# Partial Fraction Decomposition Calculator

Created by Maciej Kowalski, PhD candidate
Reviewed by Steven Wooding
Last updated: Nov 06, 2022

Welcome to Omni's partial fraction decomposition calculator, where we'll take a rational expression and write it as several simpler ones. The process can prove very useful, for instance, when computing integrals: the initial function may be tricky to deal with, but the individual summands; not so much. We'll learn all the partial fraction decomposition rules (there aren't many, don't worry) and see how to do partial fraction decomposition on a nice example.

So without further ado, let's see what this partial fraction expansion is all about!

## Polynomials, rational expressions, factorization

Polynomials are algebraic expressions that contain variables only in non-negative integer powers. In other words, they describe functions that consist of numbers, letters (i.e., variables), and basic arithmetic expressions. However, the variables cannot appear under a root, inside functions (e.g., trigonometric functions or logarithms), with fractional powers, etc.

A few examples of polynomials are:

• $x + 2y$;

• $a^2 + 2ab + b$;

• $n^3 - 0.7n + \frac{3}{8}$; and

• $1 + 3 + x^5 - x^7 + 19x^9$.

Note how there may be more than one letter in a polynomial. Nevertheless, today we'll focus only on those with a single one, just like the partial fraction decomposition calculator.

In mathematics, a quotient of two polynomials is often called a rational function or a rational expression. Needless to say, such objects are a bit more difficult to operate than simple polynomials. Just look at an example of such a monstrosity:

$\frac{3x^2 - 2x +5}{x^3 - 2x^2 - 5x + 6}$

Luckily, there are ways to write such things in a nicer way. A sum of several rational expressions, true. But prettier ones!

However, before we see how to do partial fraction decomposition, we need to go through several math properties. The very basic one concerns factoring polynomials.

When we work with real numbers (anything from $1$, through fractions, roots, up to numbers such as $\pi$ and the Euler number $\mathrm e$), every polynomial can be decomposed into factors of degree $1$ or $2$. In other words, however large the exponents of your polynomial, you can write the whole thing as a product of binomials and quadratic polynomials. For instance,

$x^4 + 4x^3 - 7x^2 - 22x + 24 \\[1em] = (x - 2)(x-1)(x + 3)(x+4)\\[2em] 2x^3 + 5x^2 + 5x + 3 \\[1em] = (2x + 3)(x^2 + x + 1)$

In fact, the quadratic factors appear (i.e., cannot be decomposed into two of order $1$) only when they have no real roots. If we were to move to complex numbers, we would obtain only binomials in the expansion. Either way, these factors are what appears in the denominators of the partial fraction formula.

In general, factoring polynomials of a high order is an extremely difficult problem. Feel free to check out Omni's algebra calculators section to find some useful tools to help with that, such as synthetic division calculator or the rational zeros calculator.

We've already mentioned what appears in the partial fraction formula, but we've yet to see it in detail. It's now time to study it in depth, together with the partial fraction decomposition rules that govern the whole thing.

## Partial fraction decomposition rules

In the above section, we've introduced factoring polynomials. However, it's important to remember that each factor can appear multiple times. For instance,

$2x^3 + x^2 - 4x - 3 \\[1em] = (2x - 3)(x+1)(x+1)\\[1em] = (2x - 3)(x+1)^2$

In general, let's say we want to find the partial fraction expansion of the quotient $P(x) / Q(x)$ of two polynomials, where $P(x)$ is of a smaller degree than $Q(x)$ (if it weren't, we first need to pull the higher exponents out of the quotient; for details see the next section). Assume that $Q(x)$ factorizes as:

$\footnotesize Q(x) \\[0.5em] =(W_1(x))^{k_1}(W_2(x))^{k_2}\ldots(W_n(x))^{k_n}$

Then the partial fraction decomposition rules for $P(x) / Q(x)$ are as follows:

• The partial fraction expansion has $k_1 + k_2 + ... + k_n$ summands.

• The denominators of the quotients take the form $(W_i(x))^t$ with $t$ ranging from $1$ to $k_i$. In other words, we have one quotient with $W_i(x)$, one with $(W_i(x))^2$, one with $(W_i(x))^3$, and so on up to $(W_i(x))^{k_i}$.

• The numerator of the summand with $(W_i(x))^t$ in the denominator is of degree one smaller than $W_i(x)$. To be precise, it can be either of degree $0$ if $W_i(x)$ is linear or of degree $1$ if $W_i(x)$ is quadratic.

All in all, the partial fraction formula we end up with looks like this:

$\footnotesize \frac{P(x)}{(W_1(x))^{k_1}(W_1(x))^{k_2} \ldots (W_n(x))^{k_n}} \\[1.5em] \!= \frac{A_{1,1}(x)}{W_1(x)} \!\! + \!\! \frac{A_{1,2}(x)}{(W_1(x))^2} \!\! + \! \ldots \! + \!\! \frac{A_{1,k_1}(x)}{(W_1(x))^{k_1}} \\[1.5em] \!+ \frac{A_{2,1}(x)}{W_2(x)} \!\! + \!\! \frac{A_{2,2}(x)}{(W_2(x))^2} \!\! + \! \ldots \! + \!\! \frac{A_{2,k_2}(x)}{(W_2(x))^{k_2}} \\[1em] \qquad \qquad \qquad \qquad \vdots \\[1em] \!+ \frac{A_{n,1}(x)}{W_n(x)} \!\! + \!\! \frac{A_{n,2}(x)}{(W_n(x))^2} \!\! + \! \ldots \! + \!\! \frac{A_{n,k_n}(x)}{(W_n(x))^{k_n}} \\[1.5em]$