Skip to content

Mathematics · Ch 11 — Integral Calculus

Decomposition by Partial Fractions

11.7.2

Decomposition by Partial Fractions

When the integrand is a proper rational (algebraic) fraction p(x)q(x)\dfrac{p(x)}{q(x)} — a ratio of polynomials with q(x)≠0q(x)\ne0 and deg⁡p(x)<deg⁡q(x)\deg p(x)<\deg q(x) — and it does not simplify by the direct decomposition tricks of §11.7.1, the standard route is to rewrite it as a sum of simpler fractions with the same denominator factors: partial fractions. If deg⁡p(x)≥deg⁡q(x)\deg p(x)\ge\deg q(x), perform polynomial long division first, so the improper fraction becomes (polynomial) ++ (a genuinely proper fraction), and apply partial fractions only to the proper remainder.

How the decomposition is set up, factor by factor of q(x)q(x):

  • a non-repeated linear factor (x−a)(x-a) contributes a term Ax−a\dfrac{A}{x-a};
  • a repeated linear factor (x−a)k(x-a)^k contributes A1x−a+A2(x−a)2+⋯+Ak(x−a)k\dfrac{A_1}{x-a}+\dfrac{A_2}{(x-a)^2}+\cdots+\dfrac{A_k}{(x-a)^k};
  • an irreducible quadratic factor (x2+bx+c)(x^2+bx+c) (no real roots) contributes a term with a linear numerator, Bx+Cx2+bx+c\dfrac{Bx+C}{x^2+bx+c}.

The unknown constants are found by clearing denominators (multiplying both sides by q(x)q(x)) and then either substituting convenient values of xx (in particular, x=x= each root of a linear factor makes every other term vanish and pins that constant down immediately) or comparing coefficients of matching powers of xx on both sides — usually a mix of both is fastest.

Two worked-style illustrations of the technique (paralleling Example 11.29). For 3x+7x2−3x+2=3x+7(x−2)(x−1)\dfrac{3x+7}{x^2-3x+2}=\dfrac{3x+7}{(x-2)(x-1)}, write 3x+7(x−2)(x−1)=Ax−2+Bx−1\dfrac{3x+7}{(x-2)(x-1)}=\dfrac{A}{x-2}+\dfrac{B}{x-1}; clearing denominators gives 3x+7=A(x−1)+B(x−2)3x+7=A(x-1)+B(x-2), and substituting x=2x=2 gives A=13A=13, x=1x=1 gives B=−10B=-10, so the integral is 13log⁡∣x−2∣−10log⁡∣x−1∣+c13\log|x-2|-10\log|x-1|+c. For a denominator with a repeated factor, x+3(x+2)2(x+1)=Ax+2+B(x+2)2+Cx+1\dfrac{x+3}{(x+2)^2(x+1)}=\dfrac{A}{x+2}+\dfrac{B}{(x+2)^2}+\dfrac{C}{x+1}; clearing denominators and substituting x=−2x=-2 and x=−1x=-1 pins down BB and CC directly, and comparing the coefficient of x2x^2 (which must vanish, since the left side has none) pins down AA; each resulting piece is then either a log⁡\log (for a first-power denominator) or a negative power (for (x+2)−2(x+2)^{-2}, integrating to −1x+2-\dfrac{1}{x+2}). …