Limit Definition of a Definite Integral

In this post I will talk about the limit definition of a definite integral. A crude definition of definite integral of a function is the area inside a function from from point $a$ to $b$. Suppose that we have a graph below.

Definite Integral
The area on shade is the definite integral of $f$ from $a$ to $b$. Since we will have difficulties on calculating the area of irregular geometry, it will be much easier if we divide the area into smaller parts with $n$ partitions (we call each partition as $x$) randomly. That way, we can approximate the area using the sum of rectangles. The smaller the parts the better approximation we can get. That way, we will have graph below.

Irregular Partitions
Point $a$ will have the same value as $x_{0}$ and point $b$ will have the same value as $x_{n}$. Note that the length of each partition is varying. I didn't even bother to use a scale drawing that! The length between each $x$ is called as subinterval, and it will be represented by a sampling point $c$. For example $c_{1}$ is in $[x_{0}, x_{1}]$, $c_{2}$ is in $[x_{1}, x_{2}]$, $c_{n}$ is in $[x_{n-1}, x_{n}]$. The $mesh$ of the partitions is the maximum value of all partition length $(\Delta x_{i}=x_{i}-x_{i-1})$, which is $mesh=\max_{1\leqslant i \leqslant n}(\Delta x_{i})$. In conclusion, the definite integral of $f$ on the interval $[a, b]$ can be defined as:
\[\int_{a}^{b}f(x)dx=\lim_{mesh\rightarrow 0}\sum_{i=1}^{n}f(c_{i})\Delta x_{i}\]
The smaller the mesh, the more accurate approximation we will get.

For convenience of computation let us consider that all subinterval will have the same partition length:
\[\Delta x_{i}=\frac{b-a}{n}\]
Then the subintervals can determined with:
\[c_{i}=a+(\frac{b-a}{n})i\]
Regular Partitions
With above consideration, we can alternatively define the integral of $f$ on the interval $[a, b]$ as:
\[\int_{a}^{b}f(x)dx=\lim_{n\rightarrow \infty}\sum_{i=1}^{n}f(c_{i})\Delta x_{i}\]
Since the length of each partition will be the same, in order to make the $mesh$ as small as possible we need to divide the area to $n\rightarrow \infty$ partitions.

Comments

Popular posts from this blog

How to Install OpenCV Package to WinPython

Preface