. Set Δf(z)=h(z),zD, where h(z)=l(z)(|Df(z)|2+1) and ||l||max{a,b}. Since hC(D), we get, by Lemma 9 that ||h||max{a,b}(L2(a,b,K)+1). Then,

    f(z)=P[k](z)G[h](z), (6.1)

    where f|T:=k. Next, the method is similar to the one given in [50]. But, here, we do not demand that hC(¯D), as we only need hC(D). By virtue of (6.1) and Lemmas 8 and 10, we have

    |f(z)|=|P[k](z)1|z|21+|z|2P[k](0)+1|z|21+|z|2P[k](0)G[h](z)||P[k](z)1|z|21+|z|2P[k](0)|+1|z|21+|z|2|P[k](0)|+|G[h](z)|4πarctan|z|+1|z|21+|z|2|G[h](0)|+|G[h](z)|4πarctan|z|+1|z|21+|z|2||h||4+(1|z|2)||h||4:=M(z). (6.2)

    Since f is differential at z=1 and satisfies that f(1)=1, we get

    f(z)=1+fz(z)(z1)+f¯z(¯z1)+o(|z1|).

    Since |f(z)|2M2(z),

    2Re(fz(1z)+f¯z(1¯z))1M2(z)+o(|z1|).

    By letting z=r(0,1) and r1, we obtain

    2Re(fz(1)+f¯z(1))limr11M2(z)1r=2limr1(4π11+r2r||h||(1+r2)2r||h||2)=2(2π3||h||4)2(2π3max{a,b}(L2(a,b,K)+1)4)>0. (6.3)

    This finishes the proof.

    By establishing the general Schwarz-Pick lemma, which is a generalization of the result of Ruscheweyh [41] and sharp when n=1 and f(0)=0, we have obtained a Schwarz-Pick-type lemma for the holomorphic self-mapping of the unit disk with respect to the q-distance. We also have established the general Schwarz-Pick lemma for self-mappings of the unit disk satisfying the Poisson differential inequality. This obtained result improved the one in [47] by giving the explicit constant L(a,b,K), and it showed that the result is still valid when we consider all of the cases for q>0. As an application, it has been proven that this mapping is Lipschitz continuous with respect to the q-distance under certain conditions; Moreover, the corresponding explicit Lipschitz constant has been given. Next, we investigated the boundary Schwarz lemma for self-mappings of the unit disk satisfying the Poisson differential inequality. We found that those types of mappings are generally not associated with the boundary Schwarz lemma. With some additional conditions, we have established a boundary Schwarz lemma for self-mappings of the unit disk satisfying the Poisson differential inequality.

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

    The authors would like to express their sincere thanks to the referees for their great efforts to improve this paper. This work was supported by the Guangdong Basic and Applied Basic Research Foundation (No. 2022A1515110967; No. 2023A1515011809) and Research Foundation of Shenzhen Polytechnic University (No. 6023312032K).

    The authors declare that they have no competing interests.



    [1] C. D. Maranas, I. P. Androulakis, C. A. Floudas, A. J. Bergerb, J. M. Mulvey, Solving long-term financial planning problems via global optimization, J. Econ. Dyn. Contr., 21 (1997), 1405–1425. https://doi.org/10.1016/S0165-1889(97)00032-8 doi: 10.1016/S0165-1889(97)00032-8
    [2] H. W. Jiao, J. Q. Ma, P. P. Shen, Y. J. Qiu, Effective algorithm and computational complexity for solving sum of linear ratios problem, J. Ind. Manag. Optim., 2022. http://dx.doi.org/10.3934/jimo.2022135 doi: 10.3934/jimo.2022135
    [3] H. W. Jiao, S. Y. Liu, An efficient algorithm for quadratic Sum-of-Ratios fractional programs problem, Numer. Func. Anal. Opt., 38 (2017), 1426–1445. https://doi.org/10.1080/01630563.2017.1327869 doi: 10.1080/01630563.2017.1327869
    [4] J. E. Falk, S. W. Palocsay, Optimizing the sum of linear fractional functions, recent advances in global optimization, New Jersey: Princeton University Press, 1992.
    [5] H. W. Jiao, J. Q. Ma, An efficient algorithm and complexity result for solving the sum of general ratios problem, Chaos Soliton. Fract., 164 (2022), 112701. https://doi.org/10.1016/j.chaos.2022.112701 doi: 10.1016/j.chaos.2022.112701
    [6] C. Bajona-Xandri, J. E. Martinez-Legaz, Lower subdifferentiability in minimax fractional programming, Optimization, 45 (1999), 1–12. https://doi.org/10.1080/02331939908844423 doi: 10.1080/02331939908844423
    [7] F. Ding, Two-stage least squares based iterative estimation algorithm for CARARMA system modeling, Appl. Math.Model., 37 (2013), 4798–4808. https://doi.org/10.1016/j.apm.2012.10.014 doi: 10.1016/j.apm.2012.10.014
    [8] F. Ding, Decomposition based fast least squares algorithm for output error systems, Signal Process., 93 (2013), 1235–1242. https://doi.org/10.1016/j.sigpro.2012.12.013 doi: 10.1016/j.sigpro.2012.12.013
    [9] H. W. Jiao, S. Y. Liu, Range division and compression algorithm for quadratically constrained sum of quadratic ratios, Comp. Appl. Math., 36 (2017), 225–247. https://doi.org/10.1007/s40314-015-0224-5 doi: 10.1007/s40314-015-0224-5
    [10] H. W. Jiao, S. Y. Liu, W. J. Wang, Solving generalized polynomial problem by using new affine relaxed technique, Int. J. Comput. Math., 99 (2022), 309–331. https://doi.org/10.1080/00207160.2021.1909727 doi: 10.1080/00207160.2021.1909727
    [11] H. W. Jiao, W. J. Wang, R. J. Chen, Y. L. Shang, J. B. Yin, An efficient outer space algorithm for generalized linear multiplicative programming problem, IEEE Access, 8 (2020), 80629–80637. https://doi.org/10.1109/ACCESS.2020.2990677 doi: 10.1109/ACCESS.2020.2990677
    [12] H. W. Jiao, Y. L. Shang, R. J. Chen, A potential practical algorithm for minimizing the sum of affine fractional functions, Optimization, 2022. https://doi.org/10.1080/02331934.2022.2032051 doi: 10.1080/02331934.2022.2032051
    [13] Y. Y. Ding, Y. H. Xiao, J. W. Li, A class of conjugate gradient methods for convex constrained monotone equations, Optimization, 66 (2017), 2309–2328. https://doi.org/10.1080/02331934.2017.1372438 doi: 10.1080/02331934.2017.1372438
    [14] Y. H. Xiao, L. Chen, D. H. Li, A generalized alternating direction method of multipliers with semi-proximal terms for convex composite conic programming, Math. Program. Comput., 10 (2018), 533–555. https://doi.org/10.1007/s12532-018-0134-9 doi: 10.1007/s12532-018-0134-9
    [15] H. W. Jiao, W. J. Wang, J. B. Yin, Y. L. Shang, Image space branch-reduction-bound algorithm for globally minimizing a class of multiplicative problems, Rairo-Oper. Res., 56 (2022), 1533–1552. https://doi.org/10.1051/ro/2022061 doi: 10.1051/ro/2022061
    [16] H. W. Jiao, Y. L. Shang, Two-level linear relaxation method for generalized linear fractional programming, J. Oper. Res. Soc., 2022. https://doi.org/10.1007/s40305-021-00375-4 doi: 10.1007/s40305-021-00375-4
    [17] H. W. Jiao, W. J. Wang, Y. L. Shang, Outer space branch-reduction-bound algorithm for solving generalized affine multiplicative problem, J. Comput. Appl. Math., 419 (2023), 114784. https://doi.org/10.1016/j.cam.2022.114784 doi: 10.1016/j.cam.2022.114784
    [18] A. I. Barros, J. B. G. Frenk, Generalized fractional programming and cutting plane algorithms, J. Optim. Theory Appl., 87 (1995), 103–120. https://doi.org/10.1007/BF02192043 doi: 10.1007/BF02192043
    [19] Q. G. Feng, H. P. Mao, H. W. Jiao, A feasible method for a class of mathematical problems in manufacturing system, Key Eng. Mater., 460 (2011), 806–809. https://doi.org/10.4028/www.scientific.net/KEM.460-461.806 doi: 10.4028/www.scientific.net/KEM.460-461.806
    [20] Q. G. Feng, H. W. Jiao, H. P. Mao, Y. Q. Chen, A deterministic algorithm for min-max and max-min linear fractional programming problems, Int. J. Comput. Int. Sys., 4 (2011), 134–141. http://dx.doi.org/10.1080/18756891.2011.9727770 doi: 10.1080/18756891.2011.9727770
    [21] R.W. Freund, F. Jarre, An interior-point method for fractional programs with convex constraints, Math. Program., 67 (1994), 407–440. https://doi.org/10.1007/BF01582229 doi: 10.1007/BF01582229
    [22] Y. Benadada, J. A. Fedand, Partial linearization for generalized fractional programming, Zeitschrift Für Oper. Res., 32 (1988), 101–106. https://doi.org/10.1007/BF01919185 doi: 10.1007/BF01919185
    [23] N. T. H. Phuong, H. Tuy, A unified monotonic approach to generalized linear fractional programming, J. Global Optim., 26 (2003), 229–259. https://doi.org/10.1023/A:1023274721632 doi: 10.1023/A:1023274721632
    [24] A. Roubi, Method of centers for generalized fractional programming, J. Optim. Theory Appl., 107 (2000), 123–143. https://doi.org/10.1023/A:1004660917684 doi: 10.1023/A:1004660917684
    [25] M. Gugat, Prox-regularization methods for generalized fractional programming, J. Optim. Theory Appl., 99 (1998), 691–722. https://doi.org/10.1023/A:1021759318653 doi: 10.1023/A:1021759318653
    [26] A. Ghazi, A. Roubi, A DC approach for minimax fractional optimization programs with ratios of convex functions, Optim. Methods Softw., 2020. https://doi.org/10.1080/10556788.2020.1818234 doi: 10.1080/10556788.2020.1818234
    [27] H. Boualam, A. Roubi, Dual algorithms based on the proximal bundle method for solving convex minimax fractional programs, J. Ind. Manag. Optim., 15 (2019), 1897–1920. https://doi.org/10.3934/jimo.2018128 doi: 10.3934/jimo.2018128
    [28] H. W. Jiao, J. Q. Ma, Y. Shang, Image space branch-and-bound algorithm for globally solving minimax linear fractional programming problem, Pac. J. Optim., 18 (2022), 195–212.
    [29] M. E. Haffari, A. Roubi, Prox-dual regularization algorithm for generalized fractional programs, J. Ind. Manag. Optim., 13 (2017). https://doi.org/10.3934/jimo.2017028 doi: 10.3934/jimo.2017028
    [30] H. W. Jiao, B. B. Li, Solving min-max linear fractional programs based on image space branch-and-bound scheme, Chaos Soliton. Fract., 164 (2022), 112682. https://doi.org/10.1016/j.chaos.2022.112682 doi: 10.1016/j.chaos.2022.112682
    [31] I. Ahmad, Z. Husain, Duality in nondifferentiable minimax fractional programming with generalized convexity, Appl. Math. Comput., 176 (2006), 545–551. https://doi.org/10.1016/j.amc.2005.10.002 doi: 10.1016/j.amc.2005.10.002
    [32] W. E. Schmitendorf, Necessary conditions and sufficient conditions for static minimax problems, J. Math. Anal. Appl., 57 (1977), 683–693. https://doi.org/10.1016/0022-247X(77)90255-4 doi: 10.1016/0022-247X(77)90255-4
    [33] S. Tanimoto, Duality for a class of nondifferentiable mathematical programming problems, J. Math. Anal. Appl., 79 (1981), 286–294. https://doi.org/10.1016/0022-247X(81)90025-1 doi: 10.1016/0022-247X(81)90025-1
    [34] S. R. Yadav, R. N. Mukherjee, Duality for fractional minimax programming problems, ANZIAM J., 31 (1990), 484–492. https://doi.org/10.1017/S0334270000006809 doi: 10.1017/S0334270000006809
    [35] V. Jeyakumar, G. Y. Li, S. Srisatkunarajah, Strong duality for robust minimax fractional programming problems, Eur. J. Oper. Res., 228 (2013), 331–336. https://doi.org/10.1016/j.ejor.2013.02.015 doi: 10.1016/j.ejor.2013.02.015
    [36] H. C. Lai, J. C. Liu, K. Tanaka, Duality without a constraint qualification for minimax fractional programming, J. Math. Anal. Appl., 101 (1999), 109–125. https://doi.org/10.1023/A:1021771011210 doi: 10.1023/A:1021771011210
    [37] I. M. Stancu-Minasian, A ninth bibliography of fractional programming, Optimization, 68 (2019) 2125–2169. https://doi.org/10.1080/02331934.2019.1632250 doi: 10.1080/02331934.2019.1632250
    [38] I. M. Stancu-Minasian, A eighth bibliography of fractional programming, Optimization, 66 (2017) 439–470. https://doi.org/10.1080/02331934.2016.1276179 doi: 10.1080/02331934.2016.1276179
    [39] C. F. Wang, Y. Jiang, P. P. Shen, A new branch-and-bound algorithm for solving minimax linear fractional programming, J. Math., 38 (2018), 113–123.
    [40] H. W. Jiao, S. Y. Liu, A new linearization technique for minimax linear fractional programming, Int. J. Comput. Math., 91 (2014), 1730–1743. https://doi.org/10.1080/00207160.2013.860449 doi: 10.1080/00207160.2013.860449
  • Reader Comments
  • © 2023 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(1609) PDF downloads(59) Cited by(4)

Figures and Tables

Tables(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog