Skip to content

Mathematics · Ch 11 — Integral Calculus

Decomposition method

11.7.1

Decomposition method

Not every integrand matches one of the standard formulas directly. The decomposition method handles this by splitting a single "hard" integrand into a sum or difference of two or more "easy" pieces, each of which integrates by a known formula, and then integrating termwise using linearity: ∫[f(x)±g(x)]dx=∫f(x)dx±∫g(x)dx\int[f(x)\pm g(x)]dx = \int f(x)dx \pm \int g(x)dx.

The functions that call for this trick are typically algebraic, trigonometric, or exponential expressions that don't have their own listed formula — e.g. (1−x3)2(1-x^3)^2, x2−x+1x3\dfrac{x^2-x+1}{x^3}, cos⁡5xsin⁡3x\cos5x\sin3x, cos⁡3x\cos^3x, e2x−1ex\dfrac{e^{2x}-1}{e^x} — but which decompose into pieces that do.

Algebraic decomposition. A power of a binomial like (1−x3)2(1-x^3)^2 is expanded first (=1−2x3+x6=1-2x^3+x^6) and then integrated term by term. A fraction whose numerator has several terms over a monomial denominator, such as x2−x+1x3\dfrac{x^2-x+1}{x^3}, is split term-by-term into 1x−1x2+1x3\dfrac1x-\dfrac1{x^2}+\dfrac1{x^3} and each piece integrated using the power rule (with ∫x−1dx=log⁡∣x∣+c\int x^{-1}dx=\log|x|+c for the middle term). A quotient like (x−1)2x3+x\dfrac{(x-1)^2}{x^3+x} is handled by expanding the numerator and splitting the resulting fraction over the factored denominator x(x2+1)x(x^2+1) into 1x−2xx2+1\dfrac1x-\dfrac{2x}{x^2+1}, each piece a known log-type integral — this is really a first taste of the partial-fractions idea developed fully in §11.7.2.

Trigonometric decomposition. A product of two different trig functions of different arguments, like cos⁡5xsin⁡3x\cos5x\sin3x, is converted to a sum using a product-to-sum identity (e.g. 2cos⁡Asin⁡B=sin⁡(A+B)−sin⁡(A−B)2\cos A\sin B=\sin(A+B)-\sin(A-B)) before integrating. A power like cos⁡3x\cos^3x is reduced using the triple-angle identity cos⁡3x=14(3cos⁡x+cos⁡3x)\cos^3x=\dfrac14(3\cos x+\cos3x). An expression like 1sin⁡2xcos⁡2x\dfrac{1}{\sin^2x\cos^2x} is attacked by writing 1=sin⁡2x+cos⁡2x1=\sin^2x+\cos^2x in the numerator and splitting into sec⁡2x+csc⁡2x\sec^2x+\csc^2x. A quotient such as sin⁡x1+sin⁡x\dfrac{\sin x}{1+\sin x} is rationalised by multiplying top and bottom by 1−sin⁡x1-\sin x, turning the denominator into 1−sin⁡2x=cos⁡2x1-\sin^2x=\cos^2x and the whole thing into tan⁡xsec⁡x−sec⁡2x+1\tan x\sec x-\sec^2x+1, three standard pieces. Expressions built from 1±cos⁡2x1\pm\cos2x or 1±sin⁡2x1\pm\sin2x are simplified first with the half-angle/double-angle identities 1+cos⁡2x=2cos⁡2x1+\cos2x=2\cos^2x, 1−cos⁡2x=2sin⁡2x1-\cos2x=2\sin^2x, and 1±sin⁡2x=(cos⁡x±sin⁡x)21\pm\sin2x=(\cos x\pm\sin x)^2, which often collapses a square-root integrand into something linear in sin⁡x,cos⁡x\sin x,\cos x. A square like (tan⁡x+cot⁡x)2(\tan x+\cot x)^2 expands to tan⁡2x+2+cot⁡2x\tan^2x+2+\cot^2x and then converts to sec⁡2x+csc⁡2x\sec^2x+\csc^2x via the Pythagorean identities tan⁡2x=sec⁡2x−1\tan^2x=\sec^2x-1, cot⁡2x=csc⁡2x−1\cot^2x=\csc^2x-1.

Exponential decomposition. A quotient like e2x−1ex\dfrac{e^{2x}-1}{e^x} is split by dividing each term of the numerator by exe^x, giving ex−e−xe^x-e^{-x}. A product of two different exponential bases, such as axexa^xe^x, is recognised as a single exponential with base aeae: axex=(ae)xa^xe^x=(ae)^x, integrating to (ae)xlog⁡(ae)+c\dfrac{(ae)^x}{\log(ae)}+c (from ∫bxdx=bxlog⁡b+c\int b^xdx=\dfrac{b^x}{\log b}+c). …