This section covers:
Introduction to Riemann Sums
I’m convinced the reason they teach you Riemann Sums is to have you “appreciate” what our former mathematicians had to go through before things got easier. I’m the first to admit that I’m not a fan of working with them, just because they are so tedious.
Riemann Sums can be used to approximate the area under curves, which will be acquired much easier by just taking the integral of the function between two different x values (we’ll do this in the Definite Integral section). But, alas, we have to learn these more difficult methods first.
So let’s say we wanted to get the area of the region between x = 1 and x = 2 for certain “curved” functions. To get an approximation, we could just add up little rectangles that we can form between x = 1 and x = 2 that are under the curve and above the xaxis. And do you see how the more rectangles we use (the “n”), the more accurate the total area will be?
Do you also see how, depending on whether the upper left or upper right (or midpoint) of the rectangles touch the curve, we’ll get slightly different areas? Do you see how, for some functions, the “right hand” sums (where the right hand sides of the rectangles touch the function) overapproximates the area (upper sum), and for others, it underapproximates the area (lower sum)? Note that “n” represents the number of rectangles.
So how does all this relate to Calculus? Soon we’ll learn to use Integration to get the area under a curve, but again, to “appreciate the math”, we’ll first learn how to approximate the area using upper, lower, and midpoint sums.
Using Upper and Lower Sums to Approximate Area
Here’s an example of using a lower sum to estimate area; in this case, it’s a lefthand sum, since the upper left part of the rectangle touches the curve. The notation can be scary looking, but it’s not that bad. Let’s use the example of \(y={{x}^{2}}\) between x = 1 and x = 2, with n = 8 (number of intervals).
Using Midpoint Rule to Approximate Area
The midpoint rule uses sums that touch the function at the center of the rectangles that are under the curve and above the xaxis. We compute the area approximation the same way, but evaluate the function right in between (the midpoint of) each of the rectangles; this will be height of the rectangles (the “y”). We’ll compute the Δx the same way. We’ll see an example below.
Upper, Lower, and Midpoint Sums Problems
Here are examples of upper, lower, and midpoint sums:
Trapezoidal Rule
The Trapezoidal Rule approximates area, but uses trapezoids instead of rectangles. Remember that a trapezoid is a quadrilateral (foursided figure) where the bases are parallel and the other two sides aren’t.
Remember from Geometry how we get the area of a trapezoid; here’s how we’ll use it in Calculus:
So we’ll want to create little trapezoids (with bases as y values) under the curve and above the xaxis between two x points, and then add up all those areas. When we do this, we come up with the definition of the Trapezoidal Rule:
The Trapezoidal Rule:
Let f be continuous on interval \(\left[ {a,\,b} \right]\). We can approximate \(\int\limits_{a}^{b}{{f\left( x \right)}}\,dx\) by the Trapezoidal Rule:
\(\displaystyle \begin{align}\int\limits_{a}^{b}{{f\left( x \right)}}\,dx\,\,\approx \,\,\frac{{ba}}{{2n}}\left[ {f\left( {{{x}_{0}}} \right)+2f\left( {{{x}_{1}}} \right)+…+2f\left( {{{x}_{{n1}}}} \right)+f\left( {{{x}_{n}}} \right)} \right],\,\,\\\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\text{where }{{x}_{0}}=f\left( a \right),\,\,{{x}_{1}}=f\left( {a+\frac{{ba}}{n}} \right),\,\text{and so on}\end{align}\)
This is because each trapezoid is “on their side” and their height is \(\frac{{ba}}{n}\) and bases are consecutive \(\displaystyle f\left( x \right)\) values. When we add up all the “averages” of the bases \(\frac{{f\left( {x{}_{1}} \right)+f\left( {x{}_{2}} \right)}}{2},\frac{{f\left( {x{}_{2}} \right)+f\left( {x3} \right)}}{2}\) and so on, we end up with dividing by 2 (thus the \(\frac{{ba}}{{2n}}\) on the outside) and having twice all the \(\displaystyle f\left( x \right)\) values, except for the first and last.
You may be given the actual function, or you may be given values at certain points so you’d have to “build” the trapezoids. Let’s do some problems:
Here are a few more Trapezoidal Rule problems. Note that when we have “n” intervals, we will get “n + 1” f(x) values.
Area by Limit Definition
The area by limit definition takes the same principals we’ve been using to find the sums of rectangles to find area, but goes one step further. We’ll be finding the area between a function and the xaxis between two x points, but doing it in a way that we’ll use as many rectangles as we can by taking the limit of the number of rectangles as that limit goes to \(\infty \). This way we’ll get the most accurate area we can, since the more rectangles we have, the closer we get the true area.
Again, this is a tough concept to grasp, but we’ll just use a formula that (hate to say it!) you’ll want to memorize how to use, as shown below. Note that the sigma sign \(\Sigma \) means to start with the first value, plug it in and go up to the last value, and taking the sum of all of those terms. Notice from the picture that this formula is closest to the midpoint rule.
Here’s the area by limit formula:
Definition of the Area of a Region by a Limit
Let f be continuous and above the yaxis (nonnegative) on interval \(\left[ {a,\,b} \right]\). The area of the region bounded by f, the xaxis and vertical lines at \(x=a\) and \(x=b\) is: \(\displaystyle \text{Area}=\underset{{n\to \infty }}{\mathop{{\lim }}}\,\sum\limits_{{i=1}}^{n}{{f\left( {{{c}_{i}}} \right)\cdot \Delta x}}\), where \(\Delta x=\frac{{ba}}{n}\), and \({{x}_{{i1}}}\le {{c}_{i}}\le {{x}_{i}}\).
Note that we will learn later that this area is \(\int\limits_{a}^{b}{{f\left( x \right)}}\,dx\). 

Here are the formulas you’ll really have to use:To simplify and get rid of summation signs, use these summation formulas (usually given):What you want to do is use the area formula with the given function and interval, then simplify as much as you can. Separate the summations, if needed, and then leave only the “i’s” in the summation (by moving everything else to the outside). Then use the above summation formulas to turn “i’s” into “n’s”. Then simplify and take the limit of what’s left (remember that \(\underset{{n\to \,\infty }}{\mathop{{\lim }}}\,\frac{a}{{{{n}^{r}}}}\,\,=\,\,0\)) . I know it’s tedious, but it works!Let’s show an example to see how this works:Here’s another Area by Limit Definition problem. This can get a little messy, so you have to be careful. And I promise you that we’ll have a much easier way to get these areas soon! They just want you to “appreciate the math” at this point:Here’s a tricky problem on area by limit definition (sums) that you might see on a test. And (spoiler alert!), you’ll see in the possible answers how we’ll do these problems so much easier using Definite Integrals in the next section.Understand these problems, and practice, practice, practice!
On to Definite Integration – you’re ready!