Sharp Bounds for the Harmonic Numbers

Similar documents
Ramanujan s Harmonic Number Expansion

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

INEQUALITIES FOR THE GAMMA FUNCTION

On the stirling expansion into negative powers of a triangular number

The Error in an Alternating Series

16.4. Power Series. Introduction. Prerequisites. Learning Outcomes

Math 0230 Calculus 2 Lectures

Some New Facts in Discrete Asymptotic Analysis

8.1 Sequences. Example: A sequence is a function f(n) whose domain is a subset of the integers. Notation: *Note: n = 0 vs. n = 1.

arxiv:math/ v1 [math.ca] 16 Jun 2003

16.4. Power Series. Introduction. Prerequisites. Learning Outcomes

Chapter 11: Sequences; Indeterminate Forms; Improper Integrals

Infinite Series Summary

Chapter 8. Infinite Series

Fractional part integral representation for derivatives of a function related to lnγ(x+1)

Section 11.1: Sequences

1 + lim. n n+1. f(x) = x + 1, x 1. and we check that f is increasing, instead. Using the quotient rule, we easily find that. 1 (x + 1) 1 x (x + 1) 2 =

Sharp inequalities and complete monotonicity for the Wallis ratio

Upper Bounds for Partitions into k-th Powers Elementary Methods

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

Series solutions to a second order linear differential equation with regular singular points

Absolute Convergence and the Ratio Test

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

Sharp inequalities and asymptotic expansion associated with the Wallis sequence

AP Calculus Chapter 9: Infinite Series

Chapter 11 - Sequences and Series

SOME INEQUALITIES FOR THE q-digamma FUNCTION

The best expert versus the smartest algorithm

Explicit formulas for computing Bernoulli numbers of the second kind and Stirling numbers of the first kind

arxiv: v1 [math.ca] 25 Jan 2011

10.1 Sequences. Example: A sequence is a function f(n) whose domain is a subset of the integers. Notation: *Note: n = 0 vs. n = 1.

Lecture 32: Taylor Series and McLaurin series We saw last day that some functions are equal to a power series on part of their domain.

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

,... We would like to compare this with the sequence y n = 1 n

The Integral Test. P. Sam Johnson. September 29, P. Sam Johnson (NIT Karnataka) The Integral Test September 29, / 39

1. Prove the following properties satisfied by the gamma function: 4 n n!

Given a sequence a 1, a 2,...of numbers, the finite sum a 1 + a 2 + +a n,wheren is an nonnegative integer, can be written

Solutions to Homework 2

Quadratic Transformations of Hypergeometric Function and Series with Harmonic Numbers

Generalized Geometric Series, The Ratio Comparison Test and Raabe s Test

Partial Fractions. Calculus 2 Lia Vas

Section Taylor and Maclaurin Series

Bounded Infinite Sequences/Functions : Orders of Infinity

INFINITE SEQUENCES AND SERIES

CHAPTER 8: EXPLORING R

Rearrangements of convergence sums at infinity

Introduction to Series and Sequences Math 121 Calculus II Spring 2015

10.1 Sequences. A sequence is an ordered list of numbers: a 1, a 2, a 3,..., a n, a n+1,... Each of the numbers is called a term of the sequence.

SJÄLVSTÄNDIGA ARBETEN I MATEMATIK

Mathematics 324 Riemann Zeta Function August 5, 2005

1. Introduction Interest in this project began with curiosity about the Laplace transform of the Digamma function, e as ψ(s + 1)ds,

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

Appendix G: Mathematical Induction

About the Gamma Function

Study of some equivalence classes of primes

ln(9 4x 5 = ln(75) (4x 5) ln(9) = ln(75) 4x 5 = ln(75) ln(9) ln(75) ln(9) = 1. You don t have to simplify the exact e x + 4e x

Assignment 4. u n+1 n(n + 1) i(i + 1) = n n (n + 1)(n + 2) n(n + 2) + 1 = (n + 1)(n + 2) 2 n + 1. u n (n + 1)(n + 2) n(n + 1) = n

APPENDIX : PARTIAL FRACTIONS

CALCULUS JIA-MING (FRANK) LIOU

On the Convergence of the Summation Formulas Constructed by Using a Symbolic Operator Approach

Solutions Quiz 9 Nov. 8, Prove: If a, b, m are integers such that 2a + 3b 12m + 1, then a 3m + 1 or b 2m + 1.

Exact formulae for the prime counting function

A SUMMATION FORMULA FOR SEQUENCES INVOLVING FLOOR AND CEILING FUNCTIONS

9.4. Mathematical Induction. Introduction. What you should learn. Why you should learn it

Successive Derivatives and Integer Sequences

On a series of Ramanujan

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

Taylor and Maclaurin Series

arxiv: v1 [math.ca] 15 Jan 2018

Here is another characterization of prime numbers.

Power series and Taylor series

Analysis II: Basic knowledge of real analysis: Part IV, Series

Review (11.1) 1. A sequence is an infinite list of numbers {a n } n=1 = a 1, a 2, a 3, The sequence is said to converge if lim

Defining Incomplete Gamma Type Function with Negative Arguments and Polygamma functions ψ (n) ( m)

The integrals in Gradshteyn and Ryzhik. Part 10: The digamma function

MATH3283W LECTURE NOTES: WEEK 6 = 5 13, = 2 5, 1 13

Part 3.3 Differentiation Taylor Polynomials

EULER AND THE FUNCTION SIN(X)/X

THREE INEQUALITIES INVOLVING HYPERBOLICALLY TRIGONOMETRIC FUNCTIONS

CHAPTER 4. Series. 1. What is a Series?

The Gamma Function. July 9, Louisiana State University SMILE REU. The Gamma Function. N. Cannady, T. Ngo, A. Williamson.

n=1 ( 2 3 )n (a n ) converges by direct comparison to

BOUNDS ON THE COEFFICIENTS OF THE CHARACTERISTIC AND MINIMAL POLYNOMIALS

SOLUTIONS FOR THE THIRD PROBLEM SET

Differential Calculus

Inequalities and Monotonicity For The Ratio of Γ p Functions

3 a = 3 b c 2 = a 2 + b 2 = 2 2 = 4 c 2 = 3b 2 + b 2 = 4b 2 = 4 b 2 = 1 b = 1 a = 3b = 3. x 2 3 y2 1 = 1.

FACTORS OF GIBBS MEASURES FOR FULL SHIFTS

A Symbolic Operator Approach to Power Series Transformation-Expansion Formulas

Equations with regular-singular points (Sect. 5.5).

Harmonic Numbers. Math. 55 Some Inequalities May 9, :54 pm

Evaluating ζ(2m) via Telescoping Sums.

Continuity. The Continuity Equation The equation that defines continuity at a point is called the Continuity Equation.

Mathematics 136 Calculus 2 Everything You Need Or Want To Know About Partial Fractions (and maybe more!) October 19 and 21, 2016

Eigenvalues, random walks and Ramanujan graphs

Bernoulli Numbers and their Applications

Number Theory, Algebra and Analysis. William Yslas Vélez Department of Mathematics University of Arizona

INFINITE SEQUENCES AND SERIES

ON A WEIGHTED INTERPOLATION OF FUNCTIONS WITH CIRCULAR MAJORANT

arxiv: v2 [math.ca] 12 Sep 2013

Transcription:

Sharp Bounds for the Harmonic Numbers arxiv:math/050585v3 [math.ca] 5 Nov 005 Mark B. Villarino Depto. de Matemática, Universidad de Costa Rica, 060 San José, Costa Rica March, 08 Abstract We obtain best upper and lower bounds for the Lodge-Ramanujan and DeTemple- Wang approximations to the nth Harmonic Number. Introduction For every natural number n the Harmonic Number, H n is the nth partial sum of the harmonic series: H n := + + 3 + + n. ) Although the asymptotics of H n were determined by Euler, see [4]), in his famous formula: H n lnn+γ + n n + [ ], ) 0n4 where γ = 0.577 is Euler s constant and each summand in the asymptotic expansion is of the form B k n k, where B k denots the kth Bernoulli number, mathematicians have continued to offer alternate approximative formulas to Euler s. We cite the following formulas, which appear in order of increasing accuracy: H n lnn+γ + n+ 3 ln nn+)+γ + 6nn+)+ 6 5 ln n+ ) +γ + 4 n+ 3) 4) ). 5) + 5

The formula 3) is the Tóth-Mare approximation, see [5]), and itunderestimates the true value of H n by terms of order 7n3; the second, 4), is the Lodge-Ramanujan approximation, and it overestimates the true value of H n by terms of order 9 350[nn+)] 3, see [6]); and the last, 5), is the DeTemple-Wang approximation, and it overestimates 07 the true value of H n by terms of order 806400 ) n+ 6, see []). In 003, Chao-Ping Chen and Feng Qi, see []), published a proof of the following sharp form of the Tóth-Mare approximation: Theorem. For any natural number n, the following inequality is valid: n+ γ H n lnn γ < n+. 6) 3 The constants =.3657 and γ 3 n =. are the best possible, and equality holds only for The first statement of this theorem had been announced ten years earlier by the editors of the Problems section of the American Mathemtical Monthyly, Vol 99, No. 7, Jul-Aug, 99), p 685, as part of a commentary on the solution of Problem 343, but they did not publish the proof. So, the first published proof is apparently that of Chen and Qi. In this paper we will prove sharp forms of the Lodge-Ramanujan approximation and the DeTemple-Wang approximation. Theorem. For any natural number n, the following inequality is valid: 6nn+)+ 6 5 < H n ln nn+) γ 6nn+)+ γ ln γ ln. 7) The constants γ ln γ ln only for n =. and =.50934 and 6 5 are the best possible, and equality holds Theorem 3. For any natural number n, the following inequality is valid: 4 n+ ) + 5 H n lnn γ < 4 ) n+. 8) + 54ln 3 +54γ 53 ln 3 γ The constants 54ln 3 +54γ 53 ln 3 γ holds only for n =. = 3.739975 and 5 are the best possible, and equality

All three theorems are corollaries of the following stronger theorem: Theorem 4. For any natural number n, define f n, λ n, and d n by: H n := lnn+γ + 9) n+f n := ln nn+)+γ + 0) 6nn+)+λ n := ln n+ ) +γ + 4 ) n+, ) +dn respectively. Then for any natural number n the sequence {f n } is monotonically decreasing while the sequences {λ n } and {d n } are monotonically increasing. Chen and Qi, see []), proved that the sequence {f n } decreases monotonically. In this paper we will prove the monotonicity of the sequences {λ n } and {d n }. Lemmas Our proof is based on inequalities satisfied by the digamma function, Ψx): Ψx) := d dx lnγx) Γ x) Γx) γ x +x nx+n), ) which is the generalization of H n to the real variable x since Ψx) and H n satisfiy the equation: Ψn+) = H n γ. 3) Lemma. For every x > 0 there exist numbers θ x and Θ x, with 0 < θ x < and 0 < Θ x <, for which the following equations are true: Ψx+) = lnx+ x x + 0x 4 5x 6 + 40x 8θ x, 4) Ψ x+) = x x + 6x 3 30x 5 + 4x 7 30x 9Θ x. 5) Proof. Both formulas are well-known. See, for example, [3], pp 4-5. n= 6) Lemma. The following inequalities are true for x > 0: 3xx+) 5x x+) < Ψx+) ln{xx+)} < 3xx+) 5x x+) + 8 35x 3 x+) 3, 7) 3

x xx+) 3x 3 + 5x 5 8x 7 < x + x+ Ψ x+) < x xx+) 3x + 3 5x5. 8) Proof. The inequalities 7) were proved in our paper, see[6]), for integers n instead of the real variable x. But the proofs are valid for real x. For 8) we start with 5) of Lemma. We conclude that x 6x 3 + 30x 5 36x 7 < x Ψ x+) < x 6x 3 + 30x 5. Now we multiply to all three components of the inequality by and add Lemma 3. The following inequalities are true for x > 0: ) x+ x + x 6x + 3 30x 5 4x < 7 x+ < Ψ x+) x+ x+ x to them. ) x + x 6x + 3 30x5, 9) 4x 4x + 3 3 960x 4 60x 5 8064x 6 896x < Ψx+) ln x+ ) 7 < 4x 4x + 3 3 960x 4 60x 5 8064x 6 896x + 43 7 3070x8. 0) Proof. Similar to the proof of Lemma. 3 Proof for the Lodge-Ramanujan approximation Proof. We solve 0) for λ n and use 3) to obtain Define λ n = Λ x := for all x > 0. Observe that Λ n = λ n. Ψn+) ln nn+) 6nn+). 3xx+). ) Ψx+) lnxx+) We will show that Λ x > 0 for x > 5. Computing the derivative we obtain Λ x = x + x+ Ψ x+) {Ψx+) ln{xx+)} 6x+3) 4

and therefore {Ψx+) ln{xx+)} Λ x = x + x+ Ψ x+) 6x+3){Ψx+) ln{xx+)}. By Lemma, this is greater than x xx+) 3x + 3 5x 5 8x 7 { 6x+3) 3xx+) 5x x+) + 8 35x 3 x+) 3 = 07x6 +840x 5 789x 4 4966x 3 50999x 78x 3675 6650x 7 x+) 6 ) 5 x + 30 3797 x+ 07 07 + 685960 07 = x 5) x 5 + 95 5 x4 + 68 5 x3 + 784 05 7 x7 x+) 6 which is obviously positive for x > 5. For x =,, 3, 4, 5, we compute directly: Λ =.56075467 Λ =.58489 Λ 3 =.5958588 Λ 4 =.5948086 Λ 5 =.5964909 } Therefore, the sequence {Λ n }, n, is a strictly increasing sequence, and therefore so is the sequence {λ n }. Moreover, in [6], we proved that where 0 < n < 38 75nn+). Therefore λ n = 6 5 n, This completes the proof. lim n λ n = 6 5. 4 Proof for the DeTemple-Wang Approximation Proof. Following the idea in the proof of the Lodge-Ramanujan approximation we solve ) for d n and define the corresponding real-variable version. Let d x := Ψx+) ln ) 4 x+ ) x+ ) 5

We compute the derivative, ask when it is positive, clear the denominator and observe that we have to solve the inequality: { x+ } Ψ x+) 48 x+ ){ Ψx+) ln x+ > 0. )} By Lemma 3, the left hand side of this inequality is > x+ ) x + x 6x 3 + 30x 5 4x 7 48 x+ ) 4x 4x 3 + 3 960x 4 60x 5 8064x 6 896x 7 + 43 3070x 8 ) for all x > 0. This last quantity is equal to 908009 374776x 4007876x +593379x 3 +4547x 4 93996x 5 +956659x 6 +6563008x 7 70008896x 8 53447408x 9 +783936x 0 )/734085600x 6 +x)) The denominator, 734085600x 6 +x), is evidently positive for x > 0 and the numerator can be written in the form where px)x 4)+r px) = 548963409+374874745x+343568883x +8564093760x 3 +385987x 4 +56684979x 5 +80800x 6 +547648x 7 +783936x 8 +783936x 9 with remainder r equal to r = 95843950359. Therefore, the numerator is clearly positive for x > 4, and therefore, the derivative, d x, too, is postive for x > 4. Finally d = 3.739975 d = 4.089544 d 3 = 4.30874 d 4 = 4.588035 Therefore {d n } is an increasing sequence for n. 6

Now, if we expand the formula for d n into an asymptotic series in powers of ), we n+ obtain d n 5 400 07 ) + n+ and we conclude that This completes the proof. References lim d n = n 5. [] Ch.-P. Chen and F. Qi, The best bounds of the n-th harmonic number, Global Journal of Mathematics and Mathematical Sciences 006), accepted. The best lower and upper bounds of harmonic sequence, RGMIA Research Report Collection 6 003), no., Article 4. The best bounds of harmonic sequence, available online at http://front.math.ucdavis.edu/math.ca/030633. [] D. DeTemple and S-H Wang Half-integer Approximations for the Partial Sums of the Harmonic Series Journal of Mathematical Analysis and Applications, 60 99), 49-56. [3] J. Edwards A Treatise on the Integral Calculus, Vol II, Chelsea, New York, 955. [4] K. Knopp Theory and Application of Infinite Series, Dover, New York, 990. [5] L. Tóth, and S. Mare E 343 American Mathematical Monthly, 98 99), no 3, 64. [6] M. Villarino, Ramanujan s Approximation to the nth Partial Sum of the Harmonic Series, preprint, arxiv.math.ca/040354 7