MathAnalysisCalculusIntegrals

Partial Fraction Expansion

8 minutes read

The process of finding antiderivatives can feel a bit daunting as integrands become more elaborate. Luckily there are certain techniques that can help us identify and properly tackle more intricate indefinite integrals. In this topic, we'll take a look at the partial fraction expansion method, and how it can help us solve integrals with integrands comprised of rational functions.

Partial Fraction Expansion

This method works best for integrals of the form

P(x)Q(x) dx\int \frac{P(x)}{Q(x)} \ dxwhere both PP and QQ are polynomials.

If the degree of PP is greater or equal than the degree of QQ, we need to perform polynomial division so that

P(x)Q(x)=C(x)+R(x)Q(x)\frac{P(x)}{Q(x)} = C(x) + \frac{R(x)}{Q(x)}where CC and RR are also polynomials, and the degree of RR must be less than the degree of QQ.

Once we have a proper rational expression where the degree of the numerator is less than the degree of the denominator, we can perform the partial fraction expansion as follows:

  1. Factor the denominator into linear factors (and/or irreducible quadratic factors)
  2. Make the rational expression equal to a sum of terms, each a partial fraction that corresponds to each factor in the denominator like so:
Factor in denominator Corresponding partial fraction
ax+bax+b Aax+b\frac{A}{ax+b}
ax2+bx+cax^2 +bx + c Ax+Bax2+bx+c\frac{Ax +B}{ax^2 +bx + c}
(ax+b)n(ax+b)^n A1ax+b+A2(ax+b)2++An(ax+b)n\frac{A_1}{ax+b} + \frac{A_2}{(ax+b)^2} + \cdots + \frac{A_n}{(ax+b)^n}
(ax2+bx+c)n(ax^2 +bx + c)^n A1x+B1ax2+bx+c+A2x+B2(ax2+bx+c)2++Anx+Bn(ax2+bx+c)n\frac{A_1 \cdot x+B_1}{ax^2 +bx + c} + \frac{A_2 \cdot x+B_2}{(ax^2 +bx + c)^2} + \cdots + \frac{A_n \cdot x+B_n}{(ax^2 +bx + c)^n}

where A,B,Ai,BiA, B, A_i, B_i are all constants that can be determined by comparison with the original expression. Afterwards, we are left with several much simpler integrals, usually with direct antiderivatives.

Factoring the denominator

Let's make this a bit less abstract with an example. Consider the following integral:

I=9x2+3x2x3+2x22x+3 dxI = \int \frac{9x^2 + 3x - 2}{x^3 + 2x^2-2x+3} \ dx

We already have a proper rational expression here, since the degree of the denominator is 3 and the degree of the numerator is 2. So, we move on to factoring the denominator,

I=9x2+3x2(x+3)(x2x+1) dxI = \int \frac{9x^2 + 3x - 2}{(x+3)(x^2 -x + 1)} \ dx

Obtaining the partial fraction expansion

Once the denominator is factored into linear factors and irreducible quadratic expressions, we can construct the integrand's partial fraction expansion as follows

9x2+3x2(x+3)(x2x+1)=Ax+3+Bx+Cx2x+1\frac{9x^2 + 3x - 2}{(x+3)(x^2 -x + 1)} = \frac{A}{x+3} + \frac{Bx+C}{x^2 -x + 1}In order to determine the constants A,B,CA, B, C, let's first get rid of the denominators by multiplying both sides by it:

9x2+3x2=Ax+3(x+3)(x2x+1)+Bx+Cx2x+1(x+3)(x2x+1)9x2+3x2=A(x2x+1)+(Bx+C)(x+3)9x2+3x2=Ax2Ax+A+Bx2+Cx+3Bx+3C\begin{align*} 9x^2 + 3x - 2 &= \frac{A}{\cancel{x+3}} \cdot \cancel{(x+3)} \cdot(x^2 -x + 1) + \frac{Bx+C}{\cancel{x^2 -x + 1}} \cdot (x+3) \cdot (\cancel{x^2 -x + 1}) \\ 9x^2 + 3x - 2 &= A \cdot(x^2 -x + 1) + (Bx+C)\cdot(x+3) \\ 9x^2 + 3x - 2 &= Ax^2 -Ax + A + Bx^2+Cx + 3Bx+ 3C \end{align*}Now, we can compare the two expressions on both sides

9x2+3x2=(A+B)x2+(A+C+3B)x+(A+3C)    {9=A+B3=A+3B+C2=A+3C9x^2 + 3x - 2 = (A+ B) \cdot x^2+ (-A + C + 3B) \cdot x + (A + 3C) \implies \begin{cases} 9 = A +B \\ 3 = -A + 3B +C \\ -2 = A + 3C \end{cases}After solving that system of equations, we are left with

A=7013B=4713C=3213A = \frac{70}{13} \qquad B = \frac{47}{13} \qquad C = -\frac{32}{13}

Finishing off using previous methods

Then,

I=9x2+3x2(x+3)(x2x+1) dx=(70/13x+3+(47/13)x(32/13)x2x+1) dxI=70131x+3 dx+4713xx2x+1 dx32131x2x+1 dx\begin{align*} I = \int \frac{9x^2 + 3x - 2}{(x+3)(x^2 -x + 1)} \ dx &= \int \left( \frac{70/13}{x+3} + \frac{(47/13)x-(32/13)}{x^2 -x + 1} \right) \ dx \\ I &= \frac{70}{13} \int \frac{1}{x+3} \ dx + \frac{47}{13} \int \frac{x}{x^2 -x + 1} \ dx - \frac{32}{13} \int \frac{1}{x^2 -x + 1} \ dx \\ \end{align*}The integrals we are left with can be solved with both uu-substitution (for the first two) and trigonometric substitution (for the last one):

For the first one, we have

I1=70131x+3 dx    I1=70131u du=7013lnu+C    I1=7013lnx+3+CI_1 = \frac{70}{13} \int \frac{1}{x+3} \ dx \implies I_1 = \frac{70}{13} \int \frac{1}{u} \ du = \frac{70}{13} \ln|u| + C \implies I_1 = \frac{70}{13} \ln|x+3| +CFor remaining two, we will need to perform some algebraic manipulation first

472132x1+1x2x+1 dx32131x2x+1 dx=47262x1x2x+1 dx+(47263213)1x2x+1 dx=47262x1x2x+1 dx17261x2x+1 dx=I2+I3\begin{align*} \frac{47}{2 \cdot 13} \int \frac{2x - 1 + 1}{x^2 -x + 1} \ dx - \frac{32}{13} \int \frac{1}{x^2 -x + 1} \ dx &= \frac{47}{26} \int \frac{2x - 1}{x^2 -x + 1} \ dx + \left( \frac{47}{26} - \frac{32}{13} \right) \int \frac{1}{x^2 -x + 1} \ dx \\ &= \frac{47}{26} \int \frac{2x - 1}{x^2 -x + 1} \ dx - \frac{17}{26} \int \frac{1}{x^2 -x + 1} \ dx = I_2 + I_3 \end{align*}So, for I2I_2 we have

I2=47262x1x2x+1 dx    I2=47261u du=4726lnu+C    I2=4726lnx2x+1+CI_2 = \frac{47}{26} \int \frac{2x - 1}{x^2 -x + 1} \ dx \implies I_2 = \frac{47}{26} \int \frac{1}{u} \ du = \frac{47}{26} \ln|u| + C \implies I_2 = \frac{47}{26} \ln|x^2 -x +1| +CAnd for I3I_3, we have

I3=17261x2x+1 dx=17261(x12)2+34 dx=(1726)(43)1(23x13)2+1 dx    I3=34391tan2(θ)+1(32)sec2(θ) dθ=(3439)(32)1sec2(θ)sec2(θ) dθ=17339dθ=17339θ+C    I3=17339arctan(23x13)+C\begin{align*} I_3 = - \frac{17}{26} \int \frac{1}{x^2 -x + 1} \ dx = - \frac{17}{26} \int \frac{1}{\left(x-\frac{1}{2} \right)^2 + \frac{3}{4}} \ dx &= \left(- \frac{17}{26}\right)\left( \frac{4}{3} \right) \int \frac{1}{\left(\frac{2}{\sqrt{3}}x-\frac{1}{\sqrt{3}} \right)^2 + 1} \ dx \\ \implies I_3 &= - \frac{34}{39} \int \frac{1}{\tan^2(\theta) + 1} \cdot \left( \frac{\sqrt{3}}{2} \right) \cdot \sec^2(\theta) \ d\theta \\ &= \left( - \frac{34}{39} \right) \left( \frac{\sqrt{3}}{2} \right) \int \frac{1}{ \cancel{\sec^2(\theta)}} \cdot \cancel{\sec^2(\theta)} \ d\theta \\ &= - \frac{17\sqrt{3}}{39} \int d\theta = - \frac{17\sqrt{3}}{39} \theta + C \\ \implies I_3 &= - \frac{17\sqrt{3}}{39} \cdot \arctan\left(\frac{2}{\sqrt{3}}x-\frac{1}{\sqrt{3}} \right) + C \end{align*}Finally,

9x2+3x2x3+2x22x+3 dx=7013lnx+3+4726lnx2x+117339arctan(33(2x1))+C\int \frac{9x^2 + 3x - 2}{x^3 + 2x^2-2x+3} \ dx = \frac{70}{13} \ln|x+3| + \frac{47}{26} \ln|x^2 -x +1| - \frac{17\sqrt{3}}{39} \cdot \arctan\left(\frac{\sqrt{3}}{3} \cdot (2x- 1) \right) + C

Conclusion

In this topic, we have learned how to use the partial fraction expansion integration method. This method can be summarized as follows:

P(x)Q(x)dx=Aax+bdx+Bcx+ddx+\int \frac{P(x)}{Q(x)} dx = \int \frac{A}{ax+b} dx + \int \frac{B} {cx+d} dx + \cdotswhere each summand in the right-hand side can be calculated using the methods mentioned in our previous topics. It should not come as a surprise to you that in order to master all these methods, you need a lot of practice. So, let's not waste time and proceed with tasks.

How did you like the theory?
Report a typo