Scribd
Upload
67 views

Approximating π: An Approach Using Numerical Methods

The document discusses approximating the value of pi (π) using numerical methods. It first examines integral methods, showing that evaluating the integral from 0 to 1 of (x4(1-x)4)/(1+x2) dx approximates π as less than 22/7. It then considers using infinite series approximations that converge to π as more terms are included, such as the Gregory-Leibniz series and Nilakantha series. The investigation aims to find accurate approximations to π using numerical techniques.

Uploaded by

Toby Reichelt
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
67 views

Approximating π: An Approach Using Numerical Methods

The document discusses approximating the value of pi (π) using numerical methods. It first examines integral methods, showing that evaluating the integral from 0 to 1 of (x4(1-x)4)/(1+x2) dx approximates π as less than 22/7. It then considers using infinite series approximations that converge to π as more terms are included, such as the Gregory-Leibniz series and Nilakantha series. The investigation aims to find accurate approximations to π using numerical techniques.

Uploaded by

Toby Reichelt
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Approximating

An Approach Using Numerical Methods

Toby Jonathan Reichelt


Email: [email protected]

May 2017

1 Introduction

The Greek letter denotes the ratio between a circles circumference and diameter. Its value is approximately

3.14159 . . ., and it is an irrational number. It has first been approximated by Archimedes who inscribed a

regular hexagon inside a circle and then circumscribing another regular hexagon outside that circle. The

circumferences and diameters of the hexagons obtained a rough approximation of .

Another method used to approximate would be to use the Gregory-Leibniz series

1 1 1 1
1 + + ...
4 3 5 7 9


which would approximate to two decimal places when the first 300 terms were evaluated. Similarly, the
4
Nilakantha Series
4 4 4
3+ + + ...
234 456 678

converges more quickly to , giving it the upper-hand over the Gregory-Leibniz series.

The investigation of this text is to find a decent approximation to employing numerical methods where

analytical methods fail. We first investigate using integrals which, when evaluated, approximate . Then

we move onto infinite series and observe their behaviour as more terms are considered.

2 Integral Methods

A integral is simply one which approximates with good accuracy. Here we will investigate two integrals,

one using an analytical method and the other using numerical methods.

1
The first result we demonstrate is the integral

1
x4 (1 x)4
Z
22
I= 2
dx = . (1)
0 1+x 7

x4 (1 x)4
Theorem 1. The curve y = is nonnegative on the interval [0, 1].
1 + x2
x4 (1 x)4
Proof. We prove by contradiction. Suppose that < 0 for 0 6 x 6 1. Then x4 (1 x)4 < 0, hence
1 + x2
either x4 < 0 (a) or (1 x)4 < 0 (b). Clearly case (a) is not true unless x C.

Investigating (b) more thoroughly shows that (1 x) < 0 hence x > 1, which contradicts our original

assertion that 0 6 x 6 1. Therefore, the integrand is nonnegative on the interval [0, 1].

Now that we have established that integrand is nonnegative on that interval, it follows that is strictly
22
less than .
7
It is easy to prove (1). Decomposing the integrand into partial fractions gives us

x4 (1 x)4 4
= x6 4x5 + 5x4 4x2 + 4
1 + x2 1 + x2

and integrating gives us

Z 1
4 1 2 4
x6 4x5 + 5x4 4x2 + 4 dx = + 1 + 4 4
0 1 + x2 7 3 3 4

from which (1) immediately follows.


22
Since now we have established that < we can use (1) to establish the error bounds on this approx-
7
imation.

You might also like