So, I'll come back, everyone, right, everyone, obviously, you know, act natural, right,
left it by usually hilarious jokes, anyway, a few orders of business, we can start it.
One, we've got the homework due tonight, and so hopefully no one's getting started right now.
And number two, project proposals will do end up next week.
I believe in this point, most people have come to talk about projects.
If you haven't done so, make sure to come in the next two opportunities of office hours,
because that's what we have left.
So, right, Tuesday is going to be online, Thursday is going to be in person.
And lastly, we have right into, not next week, the following week, Tuesday we're not meeting.
It'll be out for our conference, that'll be facing for this lecture.
So, try to pre-record that and have that online by lecture time.
What would have been lecture time on Tuesday?
Okay, there are questions about logistics before we get started.
Okay, great.
So, on Tuesday, we started making the transition from finite horizon problems to infinite horizon problems.
And there are a few kind of key insights that we developed to try and think about how we solve infinite horizon problems
and how they might be different from finite horizon.
So, if we think about that, what was the main idea of solving infinite horizon problems
that was different from finite horizon problems?
Well, I should say, why are finite horizon problems easier to solve than infinite horizon problems?
So, everyone's scared?
In the finite horizon, we can start at the terminal state and proceed backwards to find the optimal horizon for each state.
Perfect, right?
So, in the finite horizon, we can always start at the end and do the beginning, because there's only a finite amount of time.