Speaker 1: Hi, my name is Gautam. I am thankful to NPTEL for allowing me to teach a course called Decision Making Under Uncertainty. This falls under the broad umbrella of operations research and management science. We have all heard words such as descriptive analytics, predictive analytics, prescriptive analytics. This course basically is a topic that covers all these aspects. However, this is very much at the introductory level. We are going to cover a lot of breadth and we will talk about a variety of applications. So the audience that is appropriate for this course go all the way from engineering to business. This course has a prerequisite of elementary probability and statistics. In particular, something that involves a little bit of calculus. And this course will help you develop foundations in decision making through the use of data science. We will look at examples from very familiar situations such as topics like retail. We all go to stores all the time. Sports, we all love sports. TV, we all watch TV every time. Food, of course, who does not eat? And we will look at a lot of personal life situations. So, situations that we are all familiar with, we will discuss various decision making strategies and in all these situations, there will be some level of uncertainty. Now, the topic outline for this course is as follows. In week 1, we will start with some background and introduction. We will talk about some example situations where we would be having to make some decisions under uncertainty and then talk a little bit about topics such as probability, random variables, expectations, variance, etc. After that, we will talk about optimization criteria and objective functions and wrap up that week. In week 2, we will start the course with what is called one time decisions. These are one of the hardest decisions. Why are they hard? Because you have to make some major decisions just once. For example, we will talk about the famous secretary problem or also called the marriage problem where you have to make a decision of who to hire or who to get married to. After that, we will move on to a topic called utility function. This allows us to make sure that we have various objectives trapped into a single function. And after that, we will talk about the topic called decision trees where we make decisions as we go along and these could be nested along the way. Then, we will discuss the aspect of making decisions in game shows, popular ones that you see on TV. Then, we will wrap up the first, I am sorry, the second week with what is called project evaluation. In week 3, we will look at repeated decisions, okay. This is such as when a news vendor like the person who sells newspapers in the corner of two streets, they have to decide how many newspapers do I stock at the beginning of the day, every single day. After that, we will look into the aspect of buffering for both uncertainty as well as variability and this will be dealt with at a very qualitative basis. Then, we will discuss the notion of carrying safety stocks for inventory in retail type of situations. After that, we will have a discussion about how one goes about planning for routes, especially this day and age everybody has GPS, it gives you all kinds of information back to you, how do you go about making decisions about what is the best route to go from point A to point B. After that, we will wrap up week 3 with the notion of exploration versus exploitation. This is where you first learn about your environment and then you exploit, okay. That brings us to end of week 3. Then, in week 4, we will look at what is called sequential adaptive decisions, okay. This is a topic where as you go along, the uncertainty gets revealed and your environment is known and you take actions. We will start with the notion of stochastic programming where we will make strategic decisions followed by some tactical decisions. After that, we will visit the topic called Simpson's paradox and then move on to Markov decision process. This is kind of like feedback control. However, these topics could potentially be taught at an extremely high level, but we will bring this down to a very introductory level with the understanding that once you learn about these topics, you can take up courses on some of these other topics in the future. Now, I have to say one thing that differentiates this course from some of the other course that I have at least taught. One is the topics are such that they are very independent. One topic has very little to do with the next topic, is very little to do with the third topic and so on. So, as a student, the nice thing about that is that if one topic is a little bit rough, you can say, okay, this will end and something nice will happen next and then you will go back and start to work on that topic and do well. Now, this course uses not only a little bit of skills in mathematics, it also uses some software. In particular, I am going to be using a freeware called Octave and Octave is exactly like MATLAB except it is available for free. I will give you some example codes in Octave that you could use throughout the course. At the end of the course, as a student, you could ask, so what am I going to get after learning this material? And you could, well, this is my answer to that question. I would say, well, you will develop a few skills such as you will develop a probabilistic intuition, you will also gain strategies to manage uncertainty and you will be able to make effective decisions. I sincerely hope that you will consider taking this course and come with me to take this journey of how one goes about making decisions under uncertainty. I welcome you to the course. Thank you.