Indefinite integrals

This page is about evaluating indefinite integrals and using them to solve simple differential equations.

Contents

Antidifferentiation

A property of indefinite integrals is that they are antiderivatives, which “undo” differentiation.

Eg. 1xdx=lnx+C\displaystyle \int\frac{1}{x} \d x = \ln\lvert x\rvert + C, where CC is any constant of integration. Here 1x\displaystyle \frac{1}{x} is the integrand and lnx+C\ln\lvert x\rvert + C is the antiderivative or indefinite integral.

Usually, the +C+C is worth one mark!

The +C+C indicates that antiderivatives are not unique, as many answers are returned off by a constant, any constant.

There is also the notation

dxf(x)=1f(x)dx\int \frac{\d x}{f(x)} = \int \frac{1}{f(x)}\d x

Common antiderivatives

Reading the tables in Common derivatives and More derivatives (HL) in reverse yields common antiderivatives.

Here are common antiderivatives you should know

functionantiderivatives
kk, not a function of xxkx+Ckx + C
xn,n1x^n, n \neq -11n+1xn+1+C\displaystyle \frac{1}{n+1}x^{n+1} + C
x1x^{-1}lnx+C\ln\lvert x\rvert + C
ex\e^xex+C\e^x + C
sinx\sin xcosx+C-\cos x + C
cosx\cos xsinx+C\sin x + C

The absolute value in lnx+C\ln\lvert x\rvert + C aims to give the antiderivatives the same domain as the integrand.

Also remember

cos2x=2cos2x1=12sin2x\cos2x = 2\cos^2x-1 = 1-2\sin^2x

or that

cos2x=1+cos2x2\cos^2x = \frac{1 + \cos2x}{2}
sin2x=1cos2x2\sin^2x = \frac{1 - \cos2x}{2}

for integrating cos2x\cos^2 x and sin2x\sin^2 x.

At HL, the formula booklet also lists the following integrals

functionantiderivatives
axa^x1lnaax+C\displaystyle \frac1{\ln a} a^x + C
1a2+x2\displaystyle \frac{1}{a^2 + x^2}1aarctan(xa)+C\displaystyle \frac1a \arctan\left(\frac xa\right) + C
1a2x2\displaystyle \frac{1}{\sqrt{a^2 - x^2}}arcsin(xa)+C,x<a\displaystyle \arcsin \left(\frac xa\right) + C,\, \lvert x\rvert < \lvert a\rvert

Properties

Unlike differentiation, there lack formulas for antiderivatives of composite functions, or product or quotient of functions, because there lack differentiation formulas resulting in such functions.

Instead, analogous properties involving constants are presented below. F(x)F(x) is an antiderivative of f(x)f(x).

f(x)+g(x)dx=f(x)dx+g(x)dx\int f(x) + g(x)\d x = \int f(x) \d x + \int g(x) \d x
kf(x)dx=kF(x)+C\int k f(x)\d x = kF(x) + C
f(ax+b)dx=1af(x)dx=1aF(ax+b)+C\int f(ax + b)\d x = \frac 1a \int f(x)\d x = \frac 1aF(ax + b) + C

The last one is a special case of integration by substitution.

Continuous functions are always integrable (meaning antiderivatives exist), even if we often cannot easily evaluate them.

Example: Simplify

(4x+13+14x+13)2dx\int \left(\sqrt[3]{4x+1} + \frac{1}{\sqrt[3]{4x+1}}\right)^2 \d x

Rewrite radicals (surds) as fractional exponents. Then expand using (a+b)2=a2=2ab+b2{(a+b)^2 = a^2 = 2ab + b^2}

integrand=((4x+1)13+(4x+1)13)2=(4x+1)23+2+(4x+1)23\begin{align*} \text{integrand} &= \left((4x+1)^{\frac13} + (4x+1)^{-\frac13}\right)^2 \\ &= (4x+1)^{\frac23} + 2 + (4x+1)^{-\frac23} \\ \end{align*}

Now, use xndx=1n+1xn+1+C{\int x^n \d x = \frac{1}{n+1}x^{n+1} + C} and f(ax+b)=1aF(ax+b)+C{\int f(ax + b) = \frac1a F(ax + b) + C} to integrate term by term.

(4x+1)23dx=14(23+1)(4x+1)23+1+C1=320(4x+1)53+C1\begin{align*} \int (4x+1)^{\frac23} \d x &= \frac1{4\left(\frac23 + 1\right)} (4x + 1)^{\frac23 + 1} + C_1 \\ &= \frac{3}{20}(4x + 1)^{\frac53} + C_1 \end{align*}
2dx=2x+C2\begin{align*} \int 2 \d x &= 2x + C_2 \\ \end{align*}
(4x+1)23dx=14(23+1)(4x+1)23+1+C3=34(4x+1)13+C3\begin{align*} \int (4x+1)^{-\frac23} \d x &= \frac1{4\left(-\frac23 + 1\right)} (4x + 1)^{-\frac23 + 1} + C_3 \\ &= \frac{3}{4}(4x + 1)^{\frac13} + C_3 \end{align*}

Putting everything together

I=(4x+13+14x+13)2dx=320(4x+1)53+2x+34(4x+1)13+C\begin{align*} I &= \int \left(\sqrt[3]{4x+1} + \frac{1}{\sqrt[3]{4x+1}}\right)^2 \d x \\ &= \frac{3}{20}(4x + 1)^{\frac53} + 2x + \frac{3}{4}(4x + 1)^{\frac13} + C \qed \end{align*}

The constants C1C_1, C2C_2 and C3C_3 were collected into CC.


Note that we used the real root, as opposed to the complex principal root [HL] when interpreting 4x+13\sqrt[3]{4x+1}. This is typically the case as in this course we only work with real functions.

Linear over linear integration

There is a small chance that all candidates may be asked to integrate a linear over linear rational function.

The general strategy is to convert it to a constant +rational\text{constant } + \text{rational} form.

Example: Simplify

4x+12x+3dx\int \frac{4x + 1}{2x + 3} \d x

Because 2(2x+3)=4x+62(2x+3) = 4x + 6 and 4x+65=4x+14x + 6 - 5 = 4x + 1, then

4x+12x+3=2(2x+3)52x+3=252x+3\begin{align*} \frac{4x + 1}{2x + 3} &= \frac{2(2x+3) - 5}{2x+3} \\ &= 2 - \frac{5}{2x + 3} \end{align*}
252x+3dx=2x52ln2x+3+C\begin{align*} \int 2 - \frac{5}{2x + 3} \d x &= 2x - \frac{5}{2} \ln\lvert2x+3\rvert + C\qed \end{align*}

where we invoked the property of dividing by 1a\frac{1}{a} when integrating f(ax+b)f(ax + b).

For integrating linear over quadratic functions, HL candidates should refer to rational functions integration.

Simple differential equations

With antiderivatives, you can solve differential equations of the forms

dfdx=g(x)\frac{\d f}{\d x} = g(x)

or

d2fdx2=g(x)\frac{\d ^2f}{\d x^2} = g(x)

As well as at HL,

dndxnf(x)=g(x)\frac{\d ^n}{\d x^n} f(x) = g(x)

This means, given the 11st, 22nd, or nnth derivative as g(x)g(x), find the original function f(x)f(x).

Note that this technique cannot be used if there are multiple derivative functions, ie both dfdx\frac{\d f}{\d x} and d2fdx2\frac{\d ^2f}{\d x^2} in the same equation. The technique also fails when there are variables on the right side other than the variable differentiating with respect to.

Examples of differential equations that this could solve

dydx=2x\frac{\d y}{\d x} = 2x
d2xdt2=e3t\frac{\d ^2x}{\d t^2} = \e^{-3t}

Examples of differential equations that could not be solved this way

dmdx=m\frac{\d m}{\d x} = m
d2xdt2+2dydx=te3x\frac{\d ^2x}{\d t^2} + 2\frac{\d y}{\d x} = t \e^{-3x}

The solution to a differential equation is a function. In addition, nn boundary conditions, one for each level of differentiation, should be given to pin down the answer to a specific function.

Example: The position of a particle in free fall satisfies the equation

d2ydt2=9.8\frac{\d ^2y}{\d t^2} = -9.8

where yy is the height above Earth’s surface.

Given y(0)=u\displaystyle y^\prime(0) = u and y(0)=y0y(0) = y_0. Find y(t)y(t) while the particle experiences gravity.


Integrating both sides once gives

y(t)=9.8t+Cy^\prime(t) = -9.8t + C

using y(0)=u\displaystyle y^\prime(0) = u results in

y(0)=9.8(0)+C=uC=uy(t)=9.8t+u\begin{align*}y^\prime(0) &= -9.8(0) + C = u \\ C &= u \\ y^\prime(t)&= -9.8t + u \end{align*}

Integrate again gives

y(t)=4.9t2+ut+Dy(t) = -4.9t^2 + ut + D

Using a different constant of integration as it’s a different integral in the same problem. Using y(0)=y0y(0) = y_0 we get

y(0)=4.9(0)2+u(0)+D=y0D=y0y(t)=4.9t2+ut+y0\begin{align*}y(0) &= -4.9(0)^2 + u(0) + D = y_0 \\ D &= y_0 \\ y(t) &= -4.9t^2 + ut + y_0 \qed \end{align*}

Notes on absolute values

If the integrand contains absolute value f(x)\lvert f(x)\rvert, then you must find a way to re-express the integrand without absolute values, probably after some simplification first. You may need to find the zeros, and break into domains of positive or negative f(x)f(x), or use some property of the specific functions used. Do not try to invent absolute value function integration rules.

For integrating factor (HL), where applicable, use

e1xdx=elnx=x\e^{\int \frac1x \d x} = \e^{\ln x} = x