Foundations of calculus · Section 5 of 5
Taylor expansion
Approximating functions by polynomials, with error control.
about 9 min
Start from a problem
A calculator can only add, subtract, multiply and divide. How does it compute or ?
We need a way to approximate any smooth function by expressions using only those four operations (polynomials), and we must be able to say how large the error is; otherwise the approximation is worthless.
How does a calculator compute ? It can only add, subtract, multiply and divide. But near :
At these give , , , while the true value is . The longer the polynomial, the better the approximation.
If a polynomial is to "look like" near , the natural demand is that its value, slope, curvature, … at all agree with those of :
What did you observe?Play first, then answer. Read the explanation only after you have thought it through.
- 1.As n increases, what happens near the centre a?
- 2.As a changes, how does the region of best approximation move?
- 3.Why does raising the order not necessarily improve the approximation far away? Try 1/(1−x) at x = 1.5.
Definition
Let have derivatives at . The polynomial
is the -th Taylor polynomial of at , and is the remainder.
Why ? Differentiate times and set : of the terms , only survives, contributing . So
Theorem and proof
Polynomials are easy to evaluate, but an approximation without an error bound is meaningless. The theorem gives the error exactly.
Let have derivatives on an interval containing and . Then there is between and with
Fix and choose with
Then
By Rolle's theorem has a zero between and ; then vanishes at and , so has a zero ; after steps, . Since and
Approximate by for . This polynomial is in fact (the coefficient is zero), so
If is infinitely differentiable at , is its Taylor series. The series converges to if and only if .
All derivatives of these functions are uniformly bounded on any bounded interval (the derivative of is itself; those of have absolute value at most 1). Fix and let
because (from some point on each term is multiplied by a factor below ).
Not every Taylor series converges to its function. with has all derivatives equal to at ; its Taylor series is identically , yet is not. The remainder never tends to . The clause "if and only if " in the definition is not decoration.
A Taylor series converging is not the same as converging to . For (with ) the series converges everywhere (to ) yet equals only at . The only way to know "series function" is to prove the remainder .
Applications
In Numerical solutions, Euler's method has local error and RK4 has . All such statements come from expanding the true solution in a Taylor series and comparing it term by term with the numerical scheme.
Splitting the series of into real and imaginary parts gives
Exercises
0 of 2 done
✓ All exercises in this section are done.
Discussion
Found a mistake, have a question or a suggestion? You can also write to 620026600@qq.com
No comments yet. Share your understanding, ask a question, or point out an error on this page.
Log in to join the discussion