You are on page 1of 3

Taylors Polynomials

Dimitrios Mitsotakis February 4, 2013


We would like to approximate functions using polynomials since we can make computations easier with them. One way to nd good approximations of functions is by using Taylors polynomials. These polynomials are identical to the function on one point x = c and very close to the function for x near c. As x is away from c the polynomial diverges from f . Consider a smooth function f C n+1 [a, b] and let c [a, b], then f (x) = f (c) + f (c)(x c) + f (c) f (c) f (n) (c) (x c)2 + (x c)3 + + (x c)n + Rn (x), 2! 3! n!

where f (k) denotes the k-th derivative of f , Rn is the residual Rn (x) = for some [a, b]. In other words we can write the function f (x) = Pn (x) + Rn (x) where Pn (x) =
k=0 n

f (n+1) () (x c)n+1 (n + 1)!

f (k) (c) (x c)k k! (x a)n+1 (n + 1)!

and Rn (x) = f (n+1) () the residual that is expected to be small for x near c. Now we will see some applications:

Taylors approximation for n = 1 is f (x) f (c) + f (c)(x c). Taylors approximation for n = 2 is f (x) f (c) + f (c)(x c) +
f (c) 2! (x

c).

By taking in the Taylors polynomial x + h instead of x and x instead of c we have f (x) 2 f (x) 3 f (n) (x) n h + h + + h + Rn () 2! 3! n! f (n+1) () n+1 h Chn+1 (n + 1)!

f (x + h) = f (x) + f (x)h + where

|Rn ()| =

The last expression of the Taylors polynomial is the most popular in numerical analysis since usually in the discrete level two successive steps dier by h and thus they are x and x + h. This formula also provides information about the approximation error. If for example the discretization step is h then the error approximating f by Pn is of O(hn+1 ). Therefore if h 1 and n suciently large we expect the error to be small. An application of Taylors polynomial is the approximation of the derivative f (x) by nite dierences. Specically, consider the 1st order Taylors polynomial: f (x + h) = f (x) + f (x)h + O(h2 ). Solving with respect to f (x) we have that f (x + h) f (x) + O(h). f (x) = h Thus, f (x + h) f (x) f (x) . h Some very popular Taylor expansions are the following:

ex = 1 + x + sin(x) = x cos(x) = 1

xn x2 x3 + + = 2! 3! n! n=0 x3 x5 (1)n 2n+1 + = x 3! 5! (2n + 1)! n=0 (1)n 2n x4 x2 + = x 2! 4! (2n)! n=0


The Taylors polynomials of order 2, 4 and 8 approximating the function f (x) = cos(x) is presented. We observe that the approximation is better as the order of the polynomial increases.

15

cos(x)
10

P 2 (x) P 4 (x) P 8 (x)

10

15 5

You might also like