The Great Idea of Brook Taylor

1
The great idea of Brook Taylor, Taylor’s theorem
George Mpantes mathematics teacher

Mpantes_2@yahoo.gr
Taylor’s theorem
In many calculations, in applied and theoretical mathematics, it happens that a

result can not be exported directly i.e. 3/40 = 0.75, but we must approach it in
successive steps.
Where do we meet such processes?
In our daily practice, but we do not give attention.
Suppose we want to share "fairly" one euro in three parts. Each share in
decimal is
1/3 = 0,333333 ... = 3/10 +3/100 + 3/1000 + .... = 3 (1/10 +1/100 + ...) and the
parenthesis is an endless sum. We cannot find exactly each share.
The process of adding an infinite sum of numbers is the heart of the
mathematical concept of the numerical series.
We say that the value of 1/3 at 0.001 is 0.333. So we can approach it as much
as we want, and that for mathematics, this replaces precision. In this simple example
we go beyond the finite processes of elementary mathematics (algebra, and school
geometry) and capture the infinite processes in the foundations of mathematics of
continuous magnitudes.
Later in the same way we will approach functions to calculate their values.
The idea of Taylor1 was exactly this: as we approach 1/3 with decimal fractions, we
will approach functions with polynomials.
1
Taylor was brought up in a household where his father ruled as a strict disciplinarian, yet he
was a man of culture with interests in painting and music. Although John Taylor had some negative
influences on his son, he also had some positive ones, particularly giving his son a love of music and
painting. Brook Taylor grew up not only to be an accomplished musician and painter, but he applied
his mathematical skills to both these areas later in his life. In 1712 Taylor was elected to the Royal
Society. The year 1714 also marks the year in which Taylor was elected Secretary to the Royal Society.
2
Taylor’s polynomials
The calculations with unlimited series began around the end of the 17th
century along with the development of differential calculus. Since the early 18th
century, Brook Taylor has laid the foundations for "Finite Difference Calculus",
which now plays a central role in numerical analysis.
The central concept of Taylor was to use polynomials to approximate a given
function f (x). This would avoid the fragmentary encounters of previous researchers
and would refer to each function in general. This idea came to an end by studying the
findings of the Newtonian approach to differential reasoning, of which he was deeply
acquainted.
The basic idea of differential calculus is that a curve f (x) can be "locally"
approximated with its tangent line. This is because the tangent is the only line from
(α, f (α)) to the same slope as f(x) in α.

For example, we consider υ (χ) = sinx close to α = π / 4. After
(sin=π / 4)´ = cos π / 4
the tangent straight to α = π / 4 is written
y(x) = sin (π / 4) + cos(π / 4) (χ-π / 4).
In the figure that follows, the sinx is black curve, while the tangent straight
the red. Near the π / 4 the two lines are "almost" identical.
The polynomial T1 (x) = √2 / 2 + √2 / 2 (x-π / 4) is named first degree Taylor
polynomial for f(x) = sinx at α = π / 4
3
How will we define the 'appropriate' parabola which approximates f(x)

(probably better than the tangent?) Will look for a parabola that passes from the (a, f
(a), which has its (first derivative) with f (x) at α and the rate of change of the two
slopes (second derivative) in α is identical !
Let f (x) = e-(x-1) near α = 1. The parabola we are looking for will have the
formula:
p(x) = α + β (x-1) + γ (x-1)2.
Writing the parabola in this form is easier to calculate the derivatives at α = 1:
p´ (χ) = β + 2γ (χ-1) and p´´ (χ) = 2γ. Setting α = 1 we have
p(1) = α, p´(1) = β, p´´ (1) = 2γ.
From the conditions we set we have the following conditions:
f (1) = α, f´ (1) = β, f´´ (1) = 2γ and because
f (1) = 1, f´ (1) = - 1, f´´ (1) = 1
replacing in the formula for p(x) we have
p (x) = 1- (x-1) + ½ (x-1)2 = T2 (x)
is the second degree Taylor polynomial for f(x) = e-(x-1) in x0 = 1
In the figure f (x) is the black curve and the polynomial p(x) of Taylor of f (x)
at α= 1 is red.
It is easy to consider in its position 1 the α, so the general formula of the
polynomial, centered on α is
T2 (x) = f (α) + f´ (α) (χ-α) + f´´ (α)/2 .(x-α)2
4
With the same thoughts, we can build Taylor polynomials of any rank. Their
general formula is
( )
( )
( ) ( ) ∑ ( )
But how much is the error in f(χ), when approaching it with a Taylor
polynomial in (α, χ)?
The answer to this problem is given by the famous Taylor’s theorem, by
which we identify this error with enough precision and at the same time we prepare
the development of the function in power series of beyond the coincidental
development cases we have seen so far, thanks to the geometric series . 2 "If we
assume that f, f´ , f(2) ,.. .f(n), f(n+1) are all continuous in an interval containing α and x,
then there is c ϵ (α,x) that
( ) ( ) ( )
f(x)=f(α)+ ( ) f ΄΄(α)… ( ) ...(4.1)
( )( )
( )
where R n = ( )
( )
Rn is called the remainder of the Taylor-Lagrange formula.

"Historically, Taylor's proof was based on theories of differential calculus
(generalized mean value theorem, (Anastasiadis in "General Mathematics" AUTh))
Here we do not know the point c but it is in space (α, x).
So, in order to calculate the error of approaching, we must construct inequalities
for the remainder.
Example 1:
2
Articles that illuminate the work of Brook are
5
We assume that we use the approach

ex ≈ 1 + x + x2/2! + x3/3! + x4 / 4!
A. How accurate is if | x | <5?
B. how accurate is if | x | <2?
C. If | x <1?
Solution: as long as the approximate polynomial is P4 (x) with α = 0
we are investigating a limit on the remainder
( )
( )
|R4(x)| = | | | |
A. c is between 0 and x and since the exponential function is increasing, it is safe to

assume e c <e5, so e c < e5 , έτσι |R4 (x)| < 1/ 120 e 5 · 55 ≈ 3865, so we can see that
the exponential is not approached with precision by P4 in space | χ | <5
B. | x | <2: similarly we assume that x is between -2 and +2 and we assume, similarly
to the above that c <2. So we have
( )(
)
|R4(x)| = | | | | ≈1.97
C. Here we can assume that c < 1 : |R4 (x)| = 1 /120 ec |x|5 < 1/ 120 e1 · 15 ≈
0.02265.3
Example 2.
We can approximate the value of υ (x) = sinx, which has derivatives of any order with
a polynomial e.g. 7th degree in χ0 = 0 and find the value 0.3 and the error of
approximation.
We have the polynomial Taylor at x = 0
υ (0) = sin υ (0) = 0
υ´ (x) = cosx, υ´ (0) = 1
υ´´ (x) = - sinx, υ´´(0) = 0
υ´´´ (x) = - cosx, υ´´´(0) = - 1
υ(4) (χ) = sinx, υ(4) (0) = 0
3
Taylor added to mathematics a new branch now called the "calculus of finite differences",
invented integration by parts, and discovered the celebrated series known as Taylor's expansion. These
ideas appear in his book Methodus incrementorum directa et inversa. In fact the first mention by
Taylor of a version of what is today called Taylor's Theorem appears in a letter which he wrote to
Machin on 26 July 1712. In this letter Taylor explains carefully where he got the idea from.
6
υ(5) (x) = cos x υ(5) (0) = 1

υ(6) (χ) = - sinx υ(6) (0) = 0
υ(7) (x) = - cosx, υ(7) (0) = - 1
υ(8) (x) = sinx, υ(8) (ξ) = sinξ
so R7 =(x8/8!)sinξ then
sinx=x –x3/3! +x5/5! –x7/7! +(x8/8!) sinξ
and the term (x8 / 8!) sinξ is the error of the approximation of P7 i.e if x = 0.3
radians then
sin0.3=0.3- (0.3)3/3! +(0.3)5/5! –(0.3)7/7! +(0.3)8/8! sinξ
where 0 <ξ <0.3 and sinξ <1
so when calculating the sum of the first four terms of the sum we find 0.29552021,
while
(0,3)8/8!=0,00000002
hence sin0.30 = 0.29552021 with error <0.0000000002
After the examples we can understand the point
where the "transformation" of the functions occurs!
It is clear that as we increase the degree of the Taylor polynomial, both the remainder
of Rn decreases (since the approach becomes more accurate).
When (the necessary and sufficient condition) the function is
expanded in a power series known as the Taylor series.
In 1715 there are two Taylor’s books in London, Methodus incrementor directa
and inversa, and the Linear Perspective that is extremely important in the history of
mathematics.
In the first version of Methodus incrementorum, Taylor exposes for the first time the
most powerful and straightforward method of developing series functions. In a set of
27 sentences, he basically analyzes what is known as Taylor's theorem, which he
started studying since 1712. There are many great ideas in this book that were ignored
at the time of their publication. These include unique solutions to differential
equations, how to link the derivative of a function to the derivative of the inverse
function, and an analysis of the pulse chord problem that, as we shall see later, will
peak in the next century.
7
George Mpantes
www.mpantes.gr

The Great Idea of Brook Taylor

Uploaded by

Copyright:

Available Formats

The Great Idea of Brook Taylor

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Great Idea of Brook Taylor

Uploaded by

Copyright:

Available Formats

1

The great idea of Brook Taylor, Taylor’s theorem

George Mpantes mathematics teacher

In many calculations, in applied and theoretical mathematics, it happens that a

(α, f (α)) to the same slope as f(x) in α.

How will we define the 'appropriate' parabola which approximates f(x)

Rn is called the remainder of the Taylor-Lagrange formula.

We assume that we use the approach

A. c is between 0 and x and since the exponential function is increasing, it is safe to

υ(5) (x) = cos x υ(5) (0) = 1

You might also like