An archive of Mark's Spring 2018 Numerical Analysis course.

A Taylor approximation

mark

Suppose we write the estimate

e^3 \approx \sum_{k=1}^{5} \frac{3^k}{k!}.

Use Taylor’s remainder theorem to provide an upper bound for the error in that estimate.

Cornelius

Recall that the remainder from the Taylor Theorem is:
\frac{f^{(n+1)}(\xi)}{(n+1)!}(x-x_0)^{n+1}.

n=5 and x=3 and x_0=0

and f(x)=e^x so f^6(x)=e^x

So \frac{e^{(\xi)}}{(6)!}(3)^{6}. Since the largest possible value for \xi is 3, the upper bound for the error is:
\frac{e^{(3)}}{(6)!}(3)^{6}

Lorentz

Taylor’s remainder theorem states that |T_n(x)-f(x)|=|R_n(x)| \leq \frac{|x-x_0|^{n+1}}{(n+1)!} max (f^{n+1}(\xi)), for some \xi \in [x_0,x]=[0,3]. Since we are expanding about the origin x_0=0 also f^{n+1}(x)=e^x for all n. This simplifies our expression R_n(x)\leq \frac{|x|^{n+1}}{(n+1!)}e^{\xi}, the value of \xi which maximizes this function is \xi = 3. So R_5(3) \leq \frac{3^6}{6!}e^3 \approx e^3. If this is correct than you would need many more terms to be sure that the error was negligible the actual error is 2.68

mark

Looks awesome, though some folks here don’t have a clue. :frowning:

mark