We’re going to think a bit about integration with a twist: what happens when our “definite” integrals can’t actually be evaluated? First, let’s try to sink ourselves back into the context we’ve been in for a while now: what kinds of problems have we encountered so far, and how do we use our calculus intuition to get around those problems?
Let’s say that we want to find what the \(y\)-values of some function \(f(x)\) are when the \(x\)-values are “infinitely close to” some value, \(x=a\text{.}\) Since there is no single \(x\)-value that is “infinitely close to” \(a\) that we can evaluate \(f(x)\) at, we need to do something else. How do we do this?
Let’s say that we want to find the rate of change of some function instantaneously at a point with \(x=a\text{.}\) We can’t find a rate of change unless we have two points, since we need to find some differences in the outputs and inputs. How do we do this?
Suppose you want to find the total area, covered by an infinite number of infinitely thin rectangles. You have a formula for finding the dimensions and areas for some finite number of rectangles, but how do we get an infinite number of them?
So moving forward, we want to remember how we typically have solved these problems. Now, let’s try to identify the types of problems with integrals that we need to figure our way around.
Activity7.1.2.Remembering the Fundamental Theorem of Calculus.
We want to think about generalizing our notion of integrals a bit. So in this activity, section, we’re going to think about some of the requirements for the Fundamental Theorem of Calculus and try to loosen them up a bit to see what happens. We’ll try to construct meaningful approaches to these situations that fit our overall goals of calculating area under a curve.
This practice, in general, is a really good and common mathematical process: taking some result and playing with the requirements or assumptions to see what else can happen. So it might feel like we’re just fiddling with the “What if?” questions, but what we’re actually doing is good mathematics!
What do we need to be true about our setup, our function, etc. for us to be able to apply this technique to evaluate \(\displaystyle \int_{x=a}^{x=b} f(x)\,dx\text{?}\)
We are going to introduce the idea of “Improper Integrals” as kind-of-but-not-quite definite integrals that we can evaluate. They are going to violate the requirements for the Fundamental Theorem of Calculus, but we’ll work to salvage them in meaningful ways.
This should build a pretty good idea of a new “class” of integrals: ones that aren’t quite definite integrals that we can evaluate with the Fundamental Theorem of Calculus, but ones that we can use limits to get at.
An integral is an improper integral if it is an extension of a definite integral whose integrand or limits of integration violate a requirement in one of two ways:
The interval that we integrate the function over is unbounded in width, or infinitely wide.
With this definition, we can think about the strategies that we got from Activity 7.1.1: we’re going to identify the problems in our integral (infinite width of the interval or infinite height of the integrand function) and use a limit!
As long as \(t\) is some real number with \(t\gt 2\text{,}\) then our function is continuous and bounded on \([2,t]\text{,}\) and so we can evaluate this integral:
Ok, we’re going to switch our focus to the other improper integral,\(\displaystyle \int_{x=-1}^{x=2} \frac{1}{(x+1)^2}\;dx\text{.}\) again, we’ll look at a slightly different integral:
As long as \(t\) is some real number with \(-1\lt t \lt 2\text{,}\) then our function is continuous and bounded on \([t,2]\text{,}\) and so we can evaluate this integral:
We’re going to evaluate this intergal for different values of \(t\) again, but this time we’ll use values that are close to \(-1\text{,}\) but slightly bigger, since we want to be in the interval \([-1,2]\text{.}\)
Let’s start with making \(t=-\frac{9}{10}\text{.}\) So we’re going to evaluate:
SubsectionStrategies for Evaluating Improper Integrals
Evaluating Improper Integrals (Infinite Width).
For a function \(f(x)\) that is continuous on \([a,\infty)\text{,}\) we can evaluate the improper integral \(\displaystyle\int_{x=a}^\infty f(x)\;dx\text{:}\)
Finally, if \(f(x)\) is continuous on \((-\infty,\infty)\) and \(m\) is some real number, then we can evaluate the improper integral \(\displaystyle\int_{-\infty}^{\infty}f(x)\;dx\text{:}\)
Try to interpret this limit. What does it mean if this limit doesn’t exist? What does it mean if the limit does exist? What does the actual number represent?
For a function \(f(x)\) that has an unbounded discontinuity (a vertical asymptote) at \(x=m\) with \(a\lt m\lt b\text{,}\) but is otherwise continuous on \([a,b]\text{,}\) then we can evaluate the improper integrals:
Try to interpret this limit. What does it mean if this limit doesn’t exist? What does it mean if the limit does exist? What does the actual number represent?
Ok, let’s note that we can classify these improper integrals into two categories. We have already classified them based on the reason that they’re improper. But now we also can classify them based on the outcome of the limit:
Improper integrals (of any type) whose limit exists.
SubsectionConvergence and Divergence of an Improper Integral
Definition7.1.4.Convergence of an Improper Integral.
We say that an improper integral converges if the limit of the appropriate definite integral exists. If the limit does not exist, then we say that the improper integral diverges.
All we’ve done here is added some language: we’ll say that an improper integral diverges if the limit doesn’t exist. And if the limit exists, we’ll say that the improper integral “converges to .”
One of the big ideas in probability is that for a curve that defines a probability density function, the area under the curve needs to be 1. What value of \(k\) makes the function \(\dfrac{kx}{(x^2+3)^{5/4}}\) a valid probability distribution on the interval \([0,\infty)\text{?}\)
Second, use this inequality to compare the function \(\dfrac{\sqrt{x^2+1}}{x^2}\) to \(\dfrac{1}{x}\) for \(x\gt 0\text{:}\) which one is bigger? Again, use your inequality from above to help!
Now compare \(\displaystyle \int_{x=1}^\infty \dfrac{\sqrt{x^2+1}}{x^2}\;dx\) to \(\displaystyle \int_{x=1}^\infty \dfrac{1}{x}\;dx\text{.}\) Which one is bigger?
Explain how we can use this result to make a conclusion about whether our integral, \(\displaystyle \int_{x=1}^\infty \dfrac{\sqrt{x^2+1}}{x^2}\;dx\) converges or diverges.