Last time | Next time |
Today's lab continues with sequences and series, which is our current theme.
If you can't make the office hours, just email and I'll find some time to talk things over.
Here's a key.
A few comments:
There are two kinds of convergence/divergence at work here. There is the convergence of functions, and there is the convergence of integrals.
In order for an improper integral (with a limit of integration of $\infty$, say) to converge, the integrand must converge (asymptote) to 0.
We compare integrals to integrals. In order for an integral to converge, it will be necessary for an integrand to converge, too.
When comparing integrals, each result only has one "working side". For now, assume all functions are positive. If $f(x)\ge g(x)$ everywhere, and \[ A=\int_a^\infty f(x) dx \]
converges, then so must \[ B=\int_a^\infty g(x) dx \]
because \[ 0 \le \int_a^\infty g(x) dx \le \int_a^\infty f(x) dx \]
But if $A$ diverges, then we can't say anything about $B$ -- since $B$ may be small enough to converge (or it may diverge; we don't know yet!).
On the other hand, if $f(x)\le g(x)$ everywhere, and \[ A=\int_a^\infty f(x) dx \]
diverges, then so must \[ B=\int_a^\infty g(x) dx \]
because \[ 0 \le \int_a^\infty f(x) dx \le \int_a^\infty g(x) dx \]
But if $A$ converges, then we can't say anything about $B$ -- since $B$ may be small enough to converge (or it may diverge; we don't know yet!).
One of the interesting things I wanted to point out here is that we can turn an integral with an infinite domain (the $u$ integral) into one with a finite domain (the $x$ integral); and we know how to use the midpoint rule to approximate the finite domained one:
\[ \int_{0}^{e}x (\ln(x))^2 \ dx = \int_{-\infty}^{1}u^2 e^{2u} \ du \]
So this is a trick to allow us to compute numerical approximations of the infinite domained improper integrals: turn them into finite-domained ones, via some transformation (in this case $u=\ln(x)$).
A lot of folks struggled with that request. I simply wanted you to write \[ p(x)=10^{stuff} \] as \[ p(x)=\left(e^{\ln(10)}\right)^{stuff}=e^{\ln(10)stuff} \]
In calculus class we prefer to use base-e....
Make sure that you understand the difference:
We can think of it as a function $a(x)$, but with integers for arguments, and real numbers for values: \[ a: \mathbb{N} \rightarrow \mathbb{R} \]
It might be better simply to write it as $a(n)$, but mathematicians have adopted the alternative notation $a_n$, perhaps to make clear that this function has a domain restricted to integers.
And this is one of the reasons that we are not too surprised when we find ourselves comparing series and improper integrals.