How to integrate using partial fractions
The goal is to decompose f (x) = P (x)/Q(x) into a sum of simple rational
expressions with computable antiderivatives.
1. If degP (x) ≥ degQ(x) then use long division to write
R(x)
f (x) = S(x) +
Q(x)
and degR(x) < degQ(x).
2. Write Z Z Z
R(x)
f (x)dx = S(x)dx + dx.
Q(x)
The first integral is easily solved because S(x) is a polynomial. It
R R(x)
remains to compute Q(x) dx.
3. Decompose Q(x) into linear terms ax+b and irreducible ( b2 −4ac < 0)
quadratic terms ax2 + bx + c.
4. Seperate the distinct linear and quadratic terms and get a product
of those terms (possibly with repeats). Use this decomposition for
the Ansatz with summands corresponding to the distinct linear and
quadratic factors (possibly with repeats) of the following form
(a) For each factor (ax + b)r (so it comes with r repeats) in Q(x):
A1 A2 Ar
+ 2
+ ... +
ax + b (ax + b) (ax + b)r
(b) For each factor (ax2 +bx+c)s (so it comes with s repeats) in Q(x):
B1 x + C1 B 2 x + C2 B s x + Cs
2
+ 2 2
+ ... +
ax + bx + c (ax + bx + c) (ax2 + bx + c)s
5. Find the common denominator, and compare corresponding coeffi-
cients (these are linear expressions in the Ai , Bj , Ck ) with the coeffi-
cients of R(x). Solve the resulting system of linear equations for the
coefficients in the Ansatz. (Hint: The total number of coefficients to
be determined is equal to the degree of Q(x).)
R R(x
6. Write Q(x) dx as a sum of integrals using the partial fraction decom-
position and apply:
Z
dx 1
= ln|ax + b| + C
ax + b a
resp. for i > 1
Z
dx 1 1 1
i
= +C
(ax + b) a 1−i (ax + b)i−1
For the integration of the quadratic terms use
Z Z Z
(Ax + B)dx A (2ax + b)dx Ab dx
2 i
= 2 i
+(B− )
(ax + bx + c) 2a (ax + bx + c) 2a (ax + bx + c)i
2
and Z
(2ax + b)dx
= ln|ax2 + bx + c| + C
ax2 + bx + c
resp. for i > 1
Z
(2ax + b)dx 1 1
2 i
= +C
(ax + bx + c) 1 − i (ax + bx + c)i−1
2
dx
R
For the remaining integrals of the form (ax2 +bx+c) i you can assume
a > 0, use quadratic completion in the denominator (ax2 + bx + c =
√ b2 √
( ax + 2√b a )2 + (c − 4a )), substitution (u = ax + 2√b a ) and finally
for i = 1 Z
du 1 u
2 2
= tan−1 ( ) + C.
u +d a a
b2
(d2 = c − 4a > 0 since b2 − 4ac < 0). For the higher order terms with
i > 1 the resulting integrals have the form
Z
du
,
(u + d2 )i
2
use trigonometric substitutions and Good Luck (Try what a CAS tells
you about these integrals! Look into a good integration table!)