Irrational numbers

This article page is a stub, please help by expanding it.

Irrational numbers are numbers which can't be expressed as the ratio of two integers (not rational numbers); in other words they are not the root of any linear polynomial, i.e. not algebraic numbers of degree one. Irrational numbers are either

• irrational algebraic numbers (a root of a polynomial of minimal degree two or higher, i.e. algebraic numbers of degree two or higher); or
• transcendental numbers (a root of an infinite power series).

Transcendental numbers are obviously irrational; most (uncountably many:

2 ℵ0

) irrational numbers are transcendental while only countably many,

ℵ0

, are algebraic.

Rational approximations of irrational numbers

Rational numbers can be used to approximate irrational numbers. The best rational approximations of a number are obtained from the convergents from simple continued fractions.

A theorem of Hurwitz, improving on earlier work by Dirichlet and Vahlen, states that for any irrational number

ξ

, there are infinitely many rational approximations

m /n

with

$\left|\xi -{\frac {m}{n}}\right|<{\frac {1}{n^{2}{\sqrt {5}}}}$

and this theorem is sharp in the sense that

2√  5

cannot be replaced with a larger number, nor can the exponent

2

be replaced with a larger number (even allowing an arbitrarily small positive number in place of

2√  5

). However, by omitting certain classes of algebraic numbers (such as the golden ratio

φ

), the constant can be improved to

2√  9 − 4 / [A002559(n)] 2

. For example, for any irrational number

ξ

not of the form

${\frac {a\varphi +b}{c\varphi +d}},\quad ad-bc=\pm 1,$

there are infinitely many rational approximations

m /n

with

$\left|\xi -{\frac {m}{n}}\right|<{\frac {1}{n^{2}{\sqrt {8}}}}.$ For this reason

φ

is sometimes considered "the most irrational number": the partial denominators of its simple continued fraction being

1

makes it the worst case for approximation by convergents.

Irrationality of a number

The irrationality of a given number is not always known for certain. Since the time of Pythagoras, it has been known that

2√  2

is irrational, while it wasn't until the 18^th century that it was proved that

e

and

π

are irrational (and transcendental), the 20^th century for Apéry's constant

ζ (3)

, and the rationality of the Euler-Mascheroni constant

γ

is an open problem.