- Tubelator AI
- >
- Videos
- >
- Education
- >
- Understanding Linear Stochastic Programming Problems: Class 03 Overview
Understanding Linear Stochastic Programming Problems: Class 03 Overview
Explore the concepts of linear stochastic programming problems in this class lecture. Learn about formulating models with recourse, using two stages for decision making, and tackling uncertainty. Get ready to answer homework questions during the session.
Install Tubelator On Chrome
Video Summary & Chapters
No chapters for this video generated yet.
Video Transcript
OK.
Hello, everybody.
Welcome to the third lecture.
Today we'll speak about a specific class
of stochastic programming problems that are linear.
Constraints and the objective is linear,
and is formulated in the form of a model with recourse.
And that it is called in two stages,
referring to the here and now variables and the wait
and c variables. Here are now variables at first stage and weight and c variables will
be second stage that can be decided once the realization of uncertainty becomes known.
Ok, the usual.
So there are some questions about the homework that was sent and that I will display here
so that during the class you have time to think about the answer and then at the end
of the class we will talk about this, we will answer these questions.
Another thing that it's important to know is that this homework is mandatory for PhD
these students about the date of when it will have to be
delivered.
We were thinking about after Easter.
So not before Easter.
Not too much after Easter either.
So the week after Easter.
We will fix the date next week.
So it's good to start working on that if you have to or if
are interested in doing this.
So the two questions refer to how.
how to model the worst-case scenario,
and what can be done about the scenario
analysis with the solution.
The first question is, well, for the worst-case scenario,
is it OK to take a 99% confidence interval
and then add the maximum value on the interval of the oil
demand?
Is this the model we are thinking about?
So this is one question.
So you can think, or maybe if you
have the answer, you can send it through the chat in YouTube.
And then scenario analysis.
Remember I said this, in fact, scenario analysis,
in spite of being popular because it's easy,
it is rather useless because it is answering
to the wrong question.
It's not what we want to model, just one
realization of uncertainty.
And then the question is, how do we
do we evaluate and assess the quality?
of the solution provided by this methodology.
Because since we run k times, if we have k scenarios,
we will solve k different linear programs.
Then what is the response that we
