Processing math: 100%
Research article

Sharp bounds for Gauss Lemniscate functions and Lemniscatic means

  • Received: 21 December 2020 Accepted: 25 April 2021 Published: 07 May 2021
  • MSC : 26D07, 26E60, 33E05

  • For a,b>0 with ab, the Gauss lemniscate mean LM(a,b) is defined by

    LM(a,b)={a2b2[arcsl(41b2/a2)]2, a>b,b2a2[arcslh(4b2/a21)]2, a<b,

    where arcsl(x)=x0dt1t4 (|x|<1) and arcslh(x)=x0dt1+t4 (xR) is the arc lemniscate sine and hyperbolic arc lemniscate sine functions respectively. In this paper, we mainly establish sharp two-parameter bounds for four symmetric and homogeneous means derived from LM(a,b), LMGA(a,b)=LM(G(a,b),A(a,b)), LMAG(a,b)=LM(A(a,b),G(a,b)), LMAQ(a,b)=LM(A(a,b),G(a,b)) and LMQA(a,b)=LM(A(a,b),G(a,b)). The obtained results lead to several asymptotical inequalities for Lemniscate functions. Here A(a,b)=(a+b)/2, G(a,b)=ab and Q(a,b)=(a2+b2)/2 are the classical arithmetic, geometric, and quadratic means.

    Citation: Wei-Mao Qian, Miao-Kun Wang. Sharp bounds for Gauss Lemniscate functions and Lemniscatic means[J]. AIMS Mathematics, 2021, 6(7): 7479-7493. doi: 10.3934/math.2021437

    Related Papers:

    [1] Ling Zhu . Sharp bounds for Heinz mean by Heron mean and other means. AIMS Mathematics, 2020, 5(1): 723-731. doi: 10.3934/math.2020049
    [2] Wei-Mao Qian, Tie-Hong Zhao, Yu-Pei Lv . Refinements of bounds for the arithmetic mean by new Seiffert-like means. AIMS Mathematics, 2021, 6(8): 9036-9047. doi: 10.3934/math.2021524
    [3] Ling Zhu, Branko Malešević . Optimal bounds for the sine and hyperbolic tangent means by arithmetic and centroidal means in exponential type. AIMS Mathematics, 2021, 6(12): 13024-13040. doi: 10.3934/math.2021753
    [4] Muhammad Ghaffar Khan, Sheza.M. El-Deeb, Daniel Breaz, Wali Khan Mashwani, Bakhtiar Ahmad . Sufficiency criteria for a class of convex functions connected with tangent function. AIMS Mathematics, 2024, 9(7): 18608-18624. doi: 10.3934/math.2024906
    [5] Khuram Ali Khan, Shaista Ayaz, İmdat İşcan, Nehad Ali Shah, Wajaree Weera . Applications of Hölder-İşcan inequality for n-times differentiable (s,m)-convex functions. AIMS Mathematics, 2023, 8(1): 1620-1635. doi: 10.3934/math.2023082
    [6] Bakhtiar Ahmad, Muhammad Ghaffar Khan, Basem Aref Frasin, Mohamed Kamal Aouf, Thabet Abdeljawad, Wali Khan Mashwani, Muhammad Arif . On q-analogue of meromorphic multivalent functions in lemniscate of Bernoulli domain. AIMS Mathematics, 2021, 6(4): 3037-3052. doi: 10.3934/math.2021185
    [7] Zareen A. Khan, Waqar Afzal, Mujahid Abbas, Jongsuk Ro, Najla M. Aloraini . A novel fractional approach to finding the upper bounds of Simpson and Hermite-Hadamard-type inequalities in tensorial Hilbert spaces by using differentiable convex mappings. AIMS Mathematics, 2024, 9(12): 35151-35180. doi: 10.3934/math.20241671
    [8] Jiankang Wang, Zhefeng Xu, Minmin Jia . On the generalized Cochrane sum with Dirichlet characters. AIMS Mathematics, 2023, 8(12): 30182-30193. doi: 10.3934/math.20231542
    [9] Thiyam Thadoi Devi, Khundrakpam Binod Mangang, Sonika Akoijam, Lalhmangaihzuala, Phinao Ramwungzan, Jay Prakash Singh . Mean chain transitivity and almost mean shadowing property of iterated function systems. AIMS Mathematics, 2024, 9(8): 20811-20825. doi: 10.3934/math.20241012
    [10] Wenjia Guo, Yuankui Ma, Tianping Zhang . New identities involving Hardy sums S3(h,k) and general Kloosterman sums. AIMS Mathematics, 2021, 6(2): 1596-1606. doi: 10.3934/math.2021095
  • For a,b>0 with ab, the Gauss lemniscate mean LM(a,b) is defined by

    LM(a,b)={a2b2[arcsl(41b2/a2)]2, a>b,b2a2[arcslh(4b2/a21)]2, a<b,

    where arcsl(x)=x0dt1t4 (|x|<1) and arcslh(x)=x0dt1+t4 (xR) is the arc lemniscate sine and hyperbolic arc lemniscate sine functions respectively. In this paper, we mainly establish sharp two-parameter bounds for four symmetric and homogeneous means derived from LM(a,b), LMGA(a,b)=LM(G(a,b),A(a,b)), LMAG(a,b)=LM(A(a,b),G(a,b)), LMAQ(a,b)=LM(A(a,b),G(a,b)) and LMQA(a,b)=LM(A(a,b),G(a,b)). The obtained results lead to several asymptotical inequalities for Lemniscate functions. Here A(a,b)=(a+b)/2, G(a,b)=ab and Q(a,b)=(a2+b2)/2 are the classical arithmetic, geometric, and quadratic means.



    Gauss's arc lemniscate sine and the hyperbolic arc lemniscate sine functions are defined by as follows:

    arcsl(x)=x0dt1t4,|x|<1,
    arcslh(x)=x0dt1+t4,xR,

    respectively (cf. [1,p.259] or [2,(2.5) and (2.6)]).

    Another pair of arc lemniscate functions, Gauss's arc lemniscate tangent and the hyperbolic arc lemniscate tangent functions are defined in terms of the arc lemniscate sine and the hyperbolic arc lemniscate sine functions, respectively (cf. [3,(3.5) and (3.6)]):

    arctl(x)=arcsl(x41+x4),xR,
    arctlh(x)=arcslh(x41x4),|x|<1.

    It is not difficult to verify that

    ω=arcsl(1)=12K(1/2)=Γ2(1/4)42π=1.31103,
    τ=arctl(1)=arcsl(1/42)=0.89558

    and

    arctlh(1)=arcslh(+)=2ω=1.85407,

    where

    K(r)=π/20(1r2sin2θ)1/2dθ=π/20dt(1t2)(1r2t2),0<r<1

    is the complete elliptic integral of the first kind [4,5,6,7,8,9,10], and

    Γ(x)=0tx1etdt,Re(x)>0

    is the classical Euler gamma function [11,12,13,14].

    For a,b>0 with ab, the arithmetic mean A(a,b), geometric mean G(a,b), harmonic mean H(a,b), quadratic mean Q(a,b), contra-harmonic mean C(a,b) and the lemniscatic mean LM(a,b) [3,15] of a and b are respectively defined by

    A(a,b)=a+b2, G(a,b)=ab, H(a,b)=2aba+b,
    Q(a,b)=a2+b22, C(a,b)=a2+b2a+b

    and

    LM(a,b)={a2b2[arcsl(41b2/a2)]2, a>b,b2a2[arcslh(4b2/a21)]2, a<b.

    Recall that the lemniscatic mean LM(a,b) of two positive real a and b is an iterative mean, i.e.,

    LM(a,b)=limnan=limnbn,

    where

    a0=a, b0=b,
    an+1=an+bn2, bn+1=an+1an

    for n=0,1,2,. And the lemniscatic mean is non-symmetric and homogeneous of degree one with respect to its variables a and b. Let

    LMGA(a,b)=LM[G(a,b),A(a,b)], LMAG(a,b)=LM[A(a,b),G(a,b)],
    LMAQ(a,b)=LM[A(a,b),Q(a,b)], LMQA(a,b)=LM[Q(a,b),A(a,b)].

    Then one has four symmetric and homogeneous means given by explicitly as follows: (c.f. [3,15])

    LMGA(a,b)=A(a,b)(zarctlhz)2, LMAG(a,b)=A(a,b)(zarcslz)2, (1.1)
    LMAQ(a,b)=A(a,b)(zarcslhz)2, LMQA(a,b)=A(a,b)(zarctlz)2 (1.2)

    with z=|ab|/(a+b). Moreover, [3,(6.10)] showed that the derived means satisfy the following inequality chain:

    H(a,b)<G(a,b)<LMGA(a,b)<LMAG(a,b)<A(a,b)<LMAQ(a,b)<LMQA(a,b)<Q(a,b)<C(a,b) (1.3)

    for all a,b>0 with ab.

    In the recent years, the arc lemniscate functions and lemniscatic means have attracted the attention of many researchers because they are closely related to special functions. In particular, many remarkable properties and inequalities involving them can be found in the literature [16,17,18,19,20,21,22]. For example, Neuman [3] obtained

    Proposition 1.1. ([3,Proposition 6.1]) Inequalities

    A2(a,b)G(a,b)LMGA(a,b)<LMAG(a,b)4,G2(a,b)A(a,b)LMAG(a,b)<LMGA(a,b)4,A2(a,b)Q(a,b)LMQA(a,b)<LMAQ(a,b)4,Q2(a,b)A(a,b)LMAQ(a,b)<LMQA(a,b)4,LMGA(a,b)LMQA(a,b)<A2(a,b),LMAG(a,b)LMAQ(a,b)<A2(a,b)

    hold true.

    Let a,b>0, p[1,), q[1/2,+), t[0,1/2] and

    GAt,p(a,b)=Gp[ta+(1t)b,tb+(1t)a]A1p(a,b) (1.4)

    and

    CAt,q(a,b)=Cq[ta+(1t)b,tb+(1t)a]A1q(a,b). (1.5)

    Then it is easy to check that

    GAt,1(a,b)=G[ta+(1t)b,tb+(1t)a],
    GAt,2(a,b)=H[ta+(1t)b,tb+(1t)a],
    CAt,12(a,b)=Q[ta+(1t)b,tb+(1t)a],
    CAt,1(a,b)=C[ta+(1t)b,tb+(1t)a]

    and the function tGAt,p(a,b) is strictly increasing on the interval [0,1/2] for fixed a,b>0 with ab and p[1,+), while tCAt,q(a,b) is strictly decreasing on [0,1/2] for fixed a,b>0 with ab and q[1/2,+). For more inequalities involving GAt,p(a,b), CAt,q(a,b) and thier applications in special functions, the readers can refer to the literature [8,23,24,25,26,27,28,29,30,31,32,33,34,35].

    Since

    GA0,p(a,b)=A(a,b)[G(a,b)A(a,b)]pG(a,b)<LMGA(a,b)<LMAG(a,b)<A(a,b)=GA12,p(a,b)=CA12,q(a,b)<LMAQ(a,b)<LMQA(a,b)<Q(a,b)A(a,b)[C(a,b)A(a,b)]q=CA0,q(a,b)

    hold for all a,b>0 with ab, then it is nature to ask that: what are the best possible parameters λ1, λ2, λ3, λ4, μ1, μ2, μ3, μ4[0,1/2] such that the double inequalities

    GAλ1,p(a,b)<LMGA(a,b)<GAμ1,p(a,b),GAλ2,p(a,b)<LMAG(a,b)<GAμ2,p(a,b)
    CAλ3,q(a,b)<LMAQ(a,b)<CAμ3,q(a,b),CAλ4,q(a,b)<LMQA(a,b)<CAμ4,q(a,b)

    hold for all p[1,), q[1/2,+) and a,b>0 with ab.

    In order to prove our main results, we need the derivative formulas of the arc lemniscate functions and several lemmas, which we present in this section.

    darcsl(x)dx=(1x4)1/2,darctlh(x)dx=(1x4)3/4, |x|<1,darcslh(x)dx=(1+x4)1/2,darctl(x)dx=(1+x4)3/4, xR.

    Lemma 2.1. ([36,Theorem 1.25]). Let <a<b<, f,g:[a,b]R be continuous on [a,b], and differentiable on (a,b). Let g(x)0 on (a,b). Then, if f(x)/g(x) is increasing (decreasing) on (a,b), so are

    f(x)f(a)g(x)g(a)andf(x)f(b)g(x)g(b).

    If f(x)/g(x) is strictly monotone, then the monotonicity in the conclusion is also strict.

    Lemma 2.2. ([22,Lemma 3.3]). (1) The function x1x4arcsl(x)x is strictly decreasing from (0,1) onto (0,1);

    (2) The function x1+x4arcslh(x)x is strictly increasing from (0,1) onto (1,ω);

    (3) The function x41+x4arctl(x)x is strictly increasing from (0,1) onto (1,21/4τ);

    (4) The function x41x4arctlh(x)x is strictly decreasing from (0,1) onto (0,1).

    Lemma 2.3. For u[0,1], p[1,). Let

    f(u,p;x)=12plog(1ux4)2logx+2log[arctlh(x)], x(0,1). (2.1)

    Then the following statements are true:

    (1) f(u,p;x)>0 for x(0,1) if and only if u3/(5p);

    (2) f(u,p;x)<0 for x(0,1) if and only if u1[1/(4ω4)]1/p.

    Proof. It follows from (2.1) that

    f(u,p;0+)=0, (2.2)
    f(u,p;1)=12plog(1u)+2log(2ω), (2.3)
    df(u,p;x)dx=2x3[x+(p1)(1x4)3/4arctlh(x)](1x4)3/4(1ux4)arctlh(x)[fp(x)u], (2.4)

    where

    fp(x)=x(1x4)3/4arctlh(x)x4[x+(p1)(1x4)3/4arctlh(x)].

    Let ζ1(x)=x/(1x4)3/4arctlh(x) and ζ2(x)=x5/(1x4)3/4+(p1)x4arctlh(x). Then simple computations lead to

    ζ1(0+)=ζ2(0+)=0, fp(x)=ζ1(x)/ζ2(x), (2.5)
    ζ1(x)ζ2(x)=11+13(p+1)(1x4)+43(p1)(1x4)3/2(1x4)1/4arctlh(x)x. (2.6)

    Eq (2.6) and Lemma 2.2(4) show that ζ1(x)/ζ2(x) is strictly increasing on (0,1), so is fp(x) by (2.5) and Lemma 2.1. Moreover,

    fp(0+)=limx0+ζ1(x)ζ2(x)=35p,fp(1)=1. (2.7)

    Following we divide the proof into three cases.

    Case 1 0u3/(5p). Then from (2.4) and (2.7) together with the monotonicity of fp(x) lead to the conclusion that the function xf(u,p;x) is strictly increasing on (0,1). Therefore, f(u,p;x)>0 for all x(0,1) follows from (2.2).

    Case 2 1[1/(4ω4)]1/pu1(Since the function uf(u,p;1) is strictly decreasing on (0,1), and f(3/(5p),p;1)>0, f(1,p;1)=, then the function uf(u,p;1) has a unique zero-point 1[1/(4ω4)]1/p(3/(5p),1) for any fixed p[1,)). Then it follows from (2.3), (2.4) and (2.7) together with the monotonicity of fp(x) that

    f(u,p;1)0,

    and there exists x0(0,1) such that the function xf(u,p;x) is strictly decreasing on (0,x0) and strictly increasing on (x0,1). Therefore, together with (2.2), f(u,p;x)<0 for all x(0,1).

    Case 3 3/(5p)<u<1[1/(4ω4)]1/p. Then it is easy to check that the function xf(u,p;x) has the same piecewise monotone property in Case 2, and while

    f(u,p;1)>0.

    Combining with (2.2), we conclude that there exists x0(0,1) such that f(u,p;x)<0 for x(0,x0) and f(u,p;x)>0 for x(x0,1).

    Lemma 2.4. For v[0,1], p[1,). Let

    g(v,p;x)=12plog(1vx4)2logx+2log[arcsl(x)] (2.8)

    Then the following statements are true:

    (1) g(v,p;x)>0 for x(0,1) if and only if v2/(5p);

    (2) g(v,p;x)<0 for x(0,1) if and only if v1(1/ω4)1/p.

    Proof. It follows from (2.8) that

    g(v,p;0+)=0, (2.9)
    g(v,p;1)=12plog(1v)+2logω, (2.10)
    dg(v,p;x)dx=2x3[x+(p1)1x4arcsl(x)]1x4(1vx4)arcsl(x)[gp(x)v], (2.11)

    where

    gp(x)=x1x4arcsl(x)x4[x+(p1)1x4arcsl(x)].

    Let ξ1(x)=x/1x4arcsl(x) and ξ2(x)=x5/1x4+(p1)x4arcsl(x). Then simple computations lead to

    ξ1(0+)=ξ2(0+)=0, gp(x)=ξ1(x)/ξ2(x), (2.12)
    ξ1(x)ξ2(x)=11+12(p+2)(1x4)+2(p1)(1x4)1x4arcsl(x)x. (2.13)

    Eq (2.13) and Lemma 2.2(1) imply that ξ1(x)/ξ2(x) is strictly increasing on (0,1). Therefore, the conclusion that gp(x) is strictly increasing on (0,1) follows from Lemma 2.1 and (2.12). Moreover,

    gp(0+)=limx0+ξ1(x)ξ2(x)=25p, gp(1)=1. (2.14)

    Following we divide the proof into three cases.

    Case 1 0<v2/(5p). Then (2.11), (2.14) and the monotonicity of the function gp(x) lead to the conclusion that the function xg(v,p;x) is strictly increasing on (0,1). Therefore, g(v,p;x)>0 for all x(0,1) follows from (2.9).

    Case 2 1(1/ω4)1/pv<1 (With the similar argument in the proof of Lemma 2.3, we claim that 1(1/ω4)1/p(2/(5p),1)). Then it follows from (2.10), (2.11) and (2.14) together with the monotonicity of gp(x) that

    g(v,p;1)0,

    and there exists x1(0,1) such that the function xg(v,p;x) is strictly decreasing on (0,x1) and strictly increasing on (x1,1). Therefore, together with (2.9), g(v,p;x)<0 for all x(0,1).

    Case 3 2/(5p)<v<1(1/ω4)1/p. Then the function xg(v,p;x) also first decreases and then increases on (0,1), and from (2.10) we get

    g(v,p;1)>0.

    Combining with (2.9), we conclude that there exists x1(0,1) such that g(v,p;x)<0 for x(0,x1) and g(v,p;x)>0 for x(x1,1).

    Lemma 2.5. For u[0,1], q[1/2,). Let

    F(u,q;x)=qlog(1+ux4)2logx+2log[arcslh(x)], x(0,1). (2.15)

    Then the following statements are true:

    (1) F(u,q;x)>0 for x(0,1) if and only if u1/(5q);

    (2) F(u,q;x)<0 for x(0,1) if and only if u(2/ω)2/q1.

    Proof. It follows from (2.15) that

    F(u,q;0+)=0, (2.16)
    F(u,q;1)=qlog(1+u)+2logωlog2, (2.17)
    dF(u,q;x)dx=2x3[x+(2q1)1+x4arcslh(x)]1+x4(1+ux4)arcslh(x)[uFq(x)], (2.18)

    where

    Fq(x)=1+x4arcslh(x)xx4[x+(2q1)1+x4arcslh(x)].

    Let ζ1(x)=arcslh(x)x/1+x4 and ζ2(x)=x5/1+x4+(2q1)x4arcslh(x). Then simple computations lead to

    ζ1(0+)=ζ2(0+)=0, Fq(x)=ζ1(x)/ζ2(x), (2.19)
    ζ1(x)ζ2(x)=11+(q+1)(1+x4)+2(2q1)(1+x4)1+x4arcslh(x)x. (2.20)

    Eq (2.20) and Lemma 2.2(2) show that ζ1(x)/ζ2(x) is strictly decreasing on (0,1), so is Fq(x) by (2.19) and Lemma 2.1. Moreover,

    Fq(0+)=limx0+ζ1(x)ζ2(x)=15q,Fq(1)=ω11+(2q1)ω. (2.21)

    Following we divide the proof into three cases.

    Case 1 1/(5q)u1. Then (2.18) and (2.21) together with the monotonicity of Fq(x) imply that the function xF(u,q;x) is strictly increasing on (0,1). Therefore, F(u,q;x)>0 for all x(0,1) follows from (2.16).

    Case 2 0u(ω1)/[1+(2q1)ω]. Then the function xF(u,q;x) is strictly decreasing on (0,1), and consequently F(u,q;x)<0 for all x(0,1).

    Case 3 (ω1)/[1+(2q1)ω]<u<1/(5q). Then it follows from (2.18) and (2.21) together with the monotonicity of Fq(x) that and there exists x2(0,1) such that the function xF(u,q;x) is strictly decreasing on (0,x2) and strictly increasing on (x2,1).

    Since the function uF(u,q;1) is strictly increasing on [0,1], and F(1/(5q),q;1)>0, F((ω1)/[1+(2q1)ω],q;1)<0, then the function uF(u,q;1) has a unique zero-point (2/ω)2/q1((ω1)/[1+(2q1)ω],1/(5q)) for any fixed q[1/2,+). Finally we investigate two subcases (2/ω)2/q1<u<1/(5q) and (ω1)/[1+(2q1)ω]<u(2/ω)2/q1.

    Subcase 3.1 (2/ω)2/q1<u<1/(5q). Then

    F(u,q;1)>0

    and thereby there exists x2(0,1) such that F(u,q;x)<0 for x(0,x2) and F(u,q;x)>0 for x(x2,1).

    Subcase 3.2 (ω1)/[1+(2q1)ω]<u(2/ω)2/q1. Then

    F(u,q;1)0

    and therefore F(u,q;x)<0 for x(0,1).

    Lemma 2.6. For v[0,1], q[1/2,). Let

    G(v,q;x)=qlog(1+vx4)2logx+2log[arctl(x)], x(0,1). (2.22)

    Then the following statements are true:

    (1) G(v,q;x)>0 for x(0,1) if and only if v3/(10q);

    (2)G(v,q;x)<0 for x(0,1) if and only if u(1/τ)2/q1.

    Proof. It follows from (2.22) that

    G(v,q;0+)=0, (2.23)
    G(v,q;1)=qlog(1+v)+2logτ, (2.24)
    dG(v,q;x)dx=2x3[x+(2q1)(1+x4)3/4arctl(x)](1+x4)3/4(1+vx4)arctl(x)[vGq(x)], (2.25)

    where

    Gq(x)=(1+x4)3/4arctl(x)xx4[x+(2q1)(1+x4)3/4arctl(x)].

    Let ξ1(x)=arctl(x)x/(1+x4)3/4 and ξ2(x)=x5/(1+x4)3/4+(2q1)x4arctl(x). Then simple computations lead to

    ξ1(0+)=ξ2(0+)=0, Gq(x)=ξ1(x)/ξ2(x), (2.26)
    ξ1(x)ξ2(x)=11+13(2q+1)(1+x4)+43(2q1)(1+x4)3/241+x4arctl(x)x. (2.27)

    Eq (2.27) and Lemma 2.2 (3) show that ξ1(x)/ξ2(x) is strictly decreasing on (0,1), so is Gq(x) by (2.26) and Lemma 2.1. Moreover,

    Gq(0+)=limx0+ξ1(x)ξ2(x)=310q,Gq(1)=23/4τ11+23/4(2q1)τ. (2.28)

    Following we divide the proof into three cases.

    Case 1 3/(10q)v1. Then (2.25) and (2.28) together with the monotonicity of Gq(x) imply that the function xG(v,q;x) is strictly increasing on (0,1). Therefore, G(v,q;x)>0 for all x(0,1) follows from (2.23).

    Case 2 0v(23/4τ1)/[1+23/4(2q1)τ]. Then the function xG(v,q;x) is strictly decreasing on (0,1), and consequently G(v,q;x)<0 for all x(0,1).

    Case 3 (23/4τ1)/[1+23/4(2q1)τ]<v<3/(10q). Then it follows from (2.25) and (2.28) together with the monotonicity of Gq(x) that and there exists x3(0,1) such that the function xG(v,q;x) is strictly decreasing on (0,x3) and strictly increasing on (x3,1).

    With the similar argument in the proof of Lemma 2.5, we also obtain that (1/τ)2/p1((23/4τ1)/[1+23/4(2q1)τ],3/(10q)). We divide the proof into two Subcases.

    Subcase 3.1 (1/τ)2/p1<v<3/(10q). Then

    G(v,q;1)>0

    and thereby there exists x3(0,1) such that G(v,q;x)<0 for x(0,x3) and G(v,q;x)>0 for x(x3,1).

    Subcase 3.2 (23/4τ1)/[1+23/4(2q1)τ]<v(1/τ)2/p1. Then

    G(v,q;1)0

    and therefore G(v,q;x)<0 for x(0,1).

    Theorem 3.1. Let p[1,) and λ1,μ1[0,1/2]. Then the double inequalities

    GAλ1,p(a,b)<LMGA(a,b)<GAμ1,p(a,b)

    hold for all a,b>0 with ab if and only if λ11/21[1/(4ω4)]1/p/2 and μ11/215p/(10p).

    Proof. Let t[0,1/2], since GAt,p(a,b) and LMGA(a,b) are symmetric and homogeneous of degree one, without loss of generality, we assume that a>b>0. Let x=(ab)/(a+b)(0,1). Then from (1.1) and (1.4) we get

    log[GAt,p(a,b)LMGA(a,b)]=log[GAt,p(a,b)A(a,b)]log[LMGA(a,b)A(a,b)]=12plog[1(12t)2x4]2logx+2log[arctlh(x)]=f((12t)2,p;x), (3.1)

    where f(,p;x) is defined in Lemma 2.3.

    Therefore, Theorem 3.1 follows easily from Lemma 2.3 and (3.1).

    Theorem 3.2. Let p[1,) and λ2,μ2[0,1/2]. Then the double inequalities

    GAλ2,p(a,b)<LMAG(a,b)<GAμ2,p(a,b)

    hold for all a,b>0 with ab if and only if λ21/21(1/ω4)1/p/2 and μ21/210p/(10p).

    Proof. Let t[0,1/2]. Without loss of generality, we suppose that x=(ab)/(a+b)(0,1). Then from (1.1) and (1.4) we get

    log[GAt,p(a,b)LMAG(a,b)]=log[GAt,p(a,b)A(a,b)]log[LMAG(a,b)A(a,b)]=12plog[1(12t)2x4]2log(x)+2log[arcsl(x)]=g((12t)2,p;x), (3.2)

    where g(,p;x) is defined in Lemma 2.4.

    Therefore, Theorem 3.2 follows easily from Lemma 2.4 and (3.2).

    Theorem 3.3. Let q[1/2,) and λ3,μ3[0,1/2]. Then the double inequalities

    CAλ3,q(a,b)<LMAQ(a,b)<CAμ3,q(a,b)

    hold for all a,b>0 with ab if and only if λ31/2(2/ω)2/q1/2 and μ31/25q/(10q).

    Proof. Let t[0,1/2], since CAt,q(a,b) and LMGA(a,b) are symmetric and homogeneous of degree one, without loss of generality, we assume that a>b>0. Let x=(ab)/(a+b)(0,1). Then from (1.2) and (1.5) we get

    log[CAt,q(a,b)LMAQ(a,b)]=log[CAt,q(a,b)A(a,b)]log[LMAQ(a,b)A(a,b)]=qlog[1+(12t)2x4]2logx+2log[arcslh(x)]=F((12t)2,q;x), (3.3)

    where F(,q;x) is defined in Lemma 2.5.

    Therefore, Theorem 3.3 follows easily from Lemma 2.5 and (3.3).

    Theorem 3.4. Let q[1/2,) and λ4,μ4[0,1/2]. Then the double inequalities

    CAλ4,q(a,b)<LMQA(a,b)<CAμ4,q(a,b)

    hold for all a,b>0 with ab if and only if λ41/2(1/τ)2/q1/2 and μ41/230q/(20q).

    Proof. Let t[0,1/2]. Without loss of generality, we suppose that x=(ab)/(a+b)(0,1). Then from (1.2) and (1.5) we get

    log[CAt,q(a,b)LMQA(a,b)]=log[CAt,q(a,b)A(a,b)]log[LMQA(a,b)A(a,b)]=qlog[1+(12t)2x4]2log(x)+2log[arctl(x)]=G((12t)2,q;x), (3.4)

    where G(,q;x) is defined in Lemma 2.6.

    Therefore, Theorem 3.4 follows easily from Lemma 2.6 and (3.4).

    Corollary 3.5. Let α1, α2, α3, α4, β1, β2, β3, β4[0,1/2]. Then the double inequalities

    H[α1a+(1α1)b,α1b+(1α1)a]<LMGA(a,b)<H[β1a+(1β1)b,β1b+(1β1)a],
    G[α2a+(1α2)b,α2b+(1α2)a]<LMGA(a,b)<G[β2a+(1β2)b,β2b+(1β2)a],
    H[α3a+(1α3)b,α3b+(1α3)a]<LMAG(a,b)<H[β3a+(1β3)b,β3b+(1β3)a],
    G[α4a+(1α4)b,α4b+(1α4)a]<LMAG(a,b)<G[β4a+(1β4)b,β4b+(1β4)a]

    hold for all a,b>0 with ab with the best possible parameters

    α1=1211/(2ω2)2,β1=123020,α2=1211/(4ω4)2,β2=121510,α3=1211/ω22,β3=12510,α4=1211/ω42,β4=121010.

    Corollary 3.6. Let α1, α2, α3, α4, β1, β2, β3, β4[0,1/2]. Then the double inequalities

    Q[α1a+(1α1)b,α1b+(1α1)a]<LMAQ(a,b)<Q[β1a+(1β1)b,β1b+(1β1)a],
    C[α2a+(1α2)b,α2b+(1α2)a]<LMAQ(a,b)<C[β2a+(1β2)b,β2b+(1β2)a],
    Q[α3a+(1α3)b,α3b+(1α3)a]<LMQA(a,b)<Q[β3a+(1β3)b,β3b+(1β3)a],
    C[α4a+(1α4)b,α4b+(1α4)a]<LMQA(a,b)<C[β4a+(1β4)b,β4b+(1β4)a]

    hold for all a,b>0 with ab with the best possible parameters

    α1=124/ω412,β1=121010,α2=122/ω212,β2=12510,α3=121/τ412,β3=121510,α4=121/τ212,β4=123020.

    Theorem 4.1. The double inequalities

    (135px4)p/4<arctlh(x)x<[1(122/pω4/p)x4]p/4, (4.1)
    (125px4)p/4<arcsl(x)x<[1(1ω4/p)x4]p/4 (4.2)

    hold for all p[1,) and x(0,1).

    Proof. Substituting λ1=1/21[1/(4ω4)]1/p/2, μ1=1/215p/(10p), and λ2=1/21(1/ω4)1/p/2, μ2=1/210p/(10p) into Theorems 3.1 and 3.2, we obtain (4.1) and (4.2) immediately.

    Theorem 4.2. The double inequalities

    (1+15qx4)q/2<arcslh(x)x<[1(121/qω2/q)x4]q/2, (4.3)
    (1+310qx4)q/2<arctl(x)x<[1(1τ2/q)x4]q/2 (4.4)

    hold for all q[1/2,) and x(0,1).

    Proof. Substituting λ3=1/2(2/ω)2/q1/2, μ3=1/25q/(10q) and λ4=1/2 (1/τ)2/q1/2, μ4=1/230q/(20q) into Theorems 3.3 and 3.4, we obtain (4.3) and (4.4) immediately.

    Lemma 4.3. (1) The function tlog(1rx4t)t is strictly decreasing on (0,1) for any fixed x(0,1) and r(1,0)(0,1);

    (2) The function tlog[1(1rt)x4]t is strictly increasing on (0,1) for any fixed x(0,1) and r(0,+).

    Proof. For part (1), let ϕ1(t)=log(1rx4t), ϕ2(t)=t and ϕ(t)=ϕ1(t)/ϕ2(t). Then ϕ1(0)=ϕ2(0)=0,

    ϕ1(t)ϕ2(t)=rx41(rx4)t. (4.5)

    It is apparent from (4.5) to see that ϕ1(t)/ϕ2(t) is strictly decreasing on (0,1). By application of Lemma 2.1, part (1) is clear.

    For part (2), let φ1(t)=log[1(1rt)x4], φ2(t)=t and φ(t)=φ1(t)/φ2(t). Then by simple computation we get φ1(0)=φ2(0)=0, and

    φ1(t)φ2(t)=x4rtlogr1(1rt)x4, ddt[φ1(t)φ2(t)]=(logr)2x4(1x4)rt[1(1rt)x4]2>0

    for all t(0,1). So that φ1(t)/φ2(t) is strictly increasing on (0,1), applying Lemma 2.1, the assertion of part (2) follows.

    Employing Lemma 4.3, it is easy to see that the best estimates in (4.1) and (4.2) are arrived at for p=1, and while these in (4.3) and (4.4) for q=1/2.

    Theorem 4.4. The double sharp inequalities

    (125x4)1/4<arcsl(x)x<[1(11ω4)x4]1/4,(1+25x4)1/4<arcslh(x)x<[1(14ω4)x4]1/4,(1+35x4)1/4<arctl(x)x<[1(11τ4)x4]1/4,(135x4)1/4<arctlh(x)x<[1(114ω4)x4]1/4

    hold for all x(0,1).

    This research was supported by the Natural Science Foundation of China (11701176, 61673169, 11301127) and the Natural Science Foundation of Zhejiang Province (Grant YL19A010012), and the Key Project of the Scientific Research of Zhejiang Open University in 2019 (Grant no XKT-19Z02).

    The authors declare that they have no competing interests.



    [1] J. M. Borwein, P. B. Borwein, Pi and the AGM: A study in analytic number theory and computational complexity, New York: John Wiley & Sons, 1987.
    [2] B. C. Carlson, Algorithms involving arithmetic and geometric means, Am. Math. Mon., 78 (1971), 496–505. doi: 10.1080/00029890.1971.11992791
    [3] E. Neuman, On Gauss lemniscate functions and lemniscatic mean, Math. Panno., 18 (2007), 77–94.
    [4] M. Abramowitz, I. A. Stegun, Handbook of mathematical functions with formulas, graphs and mathematical tables, New York: Dover, 1965.
    [5] H. Alzer, Sharp inequalities for the complete elliptic integral of the first kind, Math. Proc. Cambridge, 124 (1998), 309–314. doi: 10.1017/S0305004198002692
    [6] H. Alzer, S. L. Qiu, Monotonicity theorems and inequalities for the complete elliptic integrals, J. Comput. Appl. Math., 172 (2004), 289–312. doi: 10.1016/j.cam.2004.02.009
    [7] S. L. Qiu, M. Vuorinen, Special functions in geometric function theory, In: Handbook of complex analysis: Geometric function theory, 2 (2005), 621–659.
    [8] M. K. Wang, Y. M. Chu, Asymptotical bounds for complete elliptic integrals of the second kind, J. Math. Anal. Appl., 402 (2013), 119–126. doi: 10.1016/j.jmaa.2013.01.016
    [9] M. K. Wang, H. H. Chu, Y. M. Li, Y. M. Chu, Answers to thress conjectues on convexity of three functions involving complete elliptic integrals of the first kind, Appl. Anal. Discr. Math., 14 (2020), 255–271. doi: 10.2298/AADM190924020W
    [10] Z. H. Yang, W. M. Qian, W. Zhang, Y. M. Chu. Notes on the complete elliptic integral of the first kind, Math. Inequal. Appl., 23 (2020), 77–93.
    [11] G. E. Andrews, R. Askey, R. Roy, Special functions, In: Encyclopedia of mathematics and its applications, Cambridge: Cambridge University Press, 70 (1999).
    [12] Y. X. Li, A. Rauf, M. Naeem, M. A. Binyamin, A. Aslam, Valency-based topological properties of linear hexagonal chain and hammer-like benzenoid, Complexity, 2021 (2021), 9939469.
    [13] Z. H. Yang, W. M. Qian, Y. M. Chu, W. Zhang, On rational bounds for the gamma function, J. Inequal. Appl., 2017 (2017), 210. doi: 10.1186/s13660-017-1484-y
    [14] T. H. Zhao, Y. M. Chu, H. Wang, Logarithmically complete monotonicity properties relating to the gamma function, Abstr. Appl. Anal., 2011 (2011), 896483.
    [15] E. Neuman, On Gauss lemniscate functions and lemniscatic mean Ⅱ, Math. Pannon., 23 (2012), 65–73.
    [16] Y. X. Li, T. Muhammad, M. Bilal, M. A. Khan, A. Ahmadian, B. A. Pansera, Fractional simulation for Darcy-Forchheimer hybrid nanoliquid flow with partial slip over a spinning disk, Alex. Eng. J., 60 (2021), 4787–4796. doi: 10.1016/j.aej.2021.03.062
    [17] M. I. Khan, S. Kadry, Y. M. Chu, W. A. Khane, A. Kumar, Exploration of Lorentz force on a paraboloid stretched surface in flow of Ree-Eyring nanomaterial, J. Mater. Res. Technol., 9 (2020), 10265–10275. doi: 10.1016/j.jmrt.2020.07.017
    [18] J. E. Deng, C. P. Chen, Sharp Shafer-Fink type inequalities for Gauss lemniscate functions, J. Inequal. Appl., 2014 (2014), 35. doi: 10.1186/1029-242X-2014-35
    [19] J. Liu, C. P. Chen, Padé approximant related to inequalities for Gauss lemniscate functions, J. Inequal. Appl., 2016 (2016), 1–16. doi: 10.1186/s13660-015-0952-5
    [20] E. Neuman, Two-sided inequalities for the lemniscate functions, J. Inequalities Special Funct., 1 (2010), 1–7.
    [21] E. Neuman, On lemniscate functions, Integr. Transf. Spec. F., 24 (2013), 164–171.
    [22] M. J. Wei, Y. He, G. D. Wang, Shafer-Fink type inequalities for arc lemniscate functions, RACSAM, 114 (2020), 53. doi: 10.1007/s13398-020-00782-x
    [23] Y. M. Chu, M. K. Wang, X. Y. Ma, Sharp bounds for Toader mean in terms of contraharmonic mean with applications, J. Math. Inequal., 7 (2013), 161–166.
    [24] Y. M. Chu, M. K. Wang, Y. F. Qiu, X. Y. Ma, Sharp two parameter bounds for the logarithmic mean and the arithmetic-geometric mean of Gauss, J. Math. Inequal., 7 (2013), 349–355.
    [25] M. Nowicka, A. Witkowski, Optimal bounds for the tangent and hyperbolic sine means Ⅱ, J. Math. Inequal., 14 (2020), 23–33.
    [26] W. M. Qian, Z. Y. He, Y. M. Chu, Approximation for the complete elliptic integral of the first kind, RACSAM, 114 (2020), 57. doi: 10.1007/s13398-020-00784-9
    [27] W. M. Qian, X. H. Zhang, Y. M. Chu, Sharp bounds for the Toader-Qi mean in terms of harmonic and geometric means, J. Math. Inequal., 11 (2017), 121–127.
    [28] Y. X. Li, M. H. Alshbool, Y. P. Lv, I. Khan, M. R. Khan, A. Issakhov, Heat and mass transfere in MHD Williamson nanofluid flow over an exponentially porous stretching surface, Case Stud. Therm. Eng., 26 (2021), 100975. doi: 10.1016/j.csite.2021.100975
    [29] W. F. Xiang, S. Y. Xu, J. W. Lu, Y. M. Li, Y. M. Chu, Z. Q. Zhang, Event-triggered filtering for discrete-time Markovian jump systems with additive time-varying delays, Appl. Math. Comput., 391 (2021), 125630.
    [30] M. K. Wang, Y. M. Chu, Y. M. Li, W. Zhang, Asymptotic expansion and bounds for complete elliptic integrals, Math. Inequal. Appl., 23 (2020), 821–841.
    [31] M. K. Wang, Y. M. Chu, Y. Q. Song, Asymptotical formulas for Gaussian and generalized hypergeometric functions, Appl. Math. Comput., 276 (2016), 44–60.
    [32] M. K. Wang, H. H. Chu, Y. M. Chu, Precise bounds for the weighted Hölder mean of the complete p-elliptic integrals, J. Math. Anal. Appl., 480 (2019), 123388. doi: 10.1016/j.jmaa.2019.123388
    [33] H. H. Chu, T. H. Zhao, Y. M. Chu, Sharp bounds for the Toader mean of order 3 in terms of arithmetic, quadratic and contraharmonic means, Math. Slovaca, 70 (2020), 1097–1112. doi: 10.1515/ms-2017-0417
    [34] W. M. Qian, Y. Y. Yang, H. W. Zhang, Y. M. Chu, Optimal two-parameter geometric and arithmetic mean bounds for the Sándor-Yang mean, J. Inequal. Appl., 2019 (2019), 287. doi: 10.1186/s13660-019-2245-x
    [35] W. M. Qian, Z. Y. He, H. W. Zhang, Y. M. Chu, Sharp bounds for Neuman means in terms of two-parameter contraharmonic and arithmetric mean, J. Inequal. Appl., 2019 (2019), 168. doi: 10.1186/s13660-019-2124-5
    [36] G. D. Anderson, M. K. Vamanamurthy, M. Vuorinen, Conformal invariants, inequalities, and quasiconformal maps, New York: John Wiley & Sons, 1997.
  • Reader Comments
  • © 2021 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(2564) PDF downloads(94) Cited by(0)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog