Printed Page 499
499
When you finish this section, you should be able to:
Rational functions are discussed in Section P.2, pp. 19-20.
In this section, we integrate rational functions using a technique called partial fractions. Although the integral of any rational function can be found using partial fractions, the process requires being able to factor the denominator of the rational function.
Recall that a rational function R in lowest terms is the ratio of two polynomial functions p and q≠0, where p and q have no common factors. The domain of R is the set of all real numbers for which q≠0. The rational function R is called proper when the degree of the polynomial p is less than the degree of the polynomial q; otherwise, R is an improper rational function.
Every improper rational function R can be reduced by long division to the sum of a polynomial function and a proper rational function. For example, the rational function R(x)=x2x−1 is improper, but by long division, R(x)=x2x−1=x+1+1x−1
Then we can find ∫R(x)dx by using properties of an indefinite integral to obtain ∫x2x−1dx=∫(x+1+1x−1)dx=∫x dx+∫dx+∫1x−1dx=x22+x+ln|x−1|+C
For these reasons, the discussion that follows deals only with proper rational functions in lowest terms. Such rational functions can be written as the sum of simpler functions, called partial fractions.
For example, since 2x−1+3x+4=5x+5(x−1)(x+4), the rational function R(x)=5x+5(x−1)(x+4)
can be expressed as the sum R(x)=2x−1+3x+4 . Then R(x)=∫5x+5(x−1)(x+4)dx=∫(2x−1+3x+4)dx=2∫1x−1dx+3∫1x+4dx=2ln|x−1|+3ln|x+4|+C
500
We use a technique called partial fraction decomposition to write a proper rational function R as the sum of simpler rational terms. The resulting partial fractions depend on the nature of the factors of the denominator q. It can be shown that any polynomial q whose coefficients are real numbers can be factored (over the real numbers) into products of linear and/or irreducible quadratic factors. This means the integral of every rational function can be expressed in terms of algebraic, logarithmic, and/or inverse trigonometric functions.
The rest of this section presents systematic methods of decomposing rational functions into sums of partial fractions, that is, into forms that we can integrate.
Printed Page 500
Case 1: If the denominator q contains only distinct linear factors, say, x−a1, x−a2,…,x−an, then pq can be written as \bbox[5px, border:1px solid black, #F9F7ED]{\dfrac{p(x)}{q(x)}= \dfrac{A_{1}}{x-a_{1}}+\dfrac{A_{2}}{x-a_{2}}+\cdots +\dfrac{A_{n}}{x-a_{n}}}\tag{1}
where A_{1}, A_{2}, \ldots , A_{n} are real numbers.
To find \int \dfrac{p( x) }{q( x) }dx, we integrate both sides of (1). Then \begin{eqnarray*} \int \dfrac{p( x) }{q( x) }dx &=&\int \dfrac{A_{1}}{ x-a_{1}}dx+\int \dfrac{A_{2}}{x-a_{2}}dx+\cdots +\int \dfrac{A_{n}}{x-a_{n}}dx \\[6pt] &=&A_{1}\ln \left\vert x-a_{1}\right\vert +A_{2}\ln \left\vert x-a_{2}\right\vert +\cdots +A_{n}\ln \left\vert x-a_{n}\right\vert +C \end{eqnarray*}
All that remains is to find the numbers A_{1}, \ldots, A_{n}.
A procedure for finding the numbers A_{1}, \ldots , A_{n} is illustrated in Example 1.
Find \int \dfrac{x\,dx}{x^{2}-5x+6}.
Solution The integrand is a proper rational function in lowest terms. We begin by factoring the denominator: x^{2}-5x+6=(x-2)(x-3). Since the factors are linear and distinct, we apply Case 1 and allow for the terms \dfrac{A}{x-2} and \dfrac{B}{x-3}. \dfrac{x}{(x-2)(x-3)}=\dfrac{A}{x-2}+\dfrac{B}{x-3}
Now we clear fractions by multiplying both sides of the equation by (x-2)(x-3). \begin{eqnarray*} \begin{array}{ll@{\qquad}l} x &= A(x-3)+B(x-2) \\[3pt] x &= (A+B)x-( 3A+2B) & {\color{#0066A7}{\hbox{Group like terms.}}} \end{array} \end{eqnarray*}
This is an identity in x, so the coefficients of like powers of x must be equal. \begin{eqnarray*} \begin{array}{ll@{\qquad}l} 1 &= A+B & {\color{#0066A7}{\hbox{The coefficient of }{x}\hbox{ equals 1.}}} \\[3pt] 0 &= -3A-2B & {\color{#0066A7}{\hbox{The constant term on the left is 0.}}} \end{array} \end{eqnarray*}
This is a system of two equations containing two variables. Solving the second equation for B, we get B = -\dfrac{3}{2}A. Substituting for B in the first equation produces the solution A = -2 from which B = 3. So, \dfrac{x}{(x-2)(x-3)}=\dfrac{-2}{x-2}+\dfrac{3}{x-3}
501
Then \begin{eqnarray*} \int \dfrac{x}{(x-2)(x-3)}dx &=& \int \dfrac{-2}{x-2}\,dx+\int \dfrac{3}{x-3} \,dx \\ &=& -2\,\ln \,\vert x-2 \vert +3\ln \,\vert x-3 \vert +C=\ln \left\vert \dfrac{(x-3)^{3}}{(x-2)^{2}}\right\vert +C \end{eqnarray*}
Alternatively, we can find the unknown numbers in the decomposition of \dfrac{p}{q} by substituting convenient values of x into the identity obtained after clearing fractions.* In Example 1, after clearing fractions, the identity is \begin{equation*} x=A(x-3)+B(x-2) \end{equation*}
*This method is discussed in detail in H. J. Straight & R. Dowds (1984, June-July), American Mathematical Monthly, 91(6), 365.
When x=3, the term involving A drops out, leaving 3=B\cdot 1, so that B=3. When x=2, the term involving B drops out, leaving 2=A\cdot (-1), so that A=-2.
Derive these formulas: \bbox[5px, border:1px solid black, #F9F7ED]{{(a)} \quad \int \dfrac{dx}{x^{2}-a^{2}}=\dfrac{1}{2a}\ln \left \vert \dfrac{x-a}{x+a}\right \vert +C \quad a ≠ 0 } \bbox[5px, border:1px solid black, #F9F7ED]{{(b)} \quad \int \dfrac{dx}{a^{2}-x^{2}}=\dfrac{1}{2a}\ln \left\vert \dfrac{x+a}{x-a}\right\vert +C \quad a ≠ 0 }
Solution (a) The factored denominator x^{2}-a^{2}=( x-a) (x+a) contains only distinct linear factors. So, \dfrac{1}{x^{2}-a^{2}} can be decomposed into partial fractions of the form \begin{eqnarray*} \begin{array}{rl@{\qquad}l} \dfrac{1}{x^{2}-a^{2}} &= \dfrac{1}{(x-a)(x+a)}=\dfrac{A}{x-a}+\dfrac{B}{x+a}\\ 1 &= A(x+a) +B (x-a)&{\color{#0066A7}{\hbox{Multiply both sides by }(x-a) (x+a).}} \end{array} \end{eqnarray*}
This is an identity in x. When x=a, the term involving B drops out. Then 1=A(2a), so A=\dfrac{1}{2a}. When x=-a, the term involving A drops out. Then 1=B (-2a), so B=-\dfrac{1}{2a}. Then, \begin{equation*} \dfrac{1}{x^{2}-a^{2}}=\dfrac{A}{x-a}+\dfrac{B}{x+a}=\dfrac{1}{2a( x-a) }-\dfrac{1}{2a(x+a) } \end{equation*} \begin{eqnarray*} \int \dfrac{dx}{x^{2}-a^{2}} &=& \dfrac{1}{2a}\int \dfrac{dx}{x-a}-\dfrac{1}{2a} \int \dfrac{dx}{x+a}=\dfrac{1}{2a}\left( \int \dfrac{dx}{x-a}-\int \dfrac{dx}{x+a}\right) \\[7pt] &=& \dfrac{1}{2a}\big(\ln \,\vert x-a\vert -\ln \,\vert x+a \vert\big) +C\\[7pt] &=&\dfrac{1}{2a}\ln \left\vert \dfrac{x-a}{x+a}\right\vert +C \end{eqnarray*}
502
(b) Using the result from (a), we get \begin{eqnarray*} \int \dfrac{dx}{a^{2}-x^{2}} &=& -\int \dfrac{dx}{x^{2}-a^{2}}=-\dfrac{1}{2a}\ln \left\vert \dfrac{x-a}{x+a}\right\vert +C=\dfrac{1}{2a}\ln \left\vert \dfrac{x-a }{x+a}\right\vert ^{-1}+C \\[6pt] &=& \dfrac{1}{2a}\ln \left\vert \dfrac{x+a}{x-a}\right\vert +C \end{eqnarray*}
Printed Page 502
Case 2: If the denominator q has a repeated linear factor (x-a)^{n}, n \geq 2 an integer, then the decomposition of \dfrac{p}{q} includes the terms \bbox[5px, border:1px solid black, #F9F7ED]{ \dfrac{A_{1}}{x-a}, \dfrac{A_{2}}{(x-a)^{2}},\ \ldots ,\ \dfrac{A_{n}}{(x-a)^{n}}}
where A_{1}, A_{2,} \ldots , A_{n} are real numbers.
Find \int \dfrac{dx}{x(x-1)^{2}}.
Solution Since x is a distinct linear factor of the denominator q, and (x-1)^{2} is a repeated linear factor of the denominator, the decomposition of \dfrac{1}{x(x-1) ^{2}} into partial fractions has the three terms \dfrac{A}{x}, \dfrac{B}{x-1}, and \dfrac{C}{(x-1) ^{2}}. \begin{eqnarray*} \begin{array}{rl@{\!\!\!\qquad}l} \dfrac{1}{x(x-1)^{2}} &= \dfrac{A}{x}+\dfrac{B}{x-1}+\dfrac{C}{(x-1)^{2}} \quad &{\color{#0066A7}{\hbox{Write the identity.}}} \\ 1 &= A(x-1)^{2}+B\cdot x(x-1)+C\cdot x\quad &{\color{#0066A7}{\hbox{Multiply both sides by }{x} (x-1)^{2}.}} \end{array} \end{eqnarray*}
We find A, B, and C by choosing values of x that cause one or more terms to drop out. When x=1, we have 1=C\cdot 1, so C=1. When x=0, we have 1=A(0-1)^{2}, so A=1. Now using A=1 and C=1, we have \begin{equation*} 1=(x-1)^{2}+B\cdot x(x-1)+ 1 \cdot x \end{equation*}
Suppose we let x=2. (Any choice other than 0 and 1 will also work.) Then \begin{eqnarray*} 1 &=&1+2B+2 \\ B &=&-1 \end{eqnarray*}
Then \begin{array}{l@{\qquad}l} \dfrac{1}{x(x-1)^{2}}=\dfrac{1}{x}+\dfrac{-1}{(x-1) }+\dfrac{1}{(x-1)^{2}} \quad{\color{#0066A7}{\hbox{\(A=1\)}}\quad \quad{\color{#0066A7}{\hbox{\(B=-1\)}}}\quad \quad{\color{#0066A7}{\hbox{\(C=1\)}}}} \end{array}
503
So, \begin{eqnarray*} \int \dfrac{dx}{x(x-1)^{2}} &=& \int \dfrac{dx}{x}- \int \dfrac{dx}{x-1}+ \int \dfrac{dx}{(x-1)^{2}}\\[7pt] &=& \ln \vert x \vert -\ln \vert x-1\vert -\dfrac{1}{x-1} +C_{1} \end{eqnarray*}
To avoid confusion, we use C_{1} for the constant of integration whenever C appears in the partial fraction decomposition.
Printed Page 503
A quadratic polynomial ax^{2}+bx+c is called irreducible if it cannot be factored into real linear factors. This happens if the discriminant b^{2}-4ac<0. For example, x^{2}+x+1 and x^{2}+4 are irreducible.
The discriminant of a quadratic equation is discussed in Appendix A.1, pp. A-3 to A-4.
Case 3: If the denominator q contains a nonrepeated irreducible quadratic factor ax^{2}+bx+c, then the decomposition of \dfrac{p}{q} includes the term \bbox[5px, border:1px solid black, #F9F7ED]{ \dfrac{Ax+B}{ax^{2}+bx+c} }
where A and B are real numbers.
Find \int \dfrac{3x}{x^{3}-1}\,dx.
Solution The denominator is x^{3}-1=(x-1)(x^{2}+x+1). Since x-1 is a nonrepeated linear factor, by Case 1 the decomposition of \dfrac{ 3x}{x^{3}-1}=\dfrac{3x}{(x-1) (x^{2}+x+1) } has the term \dfrac{A}{x-1}.
The discriminant of the quadratic equation x^{2}+x+1=0 is negative, so x^{2}+x+1 is an irreducible quadratic factor of q, and the decomposition of \dfrac{p}{q} also contains the term \dfrac{Bx+C}{x^{2}+x+1}. Then \dfrac{3x}{x^{3}-1}=\dfrac{A}{x-1}+\dfrac{Bx+C}{x^{2}+x+1}
Clearing the denominators, we have \begin{equation*} 3x=A(x^{2}+x+1)+(Bx+C)(x-1)\ \end{equation*}
This is an identity in x. When x=1, we have 3=3A, so A=1. With A=1, the identity becomes \begin{eqnarray*} 3x &=& (x^{2}+x+1) + (Bx+C) (x-1) \\[3pt] -x^{2}+2x-1 &=& (Bx+C) (x-1) \\[3pt] -(x-1) ^{2} &=& (Bx+C) (x-1) \\[3pt] -(x-1) &=& Bx+C \\[3pt] B &=& -1,\quad C=1 \end{eqnarray*}
504
So, \begin{eqnarray*} \dfrac{3x}{x^{3}-1} &=&\dfrac{1}{x-1}+\dfrac{-x+1}{x^{2}+x+1} \nonumber\\[5pt] \int \dfrac{3x}{x^{3}-1}\,dx &=&\int \left( \dfrac{1}{x-1}+\dfrac{-x+1}{ x^{2}+x+1}\right) dx=\int \dfrac{1}{x-1}\,dx+\int \dfrac{-x+1}{x^{2}+x+1}\,dx \nonumber\\[5pt] &=& \ln \vert x-1\vert -\int \dfrac{x-1}{x^{2}+x+1}\, dx\tag{2} \end{eqnarray*}
To find the integral on the right, we complete the square in the denominator and use substitution. \begin{eqnarray*} \int \dfrac{x-1}{x^{2}+x+1}dx &=& \int \dfrac{x-1}{\left( x+\dfrac{1}{2}\right)^{2}+\dfrac{3}{4}}dx \underset{\underset{\color{#0066A7}{\hbox{\(u=x+\dfrac{1}{2}\)}}}{\color{#0066A7}{\left\uparrow{\vphantom{\vrule width0pc height11pt\\ depth0pt}}\right.}}} {=} \int \dfrac{u-\dfrac{3}{2}}{u^{2}+\dfrac{3}{4}}du=\int \dfrac{u}{u^{2}+\dfrac{3}{4\\ }}du-\dfrac{3}{2}\int \dfrac{du}{u^2+\dfrac{3}{4}} \\ &=& \dfrac{1}{2}\ln \left( u^{2}+\dfrac{3}{4}\right) -\dfrac{3}{2}\left[ \dfrac{2}{\sqrt{3}}\tan ^{-1}\left( \dfrac{2}{\sqrt{3}}u\right) \right] \quad {\color{#0066A7}{\hbox{\(\int \dfrac{du}{u^{2}+a^{2}}\, =\dfrac{1}{a}\tan^{-1}\left( \dfrac{u}{a}\right)\)}}} \\ &=& \dfrac{1}{2}\ln \left( u^{2}+\dfrac{3}{4}\right) -\sqrt{3}\tan ^{-1}\left( \dfrac{2}{\sqrt{3}}u\right) \\ &=& \dfrac{1}{2}\ln (x^{2}+x+1) -\sqrt{3}\tan ^{-1}\left( \dfrac{2x+1}{\sqrt{3}}\right) \end{eqnarray*}
Then from (2), \int \dfrac{3x}{x^{3}-1}dx=\ln \vert x-1\vert -\dfrac{1}{2}\ln \,(x^{2}+x+1)+\sqrt{3}\tan ^{-1}\left( \dfrac{2x+1}{\sqrt{3}}\right) +C_{1}
Printed Page 504
Case 4: If the denominator q contains a repeated irreducible quadratic polynomial (x^{2}+bx+c)^{n}, n \geq 2 an integer, then the decomposition of \dfrac{p}{q} includes the terms \bbox[5px, border:1px solid black, #F9F7ED]{ \dfrac{A_{1}x+B_{1}}{x^{2}+bx+c}\hbox{, }\ \dfrac{A_{2}x+B_{2}}{ (x^{2}+bx+c)^{2}}\hbox{,}\ \ldots\hbox{, }\ \dfrac{A_{n}x+B_{n}}{(x^{2}+bx+c)^{n}} }
where A_{1}, B_{1}, A_{2}, B_{2}, …, A_{n}, B_{n} are real numbers.
Find \int \dfrac{x^{3}+1}{(x^{2}+4)^{2}}\,dx.
Solution The denominator is a repeated, irreducible quadratic, so the decomposition of \dfrac{x^{3}+1}{(x^{2}+4)^{2}} is \dfrac{x^{3}+1}{(x^{2}+4)^{2}}=\dfrac{Ax+B}{x^{2}+4}+\dfrac{Cx+D}{(x^{2}+4)^{2}}
505
Clearing fractions and combining terms give \begin{eqnarray*} x^{3}+1 &=& (Ax+B) (x^{2}+4) +Cx+D \\[3pt] x^{3}+1 &=& Ax^{3}+Bx^{2}+(4A+C)x+4B+D \end{eqnarray*}
Equating coefficients, we get \begin{equation*} \begin{array}{rl@{\qquad}rrr} A=1 \quad B=0\quad 4A+C &= 0 & 4B+D &= 1 \\[3pt] C &=-4 & D &=1 \end{array} \end{equation*}
Then \begin{equation*} \dfrac{x^{3}+1}{(x^{2}+4)^{2}}=\dfrac{x}{x^{2}+4}+\dfrac{-4x+1}{(x^{2}+4)^{2}} \end{equation*}
and \begin{eqnarray*} \int \dfrac{x^{3}+1}{(x^{2}+4)^{2}}dx &=& \int \dfrac{x}{x^{2}+4}dx+\int \dfrac{ -4x+1}{(x^{2}+4)^{2}}dx \nonumber\\[5pt] & =& \int \dfrac{x}{x^{2}+4}dx-4\int \dfrac{x}{(x^{2}+4)^{2} }dx+\int \dfrac{dx}{(x^{2}+4)^{2}}\tag{3} \end{eqnarray*}
In the first two integrals on the right in (3), we use the substitution u=x^{2}+4 , x \geq 0. Then du=2x\,dx, and
In the third integral on the right in (3), we use the trigonometric substitution x=2\tan \theta , -\dfrac{\pi }{2} \lt \theta \lt \dfrac{\pi }{2}. Then dx=2\sec ^{2} \theta \,d\theta , and \begin{eqnarray*} \int \dfrac{dx}{(x^{2}+4)^{2}} &=& \int \dfrac{2\sec ^{2} \theta \,d\theta }{ \left( 4\tan ^{2}\theta +4\right) ^{2}}=\dfrac{2}{16}\int \dfrac{\sec ^{2} \theta \,d\theta }{( \sec ^{2} \theta) ^{2}}=\dfrac{1}{8}\int \cos ^{2}\theta \,d\theta \\[7pt] &=& \dfrac{1}{8}\int \dfrac{1+\cos (2\theta) }{2}d\theta \\[7pt] &=& \dfrac{1}{16}\left[ \theta +\dfrac{1}{2}\sin (2\theta ) \right] =\dfrac{1}{16}(\theta +\sin \theta \cos \theta ) \end{eqnarray*}
To express the solution in terms of x, either use the triangles in Figure 10 or use trigonometric identities as follows: \begin{equation*} \sin \theta \cos \theta =\dfrac{\sin \theta }{\cos \theta }\cdot \cos ^{2}\theta =\dfrac{\tan \theta }{\sec ^{2} \theta }=\dfrac{\tan \theta }{\tan ^{2}\theta +1}=\dfrac{\dfrac{x}{2}}{\dfrac{x^{2}}{4}+1}=\dfrac{2x}{x^{2}+4} \end{equation*}
Then \begin{eqnarray*} \int \dfrac{dx}{(x^{2}+4)^{2}} &=& \dfrac{1}{16}(\theta +\sin \theta \cos \theta )\underset{\underset{\color{#0066A7}{\hbox{\(x=2\tan \theta{;} \theta =\tan^{-1}\dfrac{x}{2}\)}}}{\color{#0066A7}{\uparrow}}}{=} \dfrac{1}{16}\left( \tan ^{-1}\dfrac{x}{2}+\dfrac{2x}{x^{2}+4}\right)\tag{6} \end{eqnarray*}
Now use the results of (3), (4), (5), and (6): \int \dfrac{x^{3}+1}{(x^{2}+4)^{2}}dx=\dfrac{1}{2}\ln (x^{2}+4)+\dfrac{2}{ x^{2}+4}+\dfrac{1}{16}\tan ^{-1}\dfrac{x}{2}+\dfrac{x}{8(x^{2}+4)}+C_{1}