Integration Techniques

Integration techniques are systematic methods for finding antiderivatives (indefinite integrals) and evaluating definite integrals. Key result: every elementary integral can be attacked by a small toolkit — substitution, integration by parts, partial fractions, and trigonometric methods — that converts a hard integral into a recognisable standard form. Applies to: computing areas, volumes, work, probabilities, and solving differential equations. Who/when: developed by Newton and Leibniz (1660s–1680s) and refined by Euler, Bernoulli, and later analysts.

Integration is the inverse operation to differentiation: given a function $f$ , we seek a function $F$ whose derivative is $f$ . While differentiation is mechanical, integration is an art — there is no single algorithm that integrates every elementary function, and many simple-looking integrals (like $\int e^{- x^{2}} d x$ ) have no elementary antiderivative at all. The techniques on this page are the standard transformations that reduce difficult integrals to a handful of memorised results. They matter because integration underlies nearly every quantitative science: areas under curves, accumulated change, and the solutions of differential equations. The surprising fact: integration is genuinely harder than differentiation — there exist functions you can write down in one line whose integral provably cannot be expressed with elementary functions.

The Core Insight

Every integration technique is a disguised version of a differentiation rule run backwards. The chain rule becomes substitution, the product rule becomes integration by parts, and algebraic identities (partial fractions, trig identities) reshape an integrand until it matches a known derivative. The skill is pattern recognition: spotting which rule's "fingerprint" appears in the integrand.

Formal Statement

Antiderivative and Indefinite Integral

A function $F$ is an antiderivative of $f$ on an interval $I$ if $F^{'} (x) = f (x)$ for all $x \in I$ . The indefinite integral is the family of all antiderivatives:

\int f (x) d x = F (x) + C, F^{'} (x) = f (x)

⟳directly from the definition of antiderivative; the constant

C

appears because

\frac{d}{d x} C = 0

, so antiderivatives are unique only up to an additive constant.

Here $f$ is the integrand, $x$ the variable of integration, $d x$ the differential indicating the variable, and $C$ the constant of integration.

Definite Integral via the Fundamental Theorem

For $f$ continuous on $[a, b]$ with antiderivative $F$ :

\int_{a}^{b} f (x) d x = F (b) - F (a)

⟳this is the Fundamental Theorem of Calculus, Part 2; it follows from defining the area function

A (x) = \int_{a}^{x} f

and showing

A^{'} (x) = f (x)

Symbols: $a$ = lower limit, $b$ = upper limit, and the result is a number (often interpreted as signed area), not a family of functions — so no $+ C$ is needed.

Integration by Substitution

If $u = g (x)$ is differentiable and $f$ is continuous on the range of $g$ , then

\int f (g (x)) g^{'} (x) d x = \int f (u) d u = F (u) + C

⟳reverse the chain rule. If

F^{'} = f

, then

\frac{d}{d x} F (g (x)) = F^{'} (g (x)) g^{'} (x) = f (g (x)) g^{'} (x)

Integration by Parts

For differentiable functions $u (x)$ and $v (x)$ :

\int u d v = uv - \int v d u

⟳integrate the product rule

\frac{d}{d x} (uv) = u v^{'} + v u^{'}

over the interval and rearrange.

Partial Fraction Decomposition

Any proper rational function (degree of numerator $<$ degree of denominator) can be written as a sum of simpler fractions. For distinct linear factors:

\frac{P ( x )}{( x - r _{1} ) ( x - r _{2} ) \dots ( x - r _{n} )} = \frac{A _{1}}{x - r _{1}} + \frac{A _{2}}{x - r _{2}} + \dots + \frac{A _{n}}{x - r _{n}}

⟳clear denominators to get a polynomial identity, then solve for the constants

A_{i}

by substituting the roots

x = r_{i}

(the cover-up method).

Why It Holds / Derivation

Integration by Parts from the Product Rule

We derive equation [4] step by step.

Start with the product rule for derivatives:

\frac{d}{d x} (u (x) v (x)) = u (x) v^{'} (x) + v (x) u^{'} (x)

⟳standard product rule.

Integrate both sides with respect to $x$ . The left side antiderivative is simply $uv$ :

u (x) v (x) = \int u (x) v^{'} (x) d x + \int v (x) u^{'} (x) d x

⟳the integral of a derivative returns the original function (FTC), and integration is linear so it distributes over the sum.

Rearrange to isolate the first integral, and write $d v = v^{'} d x$ , $d u = u^{'} d x$ :

\int u d v = uv - \int v d u

⟳subtract

\int v u^{'} d x

from both sides of step 2; substitute differential notation.

This completes the derivation. The technique trades the integral $\int u d v$ for a (hopefully easier) integral $\int v d u$ . $■$

Substitution from the Chain Rule

Suppose $F$ is an antiderivative of $f$ , so $F^{'} = f$ . Consider the composite $F (g (x))$ .

By the chain rule:

\frac{d}{d x} F (g (x)) = F^{'} (g (x)) \cdot g^{'} (x) = f (g (x)) g^{'} (x)

⟳chain rule applied to

F \circ g

Since $F (g (x))$ is an antiderivative of $f (g (x)) g^{'} (x)$ , integrating both sides gives:

\int f (g (x)) g^{'} (x) d x = F (g (x)) + C = \int f (u) d u_{u = g (x)}

⟳equation [1] applied to the result of step 2, with the substitution

u = g (x)

■

Gottfried Wilhelm Leibniz | 1646–1716

Leibniz introduced the integral sign $\int$ (an elongated "S" for summa, sum) in 1675 and the differential notation $d x$ that makes substitution and integration by parts so natural to write. His notation was so well-designed that we still use it unchanged today — the form $\int u d v = uv - \int v d u$ is essentially his. Working independently of Newton, Leibniz emphasised the symbolic calculus — manipulating $d x$ and $\int$ as algebraic objects — which proved more fertile for developing techniques than Newton's geometric "fluxions."

Building Intuition

Think of differentiation as "scrambling an egg" — easy, mechanical, always possible. Integration is "unscrambling the egg" — you must guess what original function would have produced this derivative, then verify.

Each technique is a strategy for guessing:

Substitution asks: "Is there an inner function whose derivative is also lurking in the integrand?" If you see $g (x)$ and its derivative $g^{'} (x)$ multiplied together, collapse them into a single variable $u$ .
Integration by parts asks: "Is this a product of two unrelated functions?" Pick the part that gets simpler when differentiated (that's $u$ ) and the part you can easily integrate (that's $d v$ ). The mnemonic LIATE (Logarithmic, Inverse trig, Algebraic, Trigonometric, Exponential) ranks which factor to call $u$ — earlier in the list = better choice for $u$ .
Partial fractions asks: "Is this a single ugly fraction?" Break it into a sum of fractions each simple enough to integrate as a logarithm or power.

The whole game is transformation until recognition: keep reshaping the integrand until it looks like something on your list of known integrals.

Substitution as "collapsing" a composite function: the shaded area under $f(g(x))g'(x)$ equals the area under $f(u)$ in the new variable

Worked Examples

Substitution: $\int 2x\,\cos(x^2)\,dx$

Given: the integral

\int 2 x cos (x^{2}) d x

. Step 1 — Spot the inner function. The argument of cosine is

x^{2}

, and its derivative

2 x

appears as a factor. This is the substitution fingerprint. Step 2 — Substitute. Let

u = x^{2}

, so

\frac{d u}{d x} = 2 x

, giving

d u = 2 x d x

. Step 3 — Rewrite the integral entirely in $u$ :

\int 2 x cos (x^{2}) d x = \int cos (u) d u

⟳replace

x^{2} \to u

and

2 x d x \to d u

using the substitution from step 2.

Step 4 — Integrate the standard form:

\int cos (u) d u = sin (u) + C

⟳

\frac{d}{d u} sin u = cos u

Step 5 — Substitute back

u = x^{2}

\int 2 x cos (x^{2}) d x = sin (x^{2}) + C

(Indefinite integral — dimensionless; answer is a function plus constant.)

Integration by Parts: $\int_0^1 x\,e^{x}\,dx$

Given: the definite integral

\int_{0}^{1} x e^{x} d x

. Step 1 — Choose $u$ and $d v$ via LIATE. Algebraic (

x

) comes before Exponential (

e^{x}

), so let

u = x, d v = e^{x} d x .

Step 2 — Differentiate and integrate:

d u = d x, v = \int e^{x} d x = e^{x} .

Step 3 — Apply equation [4]:

\int x e^{x} d x = x e^{x} - \int e^{x} d x = x e^{x} - e^{x} + C

⟳substitute

u = x

v = e^{x}

into

\int u d v = uv - \int v d u

; the remaining

\int e^{x} d x = e^{x}

Step 4 — Evaluate from 0 to 1 using equation [2]:

[x e^{x} - e^{x}]_{0}^{1} = (1 \cdot e^{1} - e^{1}) - (0 \cdot e^{0} - e^{0}) = 0 - (- 1) = 1

⟳at

x = 1

e - e = 0

; at

x = 0

0 - 1 = - 1

; subtract.

\int_{0}^{1} x e^{x} d x = 1

(Definite integral — a pure number, the signed area under $x e^{x}$ on $[0, 1]$ .)

Common Misconceptions

The integral of a product is NOT the product of integrals. It is a frequent error to write

\int f (x) g (x) d x = ? (\int f d x) (\int g d x) .

This is false. There is no product rule for integration — that is precisely why integration by parts exists. Likewise,

\int \frac{f}{g} d x \neq = \frac{\int f}{\int g}

. Always reshape products and quotients with parts, substitution, or partial fractions. Other frequent slips:

Forgetting the $+ C$ on indefinite integrals (it represents infinitely many functions).
After substitution, forgetting to convert $d x$ into $d u$ — you cannot mix variables.
For definite integrals via substitution, either change the limits to $u$ -values or convert back to $x$ before plugging in — never apply $x$ -limits to a $u$ -expression.

Test the false "product rule." Take

f (x) = x

and

g (x) = x

, so

f g = x^{2}

The true integral: $\int x^{2} d x = \frac{x ^{3}}{3} + C$ .

The fake formula: $(\int x d x) (\int x d x) = (\frac{x ^{2}}{2}) (\frac{x ^{2}}{2}) = \frac{x ^{4}}{4}$ .

These disagree ( $\frac{x ^{3}}{3} \neq = \frac{x ^{4}}{4}$ ), proving the product rule for integrals is invalid.

Connections

Integration techniques sit at the heart of calculus and connect outward to many areas:

They invert Differentiation — every technique is a differentiation rule reversed.
They depend on the Fundamental Theorem of Calculus, which links antiderivatives to areas and makes equation [2] possible.
They are the engine for Differential Equations, where separating variables and integrating both sides yields solutions.
Substitution generalises to Trigonometric Substitution for integrands containing $a^{2} - x^{2}$ and similar radicals.
When elementary techniques fail, one turns to Numerical Integration (Simpson's rule, trapezoidal rule) for approximate values.

Real-World Applications

Physics — work and energy: $W = \int F d x$ computes work done by a variable force; integration by parts handles forces like $F = x e^{- x}$ .
Engineering — centre of mass and moments: finding the centroid of a beam or plate requires $\overset{x}{ˉ} = \frac{\int x d A}{\int d A}$ , evaluated by substitution.
Probability and statistics: the normal distribution's cumulative probability is $\int e^{- x^{2} /2} d x$ — famously not elementary, forcing numerical integration in every statistics package.
Economics — consumer surplus: the area between a demand curve and price, $\int_{0}^{Q} (P (q) - p) d q$ , measures total consumer benefit.
Electrical engineering — RC circuits: charge accumulation $Q = \int I d t$ with exponential currents uses substitution and parts.
Pharmacokinetics: the "area under the curve" (AUC) of drug concentration $\int_{0}^{\infty} C (t) d t$ measures total drug exposure, computed by parts for $C (t) = C_{0} e^{- k t}$ .
Computer graphics — light transport: rendering integrals $\int L (ω) d ω$ over solid angles are evaluated numerically (Monte Carlo) when analytic techniques fail.

Historical Development

1635 — Bonaventura Cavalieri develops the "method of indivisibles," computing areas by summing infinitely thin slices — a precursor to integration.
1660s — Isaac Newton develops his calculus of "fluxions," treating integration as the inverse of finding rates of change.
1675 — Leibniz introduces the integral sign $\int$ and the notation $d x$ , on 29 October 1675, giving integration its modern symbolic form.
1690s — Johann and Jakob Bernoulli systematise integration techniques; Johann effectively invents the term "integral" and develops many substitution methods.
1740s–1760s — Leonhard Euler compiles and extends integration techniques massively in his Institutiones calculi integralis, including partial fractions and integration of rational functions.
1820s — Augustin-Louis Cauchy gives rigorous definitions of the integral as a limit of sums, putting the techniques on firm logical ground.
1854 — Bernhard Riemann formalises the Riemann integral, precisely defining which functions are integrable.
1833–1841 — Joseph Liouville proves that certain integrals (like $\int e^{- x^{2}} d x$ ) have no elementary antiderivative — explaining why no technique can succeed for them.
1968–1970 — Robert Risch develops the Risch algorithm, a decision procedure that determines whether an elementary antiderivative exists and finds it if so, now embedded in computer algebra systems.

Subtleties & Limits

Not everything can be integrated in closed form. Liouville's theorem (1830s) and the Risch algorithm (1960s) settle which elementary functions have elementary antiderivatives — but many do not. Famous "non-elementary" integrals include:

\int e^{- x^{2}} d x, \int \frac{sin x}{x} d x, \int \frac{1}{ln x} d x, \int 1 + x^{3} d x .

These are not failures of technique but provable impossibilities. They give rise to new named functions: the error function $erf (x)$ , the sine integral $Si (x)$ , and elliptic integrals. A deep practical consequence is that integration has no universal shortcut — even with the complete toolkit, success may require recognising that the answer simply cannot be written elementarily, at which point numerical or special-function methods take over. Determining the simplest closed form for a given integral remains an active area at the intersection of symbolic computation and number theory.