Stirling s formula, n-spheres and the Gamma Function



Similar documents
THE SINE PRODUCT FORMULA AND THE GAMMA FUNCTION

Scalar Valued Functions of Several Variables; the Gradient Vector

Representation of functions as power series

Section 5.1 Continuous Random Variables: Introduction

6. Define log(z) so that π < I log(z) π. Discuss the identities e log(z) = z and log(e w ) = w.

5.3 Improper Integrals Involving Rational and Exponential Functions

Review Solutions MAT V (a) If u = 4 x, then du = dx. Hence, substitution implies 1. dx = du = 2 u + C = 2 4 x + C.

The Math Circle, Spring 2004

Sequences and Series

1 Lecture: Integration of rational functions by decomposition

Calculus 1: Sample Questions, Final Exam, Solutions

Høgskolen i Narvik Sivilingeniørutdanningen STE6237 ELEMENTMETODER. Oppgaver

AB2.5: Surfaces and Surface Integrals. Divergence Theorem of Gauss

y cos 3 x dx y cos 2 x cos x dx y 1 sin 2 x cos x dx

Section 9.5: Equations of Lines and Planes

1.7. Partial Fractions Rational Functions and Partial Fractions. A rational function is a quotient of two polynomials: R(x) = P (x) Q(x).

Math 120 Final Exam Practice Problems, Form: A

About the Gamma Function

Inverse Functions and Logarithms

Linear Equations and Inequalities

Math 115 Spring 2011 Written Homework 5 Solutions

Solutions to Homework 10

HOMEWORK 5 SOLUTIONS. n!f n (1) lim. ln x n! + xn x. 1 = G n 1 (x). (2) k + 1 n. (n 1)!

THE COMPLEX EXPONENTIAL FUNCTION

Lecture 13 - Basic Number Theory.

LINEAR INEQUALITIES. Mathematics is the art of saying many things in many different ways. MAXWELL

Solutions for Review Problems

Gradient, Divergence and Curl in Curvilinear Coordinates

3 Contour integrals and Cauchy s Theorem

Triple Integrals in Cylindrical or Spherical Coordinates

MATH PROBLEMS, WITH SOLUTIONS

Taylor and Maclaurin Series

Let H and J be as in the above lemma. The result of the lemma shows that the integral

Course Notes for Math 162: Mathematical Statistics Approximation Methods in Statistics

A QUICK GUIDE TO THE FORMULAS OF MULTIVARIABLE CALCULUS

Estimating the Average Value of a Function

Homework # 3 Solutions

CURVATURE. 1. Problems (1) Let S be a surface with a chart (φ, U) so that. + g11 g 22 x 1 and g 22 = φ

Inner Product Spaces

y cos 3 x dx y cos 2 x cos x dx y 1 sin 2 x cos x dx y 1 u 2 du u 1 3u 3 C

Intermediate Value Theorem, Rolle s Theorem and Mean Value Theorem

M5A42 APPLIED STOCHASTIC PROCESSES PROBLEM SHEET 1 SOLUTIONS Term

Section 4.5 Exponential and Logarithmic Equations

Method of Green s Functions

QUADRATIC EQUATIONS EXPECTED BACKGROUND KNOWLEDGE

Practice problems for Homework 11 - Point Estimation

n k=1 k=0 1/k! = e. Example 6.4. The series 1/k 2 converges in R. Indeed, if s n = n then k=1 1/k, then s 2n s n = 1 n

MATH 31B: MIDTERM 1 REVIEW. 1. Inverses. yx 3y = 1. x = 1 + 3y y 4( 1) + 32 = 1

Note on some explicit formulae for twin prime counting function

Triangle deletion. Ernie Croot. February 3, 2010

Math 113 HW #7 Solutions

Algebraic and Transcendental Numbers

( ) = ( ) = {,,, } β ( ), < 1 ( ) + ( ) = ( ) + ( )

Probability and Statistics Prof. Dr. Somesh Kumar Department of Mathematics Indian Institute of Technology, Kharagpur

Math 55: Discrete Mathematics

An important theme in this book is to give constructive definitions of mathematical objects. Thus, for instance, if you needed to evaluate.

12.6 Logarithmic and Exponential Equations PREPARING FOR THIS SECTION Before getting started, review the following:

1. Let X and Y be normed spaces and let T B(X, Y ).

6. Differentiating the exponential and logarithm functions

Methods of Solution of Selected Differential Equations Carol A. Edwards Chandler-Gilbert Community College

1 Sufficient statistics

An Introduction to Partial Differential Equations in the Undergraduate Curriculum

Chapter 2: Binomial Methods and the Black-Scholes Formula

Factoring Quadratic Expressions

VERTICES OF GIVEN DEGREE IN SERIES-PARALLEL GRAPHS

MATHEMATICAL INDUCTION. Mathematical Induction. This is a powerful method to prove properties of positive integers.

First, we show how to use known design specifications to determine filter order and 3dB cut-off

Applications of Fermat s Little Theorem and Congruences

36 CHAPTER 1. LIMITS AND CONTINUITY. Figure 1.17: At which points is f not continuous?

4. Binomial Expansions

Does Black-Scholes framework for Option Pricing use Constant Volatilities and Interest Rates? New Solution for a New Problem

CHAPTER 5 Round-off errors

Section 6.1 Factoring Expressions

1.3. DOT PRODUCT If θ is the angle (between 0 and π) between two non-zero vectors u and v,

Examples of Functions

General Theory of Differential Equations Sections 2.8, , 4.1

PRACTICE FINAL. Problem 1. Find the dimensions of the isosceles triangle with largest area that can be inscribed in a circle of radius 10cm.

The Goldberg Rao Algorithm for the Maximum Flow Problem

Integrals of Rational Functions

Notes and questions to aid A-level Mathematics revision

Adding vectors We can do arithmetic with vectors. We ll start with vector addition and related operations. Suppose you have two vectors

PUTNAM TRAINING POLYNOMIALS. Exercises 1. Find a polynomial with integral coefficients whose zeros include

Manual for SOA Exam FM/CAS Exam 2.

4. How many integers between 2004 and 4002 are perfect squares?

MULTIPLE INTEGRALS. h 2 (y) are continuous functions on [c, d] and let f(x, y) be a function defined on R. Then

Dynamics at the Horsetooth, Volume 2A, Focussed Issue: Asymptotics and Perturbations

Undergraduate Notes in Mathematics. Arkansas Tech University Department of Mathematics

ECG590I Asset Pricing. Lecture 2: Present Value 1

arxiv:math/ v1 [math.co] 21 Feb 2002

3.1. Solving linear equations. Introduction. Prerequisites. Learning Outcomes. Learning Style

ELECTRIC FIELD LINES AND EQUIPOTENTIAL SURFACES

Contrôle dynamique de méthodes d approximation

Continued Fractions and the Euclidean Algorithm

Polynomials. Dr. philippe B. laval Kennesaw State University. April 3, 2005

Mathematics Course 111: Algebra I Part IV: Vector Spaces

Math 1B, lecture 5: area and volume

Properties of BMO functions whose reciprocals are also BMO

The Heat Equation. Lectures INF2320 p. 1/88

Transcription:

Stirling s formula, n-spheres and the Gamma Function We start by noticing that and hence x n e x dx lim a 1 ( 1 n n a n n! e ax dx lim a 1 ( 1 n n a n a 1 x n e x dx (1 Let us make a remark in passing. Note that in general the gamma function is defined by Γ(x e t t x 1 dt ( and hence a simple integration by parts (with u t x and dv e t dt shows that Γ(x + 1 xγ(x. (3 Since it is also easy to see that Γ(1 1, it follows that equation (1 is just n! Γ(n + 1. We will also have use for the basic result ( π 1/ e ax dx (4 a (To prove this, multiply the integral by itself (letting x y, resulting in a double integral with integrand e a(x +y. Change to polar coordinates with dxdy rdrdθ, and integrate r from to, and θ from to π. Define F x n e x so that lnf n lnx x. Then lnf has a maximum at x given by d lnf dx n x 1 ]xx so that x n. Expanding x in a neighborhood of n we write x n + ξ n(1 + ξ/n where ξ n. Using ln(1 + ε ε + ε /! + we have lnf n lnn(1 + ξ/n n(1 + ξ/n ( ξ n lnn + n n 1 ξ n n ξ n lnn n 1 ξ n 1

and therefore F n n e n e ξ /n Observe that if n is very large, then n n. For example, if n 1 4, then n 1 1 n. Also, if say, ξ 1 16, then ξ n while ξ n (or alternatively, ξ 1 3 n. Since it is also obvious that F has a maximum at ξ and is very small if ξ n, this max is very sharp. As an example, if n is only equal to 1, F(ξ n n e n 3.7 1 156, while letting ξ 5 1 gives a damping factor of e 1 ξ /n e 5/ 3.7 1 6. And for typical numbers like n 1 4 and ξ 1 16, this damping factor is on the order of the fantastically small number e 18 / 8.3 1 171475. We now want to use (5 in (1. We first change variables in (1 from x to ξ x n so that the integral becomes dx n dξ since if ξ < n, then F is very small. (Again, if n 1 4 ξ, then e 1 ξ /n e 14 / 1. Therefore, equation (1 becomes or, using (4 n! Taking the logarithm yields the formula dξ (5 n n e n e ξ /n dξ n n e n e ξ /n dξ (6 n! πn n n e n (7 lnn! n lnn n + 1 ln πn (8 Finally, we note that for n large, ln πn n, and hence the last term in (8 is completely negligible. (If n 1 4, then ln n 55. We are then left with the usual form of Stirling s equation lnn! n lnn n (9 As a measure of the accuracy of this formula, let us look at a comparison of equations (8 and (9 with the exact result: n lnn! n lnn n + (1/lnπn n lnn n 1 15.144 15.961 13.59 1 363.739 363.739 36.517 1 3 591.13 591.13 597.76 1 6 1.8155 1 7 1.8155 1 7 1.8155 1 7 It is clear that for the numbers of interest in statistical mechanics, equation (9 is quite accurate.

As another application of equations ( and (4, we derive an equation for the volume of an n-dimensional sphere of radius R. First of all, since exp( n i1 x i n i1 exp( x i, it is clear from (4 that ( n I n exp dx 1 dx n π n/. i1 x i Note that dx 1 dx n is the volume element dv n in the cartesian coordinates for R n. We now want to change to spherical coordinates where x 1 + + x n R. Since an n-dimensional volume must go like R n, we can write V n C n R n for come constant C n, and hence dv n d(c n R n nc n R n 1 dr. Then I n exp( R dv n nc n e R R n 1 dr. Now let t R so that dt R dr or dr (1/t 1/ dt. Using equations ( and (3 this gives us n I n C n e t t (n 1/ t 1/ dt n C n e t t n/ 1 dt n C n C n Γ Γ + 1. But we already showed that I n π n/ and therefore C n π n/ /Γ that the volume of the n-sphere is given by + 1 so V n πn/ Γ + 1Rn (1 Let us write this in another form. If we let t u in ( we find another useful representation for the gamma function: or Γ(x e u u x u du Γ(x e u u x 1 du e t t x 1 dt (11 From equation (4 we clearly have (since the integrand is symmetric e ax dx 1 ( π a 1/ 3

and hence Γ( 1 π1/. If n is even, then Γ + 1 isn t a problem. But if n is odd, then we observe that Γ + 1 n Γ n 1 Γ 1 n 1 Γ n ( ( ( n n 1 1 1 Γ n ( ( ( ( n n 4 n (n 1 1 Γ. Note there are (n 1/ + 1 (n + 1/ terms in the product, and hence we have Γ + 1 n!! π for n odd (1 n+1 where the double factorial is defined by n!! n(n (n 4 (1. Using equation (1, we can now rewrite equation (1 for n odd as V n n+1 π n 1 n!! R n. The double factorial can be written in another form as follows: n!! 1 3 5 7 9 (n 4 (n n 1 3 4 5 6 7 8 9 (n 4 (n 3 (n (n 1 n 4 6 8 (n 3 (n 1 1 3 4 5 6 7 8 9 (n 4 (n 3 (n (n 1 n ( 1( ( 3( 4 ( n 3 n 1 ( n! n 1 1!. Then we also have (for n odd V n n π n 1 n! 1! R n. Finally, as a general comment, for any nonnegative integer m we note that m + 1 is odd and hence (m + 1!! (m + 1! m m! m 1,, 3,... 4

As another application of these techniques, we evaluate the so-called Bose- Einstein integral I ν (z z 1 e x dx (13 1 where z 1 and ν R. Recall that if w (, 1 then 1 1 w 1 + w + w + We have for the integrand of equation (13 z 1 e x 1 xν ze x 1 1 ze x xν ze x and hence the integral becomes z 1 e x 1 dx w n. n (ze x n z n e nx n z n e nx dx z n e nx dx. Now let t nx so that dx dt/n and t ν /n ν, and therefore I ν (z z 1 e x 1 dx z n n ν+1 t ν e t dt. From equation ( we see that the integral is just Γ(ν + 1, and the sum we will denote by g ν+1 (z, i.e., g ν (z z n n ν z 1. In the particular case that z 1 we have g ν (1 ζ(ν, the Riemann zeta function. In any case, we have and, in particular, I ν (z z 1 e x 1 dx g ν+1(zγ(ν + 1 e x dx Γ(ν + 1ζ(ν + 1 1 5