Home page for accesible maths 4 Chapter 4 contents

Style control - access keys in brackets

Font (2 3) - + Letter spacing (4 5) - + Word spacing (6 7) - + Line spacing (8 9) - +

4.6 Discussion of Taylor’s theorem

Consequently, when (a,b)(a,b)\, is not a stationary point, the sum of f⁢(a,b)f(a,b)\, and the term in braces gives a good approximation to f⁢(a+h,b+k)f(a+h,b+k)\, near to (a,b)(a,b)\,. However, when (a,b)(a,b)\, is a stationary point, the term in braces is zero and f⁢(a+h,b+k)f(a+h,b+k)\, may be approximated near to (a,b)(a,b)\, by a quadratic expression in hh\, and kk\, with coefficients given by the second-order partial derivatives of ff\,.

This observation will be useful in determining whether stationary points are local maxima, local minima, or another type of stationary point called a saddle point.

Before proving it, we make the following observation, using the Chain rule: if g⁢(x,y)g(x,y) is a function and G⁢(t)=g⁢(a+t⁢h,b+t⁢k)G(t)=g(a+th,b+tk) then

G′⁢(t)=h⁢∂⁡g∂⁡x⁢(a+t⁢h,b+t⁢k)+k⁢∂⁡g∂⁡y⁢(a+t⁢h,b+t⁢k).G^{\prime}(t)=h\frac{\partial g}{\partial x}(a+th,b+tk)+k\frac{\partial g}{% \partial y}(a+th,b+tk).

Concisely, G′⁢(t)=h⁢gx+k⁢gyG^{\prime}(t)=hg_{x}+kg_{y}.