Approximations to inverse tangent function

Similar documents
Research Article On New Wilker-Type Inequalities

Optimal bounds for Neuman-Sándor mean in terms of the convex combination of the logarithmic and the second Seiffert means

Lazarević and Cusa type inequalities for hyperbolic functions with two parameters and their applications

Sharp inequalities and asymptotic expansion associated with the Wallis sequence

Research Article Inequalities for Hyperbolic Functions and Their Applications

Some new sequences that converge to the Ioachimescu constant

Research Article On Jordan Type Inequalities for Hyperbolic Functions

Oscillation theorems for nonlinear fractional difference equations

THE inverse tangent function is an elementary mathematical

REFINEMENTS AND SHARPENINGS OF SOME DOUBLE INEQUALITIES FOR BOUNDING THE GAMMA FUNCTION

Wirtinger inequality using Bessel functions

A monotonic refinement of Levinson s inequality

Some new continued fraction sequence convergent to the Somos quadratic recurrence constant

THREE INEQUALITIES INVOLVING HYPERBOLICALLY TRIGONOMETRIC FUNCTIONS

arxiv: v1 [math.ca] 2 Nov 2018

Complete monotonicity of a function involving the p-psi function and alternative proofs

A general theorem on the stability of a class of functional equations including quadratic additive functional equations

Monotonicity rule for the quotient of two functions and its application

Existence and multiplicity of positive solutions of a one-dimensional mean curvature equation in Minkowski space

On complete convergence and complete moment convergence for weighted sums of ρ -mixing random variables

On some Hermite Hadamard type inequalities for (s, QC) convex functions

Fourier series of sums of products of ordered Bell and poly-bernoulli functions

INEQUALITIES FOR THE GAMMA FUNCTION

Research Article Some Monotonicity Properties of Gamma and q-gamma Functions

A class of mixed orthogonal arrays obtained from projection matrix inequalities

On the Iyengar Madhava Rao Nanjundiah inequality and its hyperbolic version

arxiv: v5 [math.ca] 28 Apr 2010

Some identities related to Riemann zeta-function

A COMPLETELY MONOTONIC FUNCTION INVOLVING THE TRI- AND TETRA-GAMMA FUNCTIONS

arxiv: v1 [math.ca] 15 Jan 2018

Stability of a Functional Equation Related to Quadratic Mappings

Horst Alzer A MEAN VALUE INEQUALITY FOR THE DIGAMMA FUNCTION

COMPLETE MONOTONICITIES OF FUNCTIONS INVOLVING THE GAMMA AND DIGAMMA FUNCTIONS. 1. Introduction

Permutation transformations of tensors with an application

Research Article New Oscillation Criteria for Second-Order Neutral Delay Differential Equations with Positive and Negative Coefficients

Certain inequalities involving the k-struve function

DOI: /j3.art Issues of Analysis. Vol. 2(20), No. 2, 2013

Research Article Convergence Theorems for Common Fixed Points of Nonself Asymptotically Quasi-Non-Expansive Mappings

Additive functional inequalities in Banach spaces

Sparse signals recovered by non-convex penalty in quasi-linear systems

Newcriteria for H-tensors and an application

Mat104 Fall 2002, Improper Integrals From Old Exams

Sharp bounds for Seiffert and Neuman-Sándor means in terms of generalized logarithmic means

Properties for the Perron complement of three known subclasses of H-matrices

Journal of Mathematical Analysis and Applications

arxiv: v1 [math.ca] 25 Jan 2011

Two new regularization methods for solving sideways heat equation

Research Article A Nice Separation of Some Seiffert-Type Means by Power Means

A regularity criterion for the generalized Hall-MHD system

Some inequalities for unitarily invariant norms of matrices

Research Article A Note on Kantorovich Inequality for Hermite Matrices

Optimal bounds for the Neuman-Sándor mean in terms of the first Seiffert and quadratic means

Bounds for sine and cosine via eigenvalue estimation

A Symbolic Operator Approach to Several Summation Formulas for Power Series

Covering functionals of cones and double cones

Simplifying Coefficients in a Family of Ordinary Differential Equations Related to the Generating Function of the Laguerre Polynomials

Norm inequalities related to the Heinz means

TOTAL NAME DATE PERIOD AP CALCULUS AB UNIT 4 ADVANCED DIFFERENTIATION TECHNIQUES DATE TOPIC ASSIGNMENT /6 10/8 10/9 10/10 X X X X 10/11 10/12

Latter research on Euler-Mascheroni constant. 313, Bucharest, Romania, Târgovişte, Romania,

Boundary value problems for fractional differential equations with three-point fractional integral boundary conditions

An improved generalized Newton method for absolute value equations

New hybrid conjugate gradient methods with the generalized Wolfe line search

On the Weak Convergence of the Extragradient Method for Solving Pseudo-Monotone Variational Inequalities

An Exact Formula for all Star-Kipas Ramsey Numbers

Research Article A New Class of Meromorphic Functions Associated with Spirallike Functions

Lyapunov-type inequalities and disconjugacy for some nonlinear difference system

arxiv: v1 [math.ca] 4 Aug 2012

Strong convergence theorems for asymptotically nonexpansive nonself-mappings with applications

October 27, 2018 MAT186 Week 3 Justin Ko. We use the following notation to describe the limiting behavior of functions.

Computing the Determinant and Inverse of the Complex Fibonacci Hermitian Toeplitz Matrix

Research Article The Dirichlet Problem on the Upper Half-Space

On the Inverting of a General Heptadiagonal Matrix

Research Article Sharp Bounds by the Generalized Logarithmic Mean for the Geometric Weighted Mean of the Geometric and Harmonic Means

Math Exam 1a. c) lim tan( 3x. 2) Calculate the derivatives of the following. DON'T SIMPLIFY! d) s = t t 3t

Schwarz lemma involving the boundary fixed point

Sharpened versions of the Erdös-Mordell inequality

The Greatest Common Divisor of k Positive Integers

A NOTE ON LINEAR FUNCTIONAL NORMS

Bounds Improvement for Neuman-Sándor Mean Using Arithmetic, Quadratic and Contraharmonic Means 1

Sums of finite products of Chebyshev polynomials of the third and fourth kinds

Half convex functions

Growth properties at infinity for solutions of modified Laplace equations

Solution of fractional differential equations via α ψ-geraghty type mappings

New families of special numbers and polynomials arising from applications of p-adic q-integrals

NSE characterization of the Chevalley group G 2 (4)

Extensions of interpolation between the arithmetic-geometric mean inequality for matrices

British Journal of Applied Science & Technology 10(2): 1-11, 2015, Article no.bjast ISSN:

On Picard value problem of some difference polynomials

Dissipative Hyperbolic Geometric Flow on Modified Riemann Extensions

Sigmoid functions for the smooth approximation to x

The iterative methods for solving nonlinear matrix equation X + A X 1 A + B X 1 B = Q

Pre-Calculus and Trigonometry Capacity Matrix

Analytic Trigonometry. Copyright Cengage Learning. All rights reserved.

Sharp Bounds for Seiffert Mean in Terms of Arithmetic and Geometric Means 1

Monotone variational inequalities, generalized equilibrium problems and fixed point methods

The Generalized Viscosity Implicit Rules of Asymptotically Nonexpansive Mappings in Hilbert Spaces

Oscillation results for certain forced fractional difference equations with damping term

A Note of the Strong Convergence of the Mann Iteration for Demicontractive Mappings

Inverse spectral problems for first order integro-differential operators

On series of functions with the Baire property

Transcription:

Qiao Chen Journal of Inequalities Applications (2018 2018:141 https://doiorg/101186/s13660-018-1734-7 R E S E A R C H Open Access Approimations to inverse tangent function Quan-Xi Qiao 1 Chao-Ping Chen 1* * Correspondence: chenchaoping@sohucom 1 School of Mathematics Informatics, Henan Polytechnic University, Jiaozuo, China Abstract In this paper, we present a sharp Shafer-type inequality for the inverse tangent function Based on the Padé approimation method, we give approimations to the inverse tangent function Based on the obtained result, we establish new bounds for arctan MSC: 26D05 Keywords: Inverse trigonometric function; Inequality; Approimation 1 Introduction In 1966, Shafer [1] posed, as a problem, the following inequality: 3 1+2 < arctan, > 0 (11 1+2 Three proofs of it were later given in [2] Shafer s inequality (11 was sharpened generalized by Qi et al in [3] A survey epository of some old new inequalities associated with trigonometric functions can be found in [4] Chen et al [5]presentedanew method to sharpen bounds of both sinc arcsin functions, the inequalities in eponential form as well For each a > 0, Chen Cheung [6]determinedthelargestnumberb the smallest number c such that the inequalities b 1+a 1+ arctan c 2 1+a (12 1+ 2 are valid for all 0 More precisely, these author proved that the largest number b the smallest number c required by inequality (12are when 0 < a 2, b a, c 1+a; 2 when 2 < a 2 2, b 4(a2 1, c 1+a; a 2 2 when 2 < a <2, b 4(a2 1, c a 2 2 a; when 2 a <, b 1+a, c 2 a The Author(s 2018 This article is distributed under the terms of the Creative Commons Attribution 40 International License (http://creativecommonsorg/licenses/by/40/, which permits unrestricted use, distribution, reproduction in any medium, provided you give appropriate credit to the original author(s the source, provide a link to the Creative Commons license, indicate if changes were made

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 2 of 14 In 1974, Shafer [7] indicated several elementary quadratic approimations of selected functions without proof Subsequently, Shafer [8] established these results as analytic inequalities For eample, Shafer [8]provedthat,for >0, 8 < arctan (13 3+ 25 + 803 2 The inequality (13 canalsobe foundin [9] The inequality (13isanimprovementofthe inequality (11 Zhu [10]developed(13 to produce a symmetric double inequality More precisely, the author proved that, for >0, 8 < arctan < 3+ 25 + 803 2 where the constants 80/3 256/ 2 are the best possible 8, (14 3+ 25 + 256 2 2 Remark 11 For > 0, the following symmetric double inequality holds: 8 < arctan < 3+ 25 + 803 2 3+ 2 15 3 25 + 80 3 2, (15 where the constants 8 2 15 3 are the best possible We here point out that, for >0, the upper bound in (14 is better than the upper bound in (15 Based on the following power series epansion: ( arctan 3+ 25 + 803 2 8 + 32 4725 7 64 4725 9 + 25,376 1,299,375 11, Sun Chen [11] presented a new upper bound proved that, for >0, 8 + 32 4725 arctan < 7 (16 3+ 25 + 80 3 2 Moreover, these authors pointed out that, for 0 < < 0 14243, the upper bound in (16 is better than the upper bound in (14 In fact, we have the following approimation formulas near the origin: 8 ( arctan O 3, 3+ 25 + 256 2 2 3 arctan 1+2 1+ O( 5, 2 8 arctan O 3+ 25 ( 7, + 803 2

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 3 of 14 32 8 + 4725 arctan 7 O 3+ 25 ( 9 + 803 2 Nishizawa [12]provedthat,for >0, 2 4+ ( 2 4 2 +(2 2 < arctan < 2 4+ 32 + (2 2, (17 where the constants ( 2 4 2 32 are the best possible Using the Maple software, we derive the following asymptotic formulas in the Appendi: arctan arctan 2 4+ 32 + (2 12 2 ( 1 2 3 2 + O, (18 4 5 3 2 24 2 + 432 24 2 + 4 12(12 2 +(6 2 + 4 ( 72 1 18 3 + O 5 6, (19 ( 2 arctan 2 + 4 ( 15 1 2 + 3 + O 5 6 (110 as In this paper, motivated by (19, we establish a symmetric double inequality for arctan Based on the Padé approimation method, we develop the approimation formula (110to produce a general result More precisely, we determine the coefficients a j b j (1 j k such that ( 2 arctan 2k + a 1 2(k 1 ( + + a k 1 + O,, 2k + b 1 2(k 1 + + b k 4k+2 where k 1 is any given integer Based on the obtained result, we establish new bounds for arctan Some computations in this paper were performed using Maple software 2 Lemma It is well known that 2n+1 ( 1 k 2k+1 2n (2k +1! < sin < ( 1 k 2k+1 (2k +1! k0 k0 (21 2n+1 ( 1 k 2k k0 (2k! < cos < 2n k0 ( 1 k 2k (2k! (22 for >0n N 0 : N {0},whereN denotes the set of positive integers

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 4 of 14 The following lemma will be used in our present investigation Lemma 21 For 0< u < /2, cos u sin 2 u > u 2 5 6 u4 + 91 360 u6 41 1008 u8 (23 sin 3 u > u 3 1 2 u5 + 13 120 u7 41 3024 u9 (24 Proof We find that where cos u sin 2 u 1 4 ( cos u cos(3u u 2 5 6 u4 + 91 360 u6 41 1008 u8 + sin 3 u 1 ( 3 sin u sin(3u 4 u 3 1 2 u5 + 13 120 u7 41 3024 u9 + ( 1 n 1 w n (u (25 n5 ( 1 n 1 W n (u, (26 n5 w n (u 9n 1 (2n! u2n W n (u 3(9n 1 4 (2n +1! u2n+1 Elementary calculations reveal that, for 0 < u < /2 n 5, w n+1 (u w n (u W n+1 (u W n (u < u 2 (9 n+1 1 2(2n +1(n + 1(9 n 1 < (/2 2 (9 n+1 1 2(2n +1(n + 1(9 n 1 3 9 n+1 2(2n +1(n + 1(9 n 1 27 27 2(2n +1(n +1 { 1+ 1 9 5 1 2(2n +1(n +1 } u 2 (9 n+1 1 2(2n +3(n + 1(9 n 1 < w n+1(u <1 w n (u { 1+ 1 9 n 1 } 1,594,323 118,096(2n +1(n +1 <1 Therefore, for fied u (0, /2, the sequences n w n (u n W n (u areboth strictly decreasing for n 5 From (25 (26, we obtain the desired results (23 (24 The proof of Theorem 31 makes use of the inequalities (21 (24

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 5 of 14 3 Sharp Shafer-type inequality Equation (19 motivated us to establish a symmetric double inequality for arctan Theorem 31 For >0,we have 3 2 24 2 + α 12(12 2 +36 2 2 < arctan < 3 2 24 2 + β 12(12 2 +36 2 2, (31 with the best possible constants α 432 24 2 + 4 292538 β 576 192 2 +16 4 239581 (32 Proof The inequality (31 can be written for >0 as ( 3 2 2 β < arctan ( 24 2 2 +12 ( 12 2 36 2 2 < α (33 Bytheelementarychangeofvariablet arctan ( >0,(33becomes β < ϑ(t<α, 0<t < 2, (34 where ( 3 2 tan 2 t ϑ(t ( 24 2 2 t +12 ( 12 2 tan t 36 2 tan 2 t Elementary calculations reveal that lim ϑ(t 576 192 2 +16 4 t 0 + lim ϑ(t 432 24 2 + 4 t /2 Inordertoprove(34, it suffices to show that ϑ(t is strictly increasing for 0 < t < /2 Differentiation yields t 3 cos 3 tϑ (t ( 24t 3 t sin t cos 2 t + ( 3 3 t 12t 3 sin t ( 3 3 + ( 24 3 t 2 ( 24 2 2 t 3 cos t +3 3 cos 3 t : λ(t We now consider two cases to prove λ(t> 0 for 0 < t < /2

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 6 of 14 Case 1: 0 < t 06 Using (21(22, we have, for 0 < t 06, λ(t (6 14 3 t sin(3t+ 3 4 3 cos(3t+ {(6 + 114 } t 3 12t 3 sin t > ( 3 4 3 ( 3 24 t 2 ( 24 2 2 t 3 cos t ( 3t 9 2 t3 + 81 40 t5 243 560 t7 (6 14 3 t + 3 4 3 ( 1 9 2 t2 + 27 8 t4 81 80 t6 + {(6 + 114 t 3 12t }(t 3 16 t3 ( 3 4 3 ( 3 24 t 2 ( 24 2 2 t (1 3 12 t2 + 124 t4 t {24 3 2 2 (28 83 t 3 ( 12 2 } t 2 { 263 + t 6 20 235 192 3 + (1 112 ( 729 2 t 280 243 } t 2240 3 2 Each function in curly braces is positive for t (0, 06] Thus, λ(t >0fort (0, 06] Case 2: 06 < t < /2 We now prove λ(t >0for06<t < /2 Replacing t by u leads to an equivalent 2 inequality: μ(u>0, 0<u < 2 06, where μ(u ( 24 3( 2 u { 3 3 + ( 24 3( 2 u Using (21 (24, we have, for 0 < u < 2 06, ( ( 2 cos u sin 2 u + {3 3 u 12 3 } 2 u cos u 2 ( 24 2 2( 3 } 2 u sin u +3 3 sin 3 u μ(u > ( 24 3( ( 2 u u 2 5 6 u4 + 91 360 u6 41 1008 u8 ( 2 3 }( + {3 3 u 12 ( 2 u { 3 3 + ( 24 3( 2 u +3 3 ( u 3 1 2 u5 + 13 120 u7 41 u 4 { 1 3 4 24+ 1 1 2 u2 + 1 24 u4 1 720 u6 2 ( 24 2 2( 3 }( 2 u 3024 u9 (12 95 ( 3 u + 2 2 11 90 4 +4 u 2 u 1 6 u3 + 1 120 u5

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 7 of 14 ( + 82 360 3 15 + 199 u 3 ( + 1 5 25 56 2 + 41 2016 4 u 4 + ( 403 420 41 } u 504 3 5 >0 We then obtain λ(t>0ϑ (t > 0 for all 0 < t < /2 Hence, ϑ(t isstrictlyincreasing for 0 < t < /2 The proof is complete From (17(31, we obtain the following approimation formulas: arctan n n 2 4+ 32 + (2n : a 2 n (35 arctan n n 3 2 24 2 + 432 24 2 + 4 12(12 2 n +(6n 2 : b n, (36 as n The following numerical computations (see Table 1 would show that, for n N,Eq(36 is sharper than Eq (35 In fact, we have, as n, arctan n n ( 1 a n + O n 4 arctan n n ( 1 b n + O n 5 4 Approimations to arctan For later use, we introduce the Padé approimant (see [13 16] Let f be a formal power series, f (tc 0 + c 1 t + c 2 t 2 + (41 ThePadéapproimationoforder(p, q of the function f is the rational function, denoted by [p/q] f (t p j0 a jt j 1+ q j1 b jt j, (42 Table 1 Comparison between approimation formulas (35 (36 n a n arctan n n arctan n n 1 7055 10 3 5259 10 3 10 595 10 6 3939 10 7 100 7066 10 10 4492 10 12 1000 7182 10 14 4546 10 17 10,000 7193 10 18 4552 10 22 b n

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 8 of 14 where p 0q 1 are two given integers, the coefficients a j b j are given by (see [13 15] a 0 c 0, a 1 c 0 b 1 + c 1, a 2 c 0 b 2 + c 1 b 1 + c 2, a p c 0 b p + + c p 1 b 1 + c p, 0c p+1 + c p b 1 + + c p q+1 b q, 0c p+q + c p+q 1 b 1 + + c p b q, (43 the following holds: [p/q] f (t f (to ( t p+q+1 (44 Thus, the first p + q + 1 coefficients of the series epansion of [p/q] f are identical to those of f From the epansion (see [17,p81] we obtain where arctan 2 + j1 ( 2 arctan ( 1 j, >1, (2j 12j 1 c j 1 1 2j 3 + 1 2 5 1 +, (45 4 76 j0 c j ( 1j 2j +1 for j 0 (46 Let f (t j0 c j t j, (47 with the coefficients c j givenin(46 Then we have f ( 2 j0 ( c j 2j 2 arctan (48 In what follows, the function f is given in(47

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 9 of 14 Based on the Padé approimation method, we now give a derivation of Eq (110 To this end, we consider [1/1] f (t 1 j0 a jt j 1+ 1 j1 b jt j Noting that c 0 1, c 1 1 3, c 2 1 5, (49 holds, we have, by (43, a 0 1, a 1 b 1 1 3, 0 1 5 1 3 b 1, that is, a 0 1, a 1 4 15, b 1 3 5 We thus obtain [1/1] f (t 1+ 4 15t 1+ 3 5t 15t +4 3(5t +3, (410 we have, by (44, f (t 15t +4 3(5t +3 + O ( 1 t 3, t (411 Replacing t by 2 in (411yields(110 From the Padé approimation method the epansion (47, we now present a general result Theorem 41 ThePadéapproimationoforder(p, q of the function f (t c j j0 (at the t j point t is the following rational function: [p/q] f (t 1+ p j1 a jt j 1+ q j1 b jt j t q p ( t p + a 1 t p 1 + + a p t q + b 1 t q 1 + + b q, (412

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 10 of 14 where p 1 q 1 are any given integers, the coefficients a j b j are given by a 1 b 1 + c 1, a 2 b 2 + c 1 b 1 + c 2, a p b p + + c p 1 b 1 + c p, 0c p+1 + c p b 1 + + c p q+1 b q, 0c p+q + c p+q 1 b 1 + + c p b q, (413 c j is given in (46, the following holds: ( 1 f (t [p/q] f (to, t (414 t p+q+1 In particular, replacing t by 2 in (414 yields ( 2 arctan 2(q p ( 2p + a 1 2(p 1 + + a p 2q + b 1 2(q 1 + + b q with the coefficients a j b j given by (413 ( + O 1 2(p+q+1,, (415 Setting (p, q(k, kin(415, we obtain the following corollary Corollary 41 As, ( 2 arctan 2k + a 1 2(k 1 ( + + a k 1 + O, (416 2k + b 1 2(k 1 + + b k 4k+2 where k 1 is any given integer, the coefficients a j b j (1 j k are given by a 1 b 1 + c 1, a 2 b 2 + c 1 b 1 + c 2, a k b k + + c k 1 b 1 + c k, 0c k+1 + c k b 1 + + c 1 b k, 0c 2k + c 2k 1 b 1 + + c k b k, (417 c j is given in (46

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 11 of 14 Setting k 2in(416yields,as, ( 2 arctan 9454 + 735 2 ( +64 1 15(63 4 +70 2 + 15 + O, (418 10 which gives arctan 2 9454 + 735 2 ( +64 1 15(63 4 +70 2 + 15 + O 11 Using the Maple software, we find, as, arctan 2 9454 + 735 2 +64 15(63 4 +70 2 + 15 + 64 ( 1856 1 464,373 + O 13 15 43,659 11 (419 Equation (419motivated usto establish new bounds for arctan Theorem 42 For >0,we have 2 9454 + 735 2 +64 15(63 4 +70 2 + 15 + 64 43,659 11 1856 464,373 13 < arctan < 2 9454 + 735 2 +64 15(63 4 +70 2 + 15 + 64 (420 43,65911 Proof For >0,let ( I(arctan 2 9454 + 735 2 +64 15(63 4 +70 2 + 15 + 64 43,659 1856 11 464,373 13 ( J(arctan 2 9454 + 735 2 +64 15(63 4 +70 2 + 15 + 64 43,659 11 Differentiation yields I ( 64(230,3918 + 372,680 6 + 236,885 4 + 65,400 2 + 6525 35,721 14 (1 + 2 (63 4 +70 2 + 15 2 <0 J ( 64(12,7898 + 15,610 6 + 8890 4 + 2325 2 + 225 3969 12 (1 + 2 (63 4 +70 2 + 15 2 >0 Hence, I( is strictly decreasing J( is strictly increasing for >0, we have I(> lim t I(t0 J(< lim t J(t0 for >0 The proof is complete

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 12 of 14 Remark 41 We point out that, for > 10213, the lower bound in (420isbetterthan the one in (17 For > 0854439, the upper bound in (420 is better than the one in (17 For > 0947273, the lower bound in (420 is better than the one in (31 For > 0792793, the upper bound in (420 is better than the one in (31 5 Conclusions In this paper, we establish a symmetric double inequality for arctan (Theorem 31 We determine the coefficients a j b j (1 j ksuchthat ( 2 arctan 2k + a 1 2(k 1 ( + + a k 1 + O,, 2k + b 1 2(k 1 + + b k 4k+2 where k 1isanygiveninteger(seeCorollary41 Based on the obtained result, we establish new bounds for arctan (Theorem 42 Appendi: A derivation of (18, (19, (110 Define the function F(by F( arctan 1 a + b + c 2 We are interested in finding the values of the parameters a, b, c such that F( converges as fast as possible to zero, as This provides the best approimations of the form arctan 1 a + b + c 2, Using the Maple software, we find, as, F( c 2 2 c + a c c + b 2a2 2 2c 3/2 3 ( 1 + 3a3 + c 2 3ab 3c 2 4 + O 5 The three parameters a, b, c, which produce the fastest convergence of the function F(, are given by c 20, a c 0, b 2a 2 0, namely, if a 4 2, b 32 4, c 4 2

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 13 of 14 We then obtain, as, arctan 1 12 2 4 32 + + 4 2 4 2 2 3 2 4 2 4+ 32 + (2 12 2 2 3 2 + O 4 Define the function G(by G( arctan 1 p + q + r + s 2 Using the Maple software, we find, as, For ( 1 + O 5 ( 1 G( s 2 2 s + r 2s3/2 +2p s + 4qs 3r2 8sp 2 8 spr 2s 3/2 2 8s 5/2 3 + 48s3/2 pq +48 spr 2 36rqs +15r 3 +48s 3/2 p 3 +72sp 2 r +16s 7/2 48s 7/2 4 + ( 120sqr 2 128 spr 3 128s 2 p 4 240sp 2 r 2 + 256rs 3/2 pq + 192s 2 p 2 q 256s 3/2 p 3 r 35r 4 48q 2 s 2 / ( 128s 9/2 5 ( 1 + O 6 p 24 2 3 2, q 432 24 2 + 4 9 4, r 4(12 2 3 3, s 4 2, we obtain, as, arctan 3 2 24 2 + 432 24 2 + 4 12(12 2 +(6 2 + 4 ( 72 1 18 3 + O 5 6 Define the function H(by ( H( 2 arctan 2 + a 1 + a 2 2 + b 1 + b 2 Using the Maple software, we find, as, H( b 1 a 1 5 3a 2 3b 2 3a 1 b 1 +3b 2 1 +1 + a 1b 2 2b 1 b 2 + a 2 b 1 a 1 b 2 1 + b3 1 3 2 3 1 5a 2b 2 +5b 2 2 +10a 1b 1 b 2 15b 2 1 b 2 +5a 2 b 2 1 5a 1b 3 1 +5b4 1 5 4 + a 1b 2 2 +3b 1b 2 2 2a 2b 1 b 2 +3a 1 b 2 1 b 2 4b 3 1 b 2 + a 2 b 3 1 a 1b 4 1 + b5 1 5 ( 1 + O 6

Qiao Chen Journal of Inequalities Applications (2018 2018:141 Page 14 of 14 For a 1 0, b 1 0, a 2 4 15, b 2 3 5, we obtain, as, ( 2 arctan 2 + 4 ( 15 1 2 + 3 + O 5 6 Acknowledgements We thank the editor referees for their careful reading valuable suggestions to make the article more easily readable Funding Not applicable Competing interests The authors declare that they have no competing interests Authors contributions Both authors contributed equally to this work Both authors read approved the final manuscript Publisher s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps institutional affiliations Received: 26 January 2018 Accepted: 13 June 2018 References 1 Shafer, RE: Problem E1867 Am Math Mon 73, 309 (1966 2 Shafer, RE: Problems E1867 (solution Am Math Mon 74, 726 727 (1967 3 Qi, F, Zhang, SQ, Guo, BN: Sharpening generalizations of Shafer s inequality for the arc tangent function J Inequal Appl 2009, Article ID 930294 (2009 https://doiorg/101155/2009/930294 4 Qi, F, Niu, DW, Guo, BN: Refinements, generalizations, applications of Jordan s inequality related problems J Inequal Appl 2009, Article ID 271923 (2009 https://doiorg/101155/2009/271923 5 Chen, XD, Shi, J, Wang, Y, Pan, X: A new method for sharpening the bounds of several special functions Results Math 72, 695 702 (2017 https://doiorg/101007/s00025-017-0700-6 Chen, CP, Cheung, WS, Wang, W: On Shafer Carlson inequalities J Inequal Appl 2011, Article ID840206 (2011 https://doiorg/101155/2011/840206 7 Shafer, RE: On quadratic approimation SIAM J Numer Anal 11, 447 460 (1974https://doiorg/101137/0711037 8 Shafer, RE: Analytic inequalities obtained by quadratic approimation Publ Elektroteh Fak Univ Beogr, Ser Mat Fiz 57(598, 96 97 (1977 9 Shafer, RE: On quadratic approimation, II Publ Elektroteh Fak Univ Beogr, Ser Mat Fiz 602(633,163 170 (1978 10 Zhu, L: On a quadratic estimate of Shafer J Math Inequal 2, 571 574 (2008 http://filesele-mathcom/articles/jmi-02-51pdf 11 Sun, J-L, Chen, C-P: Shafer-type inequalities for inverse trigonometric functions Gauss lemniscate functions J Inequal Appl 2016, 212 (2016 https://doiorg/101186/s13660-016-1157-2 12 Nishizawa, Y: Refined quadratic estimations of Shafer s inequality J Inequal Appl 2017, 40 (2017 https://doiorg/101186/s13660-017-1312-4 13 Bercu, G: Padé approimant related to remarkable inequalities involving trigonometric functions J Inequal Appl 2016, 99 (2016 https://doiorg/101186/s13660-016-1044-14 Bercu, G: The natural approach of trigonometric inequalities Padé approimant J Math Inequal 11, 181 191 (2017 https://doiorg/107153/jmi-11-18 15 Bercu, G, Wu, S: Refinements of certain hyperbolic inequalities via the Padé approimation method J Nonlinear Sci Appl 9,5011 5020 (2016https://doiorg/1022436/jnsa0090705 16 Brezinski, C, Redivo-Zaglia, M: New representations of Padé, Padé-type, partial Padé approimants J Comput Appl Math 284, 69 77 (2015 https://doiorg/101016/jcam201407007 17 Abramowitz, M, Stegun, IA (eds: Hbook of Mathematical Functions with Formulas, Graphs, Mathematical Tables, 9th printing edn Applied Mathematics Series, vol 55 National Bureau of Stards, Washington (1972