Artificial Intelligence 🤖
Integration
Notation & concepts

Notation & Concepts

We start off with the definition of the anti-derivative:

Given a function f(x)f(x) an anti-derivative of f(x)f(x) is a function, say F(x)F(x), whose derivative is f(x)f(x) such that

F(x)=f(x)F^{\prime}(x)=f(x)

The most general anti-derivative of f(x)f(x) is called an indefinite integral,

f(x)dx=F(x)+c\int f(x) d x=F(x)+c

where cc is the constant of integration. Here, f(x)f(x) is called the integrand. Indefinite integrals therefore are defined up to an additive constant: if F(x)F(x) is an antiderivative for f(x)f(x) so is F(x)+cF(x)+c.

Properties of the indefinite integral

I. kf(x)dx=kf(x)dx\int k f(x) d x=k \int f(x) d x where kk is any number;

II. f(x)±g(x)dx=f(x)dx±g(x)dx\int f(x) \pm g(x) d x=\int f(x) d x \pm \int g(x) d x.

Note that in integration, we do not have rules similar to the product and quotient rules in differentiation. The technique of integration followed in such cases depends on the particular integrand in question.

There are three ways of thinking about integrals:

  1. Geometrically, as the area under a graph;
  2. Analytically, as the limit of a sum;
  3. Deriving from anti-derivatives given by the Fundamental Theorem of Calculus.

In what follows, we look at each case in more detail.

1. Geometric definition

Suppose f(x)f(x) is defined for axba \leq x \leq b. The notation

abf(x)dx\int_{a}^{b} f(x) d x

with b>ab>a denotes the area under the curve y=f(x)y=f(x) and above the xx-axis for axba \leq x \leq b, as shown below.

area under the curve

If the graph falls below the xx-axis, the area above the curve and below the xx-axis counts negatively (i.e. A2A_{2} in the plot below).

area under the curve negative

Therefore,

abf(x)dx=A1A2+A3.\int_{a}^{b} f(x) d x=A_{1}-A_{2}+A_{3} .

Note that, if b<ab<a,

abf(x)dx=baf(x)dx\int_{a}^{b} f(x) d x=-\int_{b}^{a} f(x) d x

Integrals can sometimes be computed via simple geometric reasoning. Some examples are given below.

(i)

ab1dx=ba\int_{a}^{b} 1 d x=b-a

gives the area of a rectangle with sides 1 and bab-a;

(ii)

111x2dx=π2\int_{-1}^{1} \sqrt{1-x^{2}} d x=\frac{\pi}{2}

gives half the area of a circle of radius 1 ;

(iii)

ππsinxdx=0\int_{-\pi}^{\pi} \sin x d x=0

by symmetry since

 area above x-axis = area below x-axis. \text { area above } x \text {-axis }=\text { area below } x \text {-axis. }

2. Analytic definition

Try to approximate

I=abf(x)dxI=\int_{a}^{b} f(x) d x

as the area of a finite number NN of rectangles. Let x1,x2,,xN1x_{1}, x_{2}, \cdots, x_{N-1} satisfy

a<x1<x2<<xN1<b,a<x_{1}<x_{2}<\cdots<x_{N-1}<b,

and define x0=ax_{0}=a and xN=bx_{N}=b. This is shown in the next figure from which we can see that,

If(x0)(x1x0)+f(x1)(x2x1)++f(xN1)(xNxN1),=m=1Nf(xm1)(xmxm1);\begin{aligned} I & \approx f\left(x_{0}\right)\left(x_{1}-x_{0}\right)+f\left(x_{1}\right)\left(x_{2}-x_{1}\right)+\cdots+f\left(x_{N-1}\right)\left(x_{N}-x_{N-1}\right), \\ & =\sum_{m=1}^{N} f\left(x_{m-1}\right)\left(x_{m}-x_{m-1}\right) ; \end{aligned}

the above sum is known as the Riemann sum.

Again, rectangles below the xx-axis count negatively. As the intervals between successive points in xx get smaller and smaller, NN \rightarrow \infty; one can define the integral

abf(x)dx\int_{a}^{b} f(x) d x

as a limit where the number of rectangles is taken to infinity.

3. Fundamental Theorem of Calculus

The Fundamental Theorem of Calculus states that if F(x)=f(x)F^{\prime}(x)=f(x) and f(x)f(x) is a continuous function on [a,b][a, b] then,

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

We consider

a<x1<x2<<xN1<b,a<x_{1}<x_{2}<\cdots<x_{N-1}<b,

where a=x0,b=xNa=x_{0}, b=x_{N}, as above; then F(b)F(a)F(b)-F(a) can be written as

F(b)F(a)=[F(xN)F(xN1)]+[F(xN1)F(xN2)]++[F(x1)F(x0)]=m=1NF(xm)F(xm1)=m=1NF(xm)F(xm1)xmxm1(xmxm1).\begin{aligned} F(b)-F(a) & =\left[F\left(x_{N}\right)-F\left(x_{N-1}\right)\right]+\left[F\left(x_{N-1}\right)-F\left(x_{N-2}\right)\right]+\cdots+\left[F\left(x_{1}\right)-F\left(x_{0}\right)\right] \\ & =\sum_{m=1}^{N} F\left(x_{m}\right)-F\left(x_{m-1}\right) \\ & =\sum_{m=1}^{N} \frac{F\left(x_{m}\right)-F\left(x_{m-1}\right)}{x_{m}-x_{m-1}}\left(x_{m}-x_{m-1}\right) . \end{aligned}

We can approximate the term in red above by F(xm1)F^{\prime}\left(x_{m-1}\right) which in turn is equal to f(xm1)f\left(x_{m-1}\right) given that F(x)=f(x)F^{\prime}(x)=f(x). Therefore, Eq. (8.8) gives

F(b)F(a)m=1Nf(xm1)(xmxm1).F(b)-F(a) \approx \sum_{m=1}^{N} f\left(x_{m-1}\right)\left(x_{m}-x_{m-1}\right) .

Letting NN \rightarrow \infty

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

To compute the integral above, we need to find the anti-derivative, F(x)F(x) such that F(x)=f(x)F^{\prime}(x)=f(x). For instance, if we take the integral y=f(x)=x2y=f(x)=x^{2}, we note that F(x)=x3/3F(x)=x^{3} / 3 is an anti-derivative for f(x)=x2f(x)=x^{2} so that

01x2dx=F(1)F(0)=13.\int_{0}^{1} x^{2} d x=F(1)-F(0)=\frac{1}{3} .

Some common indefinite integrals are:

xndx=xn+1n+1+c,n1x1,dx=lnx+csinxdx=cosx+ccosxdx=sinx+cexdx=ex+c11+x2dx=tan1x+c11x2dx=sin1x+c\begin{gathered} \int x^{n} d x=\frac{x^{n+1}}{n+1}+c, \quad n \neq-1 \\ \int x^{-1}, d x=\ln x+c \\ \int \sin x d x=-\cos x+c \\ \int \cos x d x=\sin x+c \\ \int e^{x} d x=e^{x}+c \\ \int \frac{1}{1+x^{2}} d x=\tan ^{-1} x+c \\ \int \frac{1}{\sqrt{1-x^{2}}} d x=\sin ^{-1} x+c \end{gathered}

Definite & improper integrals

Computing definite integrals is straightforward once an anti-derivative integral is known. We have already carried out an example when discussing the Fundamental Theorem of Calculus above. For the integral

01exdx\int_{0}^{1} e^{x} d x

an anti-derivative is F(x)=exF(x)=e^{x} so that

01exdx=F(1)F(0)=e1\int_{0}^{1} e^{x} d x=F(1)-F(0)=e-1

It is customary to use the bar notation rather than the anti-derivative FF,

01exdx=ex01=e1\int_{0}^{1} e^{x} d x=\left.e^{x}\right|_{0} ^{1}=e-1

Here, g(x)ab\left.g(x)\right|_{a} ^{b} is shorthand notation for g(b)g(a)g(b)-g(a).

Earlier we interpreted

abf(x)dx\int_{a}^{b} f(x) d x

as the limit of a sum. Such definitions do not work for integrals of the form,

af(x)dx or bf(x)dx\int_{a}^{\infty} f(x) d x \text { or } \int_{-\infty}^{b} f(x) d x

These can be defined as limits of definite integrals,

af(x)dx=limbabf(x)dx or bf(x)dx=limaabf(x)dx.\int_{a}^{\infty} f(x) d x=\lim _{b \rightarrow \infty} \int_{a}^{b} f(x) d x \quad \text { or } \quad \int_{-\infty}^{b} f(x) d x=\lim _{a \rightarrow-\infty} \int_{a}^{b} f(x) d x .

These are called improper integrals. Consider

0exdx=limb0bexdx=limbex0b=limb(eb+1)=1,\int_{0}^{\infty} e^{-x} d x=\lim _{b \rightarrow \infty} \int_{0}^{b} e^{-x} d x=\lim _{b \rightarrow \infty}-\left.e^{-x}\right|_{0} ^{b}=\lim _{b \rightarrow \infty}\left(-e^{-b}+1\right)=1,

since limbeb=0\lim _{b \rightarrow \infty} e^{-b}=0.