Taylor Expansion
in Several Real Variables

This page is a sub-page of our page on Calculus of Several Real Variables.

///////

Related KMR-pages:

Differentiation and Affine Approximation
Taylor Expansion in One Real Variable

///////

Other relevant sources of information:

Taylor’s theorem
The multinomial theorem
Pascal’s simplex

James Gregory
Colin Maclaurin
Brook Taylor

///////

The interactive simulations on this page can be navigated with the Free Viewer
of the Graphing Calculator.

///////

Anchors into the text below:

http://kmr.csc.kth.se/wp/research/math-rehab/learning-object-repository/calculus/calculus-of-several-real-variables/taylor-expansion-in-3d/

Differentials of higher order
History of Taylor’s theorem
Taylor’s theorem for multivariate functions
Multi-index notation
The Multinomial theorem
Taylor’s formula with remainder term
The Taylor series
General Leibniz’ rule
Taylor expansions in visual and interactive form
The Taylor polynomial of degree 1 for the function f(x,y) at the point (a,b)
The Taylor polynomial of degree 2 for the function f(x,y) at the point (a,b)
The Taylor polynomial of degree 3 for the function f(x,y) at the point (a,b)
The Taylor polynomial of degree 4 for the function f(x,y) at the point (a,b)
The Taylor polynomials of degrees 1 and 2 for the function f(x,y) at the point (a,b)
The Taylor polynomials of degrees 1 and 3 for the function f(x,y) at the point (a,b)
The Taylor polynomials of degrees 1 and 4 for the function f(x,y) at the point (a,b)
The Taylor polynomials of degrees 1, 2, 3 for the function f(x,y) at the point (a,b)
The Taylor polynomials of degrees 1, 2, 3, 4 for the function f(x,y) at the point (a,b)

Differentials of higher order

Let \, X \, and \, Y \, be two metric, linear spaces
and let \, f : X \rightarrow Y \, be an \, n \, times differentiable function.

Then \, L(X, Y) \, is a linear space if we define
\, (A + B) x \, \stackrel {\mathrm{def}}{=} \, Ax + Bx and \, (\alpha A) x \, \stackrel {\mathrm{def}}{=} \, \alpha A x .

Moreover, \, L(X, Y) \, is a metric space if we define
\, || A || \, \stackrel {\mathrm{def}}{=} \, \underset{|x|=1}{\mathrm{sup}} | A x | \, and \, d(A, B) \, \stackrel {\mathrm{def}}{=} \, || A - B || .

It follows that the function \, X \xrightarrow[]{f'} L(X, Y) \, is a mapping between two metric, linear spaces. Now, \, f' \, is differentiable at the point \, x \in X \, if there exists a linear function \, (f')'(x) \in L(X, L(X,Y)) \, such that

\, f'(x + h) = f'(x) + (f')'(x) h + o(h) .

Theorem: \, L(X, L(X,Y)) \, \simeq \, L_2(X,Y) \, where
\, L_2(X,Y) = \{ \mathrm{bilinear \, maps} : X × X \rightarrow Y \} .

That \, g : X × X \rightarrow Y \, is bilinear means that
\, g({\lambda}_1 \, a_1 + {\lambda}_2 \, a_2 \, , b) = {\lambda}_1 \, g(a_1, b) + {\lambda}_2 \, g(a_2, b) \, and
\, g(a \, , {\lambda}_1 \, b_1 + {\lambda}_2 \, b_2) = {\lambda}_1 \, g(a, b_1) + {\lambda}_2 \, g(a, b_2) .

Hence we have \, f''(x) \in L_2(X,Y) .

Theorem: If \, x \mapsto f''(x) \, is continuous, then we have
\;\;\;\;\;\;\;\;\;\;\;\;\; f''(x)(h_1, h_2) = f''(x)(h_2, h_1) .

Repeating the argument,
it follows that the function \, X \xrightarrow[]{f''} L(X, L(X, Y)) \, is a mapping between two metric, linear spaces. Now, \, f'' \, is differentiable at the point \, x \in X \, if there exists a linear function \, (f'')'(x) \in L(X, L(X, L(X,Y))) \, such that

\, f''(x + h) = f''(x) + (f'')'(x) h + o(h) .

And so on … for as many times (assumed to be \, n \, above) that the original function \, f \, is differentiable.

NOTE: It is important to remember that in these differentiations the point \, x \, is always held constant, whereas \, h \, provides the variability (in terms of small variations around the fixed point \, x . In general, the functions \, f, f', f'', f''', \cdots, f^{(n)} \, are non-linear, whereas the functions \, f(x), f'(x), f''(x), f'''(x), \cdots \, , f^{(n)}(x) \, are respectively constant, linear, bilinear, trilinear, … , n -linear.

History of Taylor’s theorem

/////// Quoting Wikipedia (on Taylor’s theorem):

In calculus, Taylor’s theorem gives an approximation of a k-times differentiable function around a given point by a polynomial of degree k, called the kth-order Taylor polynomial. For a smooth function, the Taylor polynomial is the truncation at the order k of the Taylor series of the function. The first-order Taylor polynomial is the linear approximation of the function, and the second-order Taylor polynomial is often referred to as the quadratic approximation.[1] There are several versions of Taylor’s theorem, some giving explicit estimates of the approximation error of the function by its Taylor polynomial.

Taylor’s theorem is named after the mathematician Brook Taylor, who stated a version of it in 1715,[2] although an earlier version of the result was already mentioned in 1671 by James Gregory.[3]

Taylor’s theorem is taught in introductory-level calculus courses and is one of the central elementary tools in mathematical analysis. It gives simple arithmetic formulas to accurately compute values of many transcendental functions such as the exponential function and trigonometric functions. It is the starting point of the study of analytic functions, and is fundamental in various areas of mathematics, as well as in numerical analysis and mathematical physics. Taylor’s theorem also generalizes to multivariate and vector valued functions.

/////// End of Quote from Wikipedia

Taylor’s theorem for multivariate functions

///////

Multi-index notation

\, {\alpha} \, \stackrel {\mathrm{def}}{=} \, ({\alpha}_1, {\alpha}_2, ... \, , {\alpha}_n) \; , \; {\alpha}_i ≥ 0 \,

\, | {\alpha} | \, \stackrel {\mathrm{def}}{=} \, {\alpha}_1 + {\alpha}_2 + ... + {\alpha}_n \,

\, {\alpha} ! \, \stackrel {\mathrm{def}}{=} \, ({\alpha}_1 !)({\alpha}_2 !) ... ({\alpha}_n !) \,

\, \alpha + \beta \, \stackrel {\mathrm{def}}{=} \, ({\alpha}_1 + {\beta}_1, ... \, , {\alpha}_n + {\beta}_n) \,

\, \alpha ≤ \beta \, means that \, {\alpha}_i ≤ {\beta}_i \, for all \, i .

///////

The Multinomial theorem:

\, (y_1 + y_2 + ... + y_n)^m \, = \, \sum\limits_{| \alpha | = m}^{ } \dfrac{ m !}{( {\alpha}_1 !) \cdot ... \cdot ( {\alpha}_n !) } \, {y_1}^{{\alpha}_1} \cdot ... \cdot {y_n}^{{\alpha}_n} \,

///////

Taylor’s formula with remainder term:

\, \mathbb{R} \xrightarrow[]{\phi} {\mathbb{R}}^n \xrightarrow[]{f} \mathbb{R} \, gives \, \mathbb{R} \xrightarrow[]{\psi} \mathbb{R} \, where \, \psi = f \circ \phi .

The chain rule gives

\, {\psi}'(t) = f'( \phi(t)) \circ {\phi}'(t) = f'( x+th) h. \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; (*)

Taylor’s formula on \, \psi gives

\, \psi(1) = \sum\limits_{\nu = 0}^{N-1} \, \frac{1}{\nu !} {\psi}^{(\nu)}(0) + \frac{1}{N !} {\psi}^{(N)}(\theta) \; , \; 0 \; < \theta \, < 1 .

Repeated use of \, (*) \, followed by applying the multinomial theorem gives

\, {\psi}^{(\nu)}(t) \, = \, \sum\limits_{i_1 = 1}^{n} \, \sum\limits_{i_2 = 1}^{n} ... \sum\limits_{i_{\nu} = 1}^{n} \, f_{i_1 i_2 ... i_{\nu}}(x+th) \, h_{i_1} h_{i_2} ... \, h_{i_{\nu}} \, =

\;\;\;\;\;\;\;\;\;\;\;\;\, = \, (h_1 {\partial}_1 + ... + h_n {\partial}_n)^{\nu} f|_{x+th} \, = \, d^{ \, \nu} f|_{x+th} .

Summing up, we arrive at Taylor's formula from \, {\mathbb{R}}^n \, to \,\mathbb{R} \, :

\, f(x + h) \, = \, \sum\limits_{\nu = 0}^{N-1} \, \frac{1}{\nu !} \, d^{ \, \nu} f|_x + \frac{1}{N !} \, d^{ \, N} f|_{x+ \theta h} .

///////

The Taylor series:

\, f(x+h) = e^d f|_x \,

\, f(x+h) \, = \, \sum\limits_{\alpha = 0}^{\infty} \, \frac{1}{\alpha !} {f}^{( \alpha)}(x) h^{\alpha} \, = \, \sum\limits_{\alpha = 0}^{\infty} \, \frac{1}{\alpha !} h^{\alpha} {(\frac{d}{dx})}^{\alpha}f(x) .

\, f(x+h) \, = \, \sum\limits_{\alpha = 0}^{\infty} \, \frac{1}{\alpha !} h^{\alpha} {\partial}^{\alpha} f(x) \, .

NOTE: \, {\partial}^{ \, \alpha} {\partial}^{ \, \beta} f \, = \, {\partial}^{ \, \alpha + \beta} f .

One variable:

\, x \,
\, h \,
\, \partial \stackrel {\mathrm{def}}{=} \frac{d}{dx} \,
\, d \stackrel {\mathrm{def}}{=} h \cdot \partial = h \frac{d}{dx} \,

\, f(x+h) = e^{h \frac{d}{dx}} f(x) = \sum\limits_{\nu = 0}^{\infty} \frac{1}{\nu !} (h \frac{d}{dx})^{\nu} f|_x \,

Several variables:

\, x \stackrel {\mathrm{def}}{=} (x_1, x_2, ... \, , x_n) \,
\, h \stackrel {\mathrm{def}}{=} (h_1, h_2, ... \, , h_n) \,
\, \partial \stackrel {\mathrm{def}}{=} ({\partial}_1, {\partial}_2, ... \, , {\partial}_n) \,
\, d \stackrel {\mathrm{def}}{=} h \cdot \partial = h_1 {\partial}_1 + h_2 {\partial}_2 + ... + h_n {\partial}_n \,

\, f(x+h) = e^{(h_1 {\partial}_1 + ... + \, h_n {\partial}_n)} f(x) = \sum\limits_{\nu = 0}^{\infty} \frac{1}{\nu !} (h_1 {\partial}_1 + ... + h_n {\partial}_n)^{\nu} f|_x \,

///////

General Leibniz' rule:

\, {\partial}^{ \, \alpha} f g \, = \, \sum\limits_{\beta + \gamma = \alpha}^{} \frac{\alpha !}{\beta ! \, \gamma ! } \, {\partial}^{ \, \beta} f \, {\partial}^{ \, \gamma} g \, = \, \sum\limits_{ \beta ≤ \alpha}^{} \binom{\alpha}{\beta} \, {\partial}^{ \, \beta} f \, {\partial}^{ \, \alpha - \beta} g \,

///////

Taylor expansion in visual and interactive form

The Taylor polynomial of degree 1 for the function \, f(x, y) \, at the point \, (a, b) \, :
(also known as the Affine Approximation):

{T_{aylor}^1 (f)}_{(a,b)} = f(a,b) + \frac{\partial f}{\partial x}_{(a,b)} (x-a) + \frac{\partial f}{\partial y}_{(a,b)} (y-b)

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Interactive simulation of Taylor expansion of order 1.
Drag the red point in the left window to move the point of expansion.

///////

The Taylor polynomial of degree 2 for the function \, f(x, y) \, at the point \, (a, b) \, :

{T_{aylor}^2(f)}_{(a,b)} = {T_{aylor}^1 (f)}_{(a,b)} + \dfrac{1}{2} (\frac{\partial^2 f}{\partial x^2}_{(a,b)} (x-a)^2 + 2 \frac{\partial^2 f}{\partial x \partial y}_{(a,b)} (x-a)(y-b) + \frac{\partial^2 f}{\partial y^2}_{(a,b)} (y-b)^2)

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Interactive simulation of Taylor expansion of order 2.
Drag the red point in the left window to move the point of expansion.

///////

The Taylor polynomial of degree 3 for the function \, f(x, y) \, at the point \, (a, b) \, :

{T_{aylor}^3(f)}_{(a,b)} = {T_{aylor}^2 (f)}_{(a,b)} + \dfrac{1}{6} (\frac{\partial^3 f}{\partial x^3}_{(a,b)} (x-a)^3 + 3 \frac{\partial^3 f}{\partial x^2 \partial y}_{(a,b)} (x-a)^2 (y-b) + 3 \frac{\partial^3 f}{\partial x \partial y^2}_{(a,b)} (x-a)(y-b)^2 + \frac{\partial^3 f}{\partial y^3}_{(a,b)} (y-b)^3)

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Interactive simulation of Taylor expansion of order 3.
Drag the red point in the left window to move the point of expansion.

///////

The Taylor polynomial of degree 4 for the function \, f(x, y) \, at the point \, (a, b) \, :

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Interactive simulation of Taylor expansion of order 4.
Drag the red point in the left window to move the point of expansion.

///////

The Taylor polynomials of degrees 1 and 2 for the function \, f(x, y) \, at the point \, (a, b) \, :

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Gammal:Interactive simulation of Taylor expansion of orders 1 and 2.

///////

The Taylor polynomials of degrees 1 and 3 for the function \, f(x, y) \, at the point \, (a, b) \, :

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

///////

The Taylor polynomials of degrees 1 and 4 for the function \, f(x, y) \, at the point \, (a, b) \, :

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

///////

The Taylor polynomials of degrees 1, 2, 3 for the function \, f(x, y) \, at the point \, (a, b) \, :

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Interactive simulation of Taylor expansion of orders 1, 2 and 3.

///////

The Taylor polynomials of degrees 1, 2, 3, 4 for the function \, f(x, y) \, at the point \, (a, b) \, :

The interactive simulation that created this movie.

Den interaktiva simulering som skapade denna film.

Interactive simulation of Taylor expansions of orders 1, 2, 3, and 4.

///////

Leave a Reply