I have a function $g(x)=f(x)e^{-x}$ and i want to consider the following integral: $\int_{0}^{\infty}g(x)dx$.

Since $f(x)$ is a complicated, but monotonic decreasing, function in the interval $[0,\infty[$ and it is multiplied by $e^{-x}$, the integrand $g(x)$ is a very rapidly decreasing function.

If I consider $T_{n}(x)$ a Taylor expansion of the function $f(x)$ around $x=0$ truncated at a certain order $n$, the integral $\int_{0}^{\infty}T_{n}(x)e^{-x}dx$ converges for every $n$.

What is the error committed in approximating $\int_{0}^{\infty}g(x)dx$ with $\int_{0}^{\infty}T_{n}(x)e^{-x}dx$?

and what would it be if $g(x)=f(x)\cdot x(e^{-x})$?

thanks in advance



If you have the Taylor expansion $$f(x)=\sum_{i=0}^n a_i x^i$$, you then have $$\int_0^\infty f(x)e^{-x}dx=\sum_{i=0}^n a_i i!$$ since $\int_0^\infty x^n e^{-x}=n!$.

I am sure that you can take from here.

  • $\begingroup$ if i consider the error $E$ as: $ E = \int_{0}^{\infty} f(x) e^{-x} dx - \sum_{i=0}^n a_i i!$ rewriting a little the expression and using the following equivalence: $\int_{0}^{\infty} f(x) e^{-x} dx = \lim_{N \rightarrow \infty} \sum_{k=0}^N a_k k!$ I can rewrite the total error as: $E = 1 -\lim_{N \rightarrow \infty} (\frac{\sum_{i=0}^n a_i i!}{\sum_{k=0}^N a_k k!})$ now of course this is a decreasing function of $n$ and this should be compared with the value of the defined integral $\int_{0}^{\infty}g(x)$ is this correct? $\endgroup$
    – SSC Napoli
    Sep 23 '14 at 8:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.