Let me bring out a very old question first:
$$ \displaystyle {f\left(x\right)=\frac{x^{2}+3}{x^{6}\left(x^{2}+1\right)}} $$
(Source: Evaluating the rational integral $\int \frac{x^2+3}{x^6(x^2+1)}dx $)
I think by now everyone knows that we can rewrite the function as:
$$ f\left(x\right)=\frac{a_{2}}{x^{2}}+\frac{a_{4}}{x^{4}}+\frac{a_{6}}{x^{6}}+\frac{bx+c}{x^{2}+1} $$
By using the Heaviside cover-up method, it is easy to obtain $a_{6}$, $b$ and $c$ (which is $3$, $0$ and $-2$ respectively). However, when it comes to other coefficients, it is (almost) impossible to use this method. (No, I do not want to use other methods like subtracting the found fractions from $f(x)$, when it comes to other very long scenarios, it can be very difficult to simplify). So, I started searching for an alternative method that at least a handheld calculator could solve quickly, and I found this formula for those coefficients.
$$a_{k}=\frac{1}{\left ( k-1 \right )!}\left\{ \frac{d^{k-1}}{dx^{k-1}} \left [ \left ( x+x_{0} \right )^{n}f\left ( x \right ) \right ]\right\}\left|\begin{matrix}\\_{x=-x_{0}}\end{matrix}\right. (2 \leq k \leq n-1)$$
Source:
- https://lpsa.swarthmore.edu/BackGround/PartialFraction/RootsRepeat.html
- https://www.youtube.com/watch?v=myERv3-Lqz0
- https://www.iaeng.org/publication/WCE2008/WCE2008_pp978-980.pdf
My question is "How can anyone come up with this???" Surely, it makes sense for the highest order since you just multiply everything with that term and calculate the limit, but it doesn't make sense how differentiation can lead you straight up to the coefficient of lower orders. For example, in the problem I proposed initially, let's say I want to find $a_{4}$ first. First, multiply everything by $x^{6}$, then I try to differentiate the RHS, twice, plug in $x = 0$, and all fractions magically equal 0, except for that specific $a_{4}$ coefficient. How does this happen?
(should've made it a instead of A)
– TargetVN Jul 24 '24 at 04:59