6.3 - Riemann Sums, Summation Notation, and Definite Integral Notation
From Approximation to Exact Area
As the number of rectangles goes up, the approximation gets better. With infinitely many rectangles, we get the exact area.
Width of Sub-intervals
For [a,b] divided into n sub-intervals:
Δx=nb−a
As n→∞, the width Δx→0, meaning the rectangles become infinitesimally thin.
Summation Notation
The sum of all rectangle areas, written out:
Δx⋅f(x1)+Δx⋅f(x2)+Δx⋅f(x3)+⋯+Δx⋅f(xn)
Each term is (width) × (height), the area of one rectangle. In summation notation:
k=1∑nΔx⋅f(xk)
Reading it:
- ∑ means “sum of”
- k=1 means start at k=1
- n means end at k=n
- Add up all the terms as k runs from 1 to n
The Limit Definition of Area
To get the exact area, take the limit as the number of rectangles goes to infinity:
n→∞limk=1∑nΔx⋅f(xk)
Components:
- Δx=nb−a is the width of each rectangle
- f(xk) is the height of the k-th rectangle
- xk=a+nb−a⋅k is the x-position of the k-th rectangle
Putting it all together:
n→∞limk=1∑nnb−a⋅f(a+nb−a⋅k)
What’s happening:
- Start at x=a
- Add k widths, so a+Δx⋅k gives the position of the k-th rectangle
- Evaluate the function there for the height
- Multiply width by height for each rectangle
- Sum them all
- Take the limit as n→∞
Alternative notation
Instead of n→∞, you can write:
Δx→0lim∑f(xk)Δx
which emphasizes that the rectangle widths shrink to zero.
The Definite Integral
That summation notation is a mouthful. Calculus gives us something much cleaner:
∫abf(x)dx
This means the exact area under the curvef(x) from x=a to x=b.
∫abf(x)dx=n→∞limk=1∑nnb−a⋅f(a+nb−a⋅k)
Parts of the notation:
- ∫ is the integral symbol (an elongated S, for “sum”)
- a is the lower limit of integration
- b is the upper limit of integration
- f(x) is the function, called the integrand
- dx says we’re integrating with respect to x
The general definition
The formulas above use equal widths and right endpoints, because that version is the easiest to write. The official definition is more flexible:
Any widths: the sub-intervals don’t have to be the same size. The i-th one has its own width Δxi.
Any point: the height can come from any point xi∗ in the i-th sub-interval (left endpoint, right endpoint, midpoint, or anywhere in between).
Sum of products: each term is the value of f at that point times the width of that sub-interval.
∫abf(x)dx=maxΔxi→0limi=1∑nf(xi∗)Δxi
Taking the limit as the widest sub-interval shrinks to 0 forces every rectangle to get thin. For a continuous function, the limit is the same no matter which widths or points you use, so the right-endpoint version with xk=a+nb−a⋅k works fine. (Left endpoints would use a+nb−a⋅(k−1) instead.)
If f is negative on part of the interval, those products f(xi∗)Δxi are negative. That’s why area below the x-axis counts as negative in a definite integral.
One summation can correspond to multiple integral forms, depending on where you decide the interval starts.
Given:n→∞limk=1∑nn6(4+n6k)2
These three integrals are all equal to that same limit:
∫06(4+x)2dx=∫410x2dx=∫016(4+6x)2dx
Form 1:∫06(4+x)2dx
- Width n6 means b−a=6
- Starting at a=0: xk=n6k
- Function: (4+x)2
Form 2:∫410x2dx
- Width n6 means b−a=6
- Starting at a=4: xk=4+n6k
- That’s exactly the expression being squared, so the function is just x2
Form 3:∫016(4+6x)2dx
- Rewrite n6=6⋅n1, so the width is n1 and b−a=1
- Starting at a=0: xk=nk
- The factored-out 6 becomes part of the function
Converting: Series to Integral
Example: a written-out sum
n2[3+n2⋅1]+n2[3+n2⋅2]+⋯+n2[3+n2⋅n]
Identify the components:
- Width: Δx=n2, so b−a=2
- With a=0, that gives b=2
- The changing part is 3+n2k, where the 3 is a constant shift
- n2k is playing the role of x, running from 0 to 2
Since xk=n2k, the expression 3+n2k is just 3+x, so f(x)=x+3:
∫02(x+3)dx
Example: with a trig function
101[cos(101)+cos(102)+cos(103)+⋯+cos(1010)]
Analysis:
- Width 101, so with n=10 we get b−a=1
- The arguments are 10k for k=1 through 10
- Those run from 101 up to 1010=1
- Starting point a=0, ending point b=1
∫01cos(x)dx
Key insight: the 10 here represents n, the number of sub-intervals. It isn’t part of the integral. Any n would give the same integral.
Careful with the limit. A sum with a fixed number of terms, like this one with n=10, is a right Riemann sum, so it only approximates∫01cos(x)dx. (The sum is about 0.818, and the integral is about 0.841.) A Riemann sum equals the integral exactly only in the limit as n→∞, so an expression that equals a definite integral has limn→∞ in front of the sum. The same goes for the written-out sum in the first example.
Example: polynomial
n5k=1∑n[(n5k)3−2]
Width: n5, so b−a=5
Starting at a=0: xk=n5k
Function: x3−2
∫05(x3−2)dx
Strategy for Conversions
Summation → integral
Identify Δx: look for the nb−a width term
Determine the interval: if you’re given the starting point, get b−a from the width; otherwise pull it from the expression
Find xk: look for the pattern a+nb−a⋅k
Extract the function: whatever xk is being plugged into becomes f(x)
Integral → summation
Identify a and b
Calculate the width:Δx=nb−a
Write xk:xk=a+nb−a⋅k
Substitute into the function: replace x with xk
Build the sum:n→∞limk=1∑nΔx⋅f(xk)
Common Patterns
Summation Element
Meaning
nc
Width of rectangles, where c=b−a
a+nb−a⋅k
Position of the k-th rectangle
∑k=1n
Sum from the 1st to the n-th rectangle
limn→∞
Take infinitely many rectangles
Key Takeaways
More rectangles means a better approximation. As n→∞, the approximation becomes the exact area
Width formula:Δx=nb−a, and as n→∞, Δx→0
Summation notation:k=1∑n(width)×(height)
Position formula:xk=a+nb−a⋅k, which tells you where to evaluate the height
Limit of a Riemann sum:n→∞limk=1∑nnb−a⋅f(a+nb−a⋅k)
Definite integral notation:∫abf(x)dx is the exact area under the curve
Equivalence: the definite integral IS the limit of a Riemann sum
Multiple forms possible: one integral can have several summation representations
The n in summations is the number of rectangles, not part of the integral
On exams: these conversions usually show up as multiple choice, so test each option
Foundation concept: integrals are the basis for everything in integral calculus