Taylor's Inequality                                            

 

 

In a Nut Shell:  Use Taylor's Inequality  to estimate the accuracy of a Taylor's series

expansion of a function.

 

 

Taylor's Inequality: If  | f(n+1) (x) | ≤  M  for  | x ˗ a |    d,  then the remainder

Rn(x) of the Taylor series satisfies the inequality

 

    | Rn(x) |      M / ( n + 1)!  | x ˗ a | n + 1    for  |  x ˗ a | ≤ d

 

 

Example:  a.  Find the Taylor series expansion of f(x) = √ x  for  a = 4,  n = 2  where

4  ≤ x    4.1 .        b.  Estimate the accuracy to six decimal places.  c.  Check result by

graphing  | R2(x) | .

 

    f(x)  =  x  1/2  ,  f(4)  =  2,                               f ' (x)  =  1/2 x ˗ 1/2   ,  f ' (4)  =  1/4

 

   f '' (x)  =  ˗ 1/4 x ˗ 3/2   ,  f '' (4)  =   1 / 32 ,      f ''' (x)  =   3/8 x ˗ 5/2   ,  f ''' (4)  =   3/256

 

 T2(x)  =  2  +  (1/4) (x ˗ 4)  +  (˗ 1 / 32) ( x ˗ 4)2 / (2!)      (result for part a)

 

    | R2(x) |      M / ( 2 + 1)!  | x ˗ 4 | 2+ 1    =   M / (3!)  | x ˗ 4 |3   = 

 

  Pick the largest value of M in 4  ≤ x    4.1,  M  =  |  3/256 |   and  | x ˗ 4 |  =  0.1

 

     | R2(x) |      ( 3/256) / ( 3!)  | 0.1 | 3   =  0.00000195  =  0.000002  (result for part b)

 

Finally graph the function    (part c)      | R2(x) |  =   | √ x ˗  T2(x) |    for  4    x  4.1

 

                            

 

 

From the above graph, the maximum error to six decimal places for the second order

Taylor series expansion of  f(x)  is   R2(4.1) = 0.000001923.    (result for part c)

Click here for another example.

 

 

 




Copyright © 2017 Richard C. Coddington

All rights reserved.