 All right, welcome back everyone. Hopefully you had some good discussion with your other peers and other learners. So while I'm going to talk for just a minute or two, but this will give you time to enter in on chat what your business problem was, what was the problem that you faced. So then we'll talk about reading and we'll say what we think, but then it'd be great to get your feedback. Before we do that, now's the time for you to write that in, let me make two comments. First, one of the things that we wanted to stress and we try to stress throughout the course, it's important for you to understand these fundamental methodologies that we've taught and not just have the spreadsheets or the models to mimic. It's an easy thing to try to mimic and that's what we try to do in the midterms and the final to make sure you're not just mimicking what we do because that'll only take you so far. If we give you a linear program and all of them fit into that same profile and all you do is change the numbers, you're not really getting the fundamentals of it. So I encourage you, and I always do this, even though I have all the spreadsheets for different ones, I would start from scratch because it's better to understand exactly what's going and it's a good way to make sure you understand what's happening. So I encourage you as you go through the courses, try to always take things a little further because the problem you get in practice will never be exactly the one that you were shown in the class. So it's important you understand how to apply and truly understand the concepts, not just the ability to mimic and plug in numbers in an existing spreadsheet or model. The second thing, a question that came up, someone asked about vehicle routing. Actually they asked about the shortest routing algorithms and what is used in practice. There's a whole series of different tools out there and so you can think of the three models I think I talked about were shortest path, traveling salesman and vehicle routing and traveling salesman and vehicle routing are really together. No one really solves the traveling salesman problem by itself. It's kind of a component of the vehicle routing, the BRP. And for shortest path, actually that's embedded in almost every mapping system you have. I don't know any companies that really invest in developing their own shortest path anymore, that's already embedded in a lot of systems. Vehicle routing though is a little bit different, it's tricky. And that's an area that's still evolving because people are coming up with more innovative ways of doing this because you can think of that problem, there's actually whole literatures that go into this. And this is what your dissertation deals with different people routing. Because you have vehicle routing under uncertainty with schedules. And there's millions of these problems and they have like- That's why we call it a family of routing problems. And so are there package software out there? Sure, UPS uses some, there's some available out there. Chances are that if you're doing this, you will probably end up doing something almost customized. We have a lab here called the Megat City Logistics Lab that looks at this is one of the areas they look at, how do you routing in urban areas? And so it's an open area, there isn't one set package. Like when we look at a mixed engineering program, we can tell you a handful of ones to use for vehicle routing. They're usually specialized because the situations differ. And if you are interested, you can go to our YouTube channel. So I had one conversation with Matthias before. Oh, great. Yeah, so it's there. You can watch that YouTube. So it's about less monthly degree problems, which is totally fit to the working problems. That's with Dr. Matthias Winkenbach, who heads up the Megat City Logistics Lab. Yeah. So I have one that comes from William Casper. And he says that they use regression for modeling and simulation to develop the staffing scenarios for startup consulting. Boom. What do you think? Yeah, I mean, I've done this before because for a startup, I was part of a software company startup. And the thing with that, all the forecasting stuff you'll learn in SC1X, for the most part, assume that you have some history. And in startups, you have no history. So you're probably going to start with, you do a pilot, and you know an engagement that, say, to implement software. You learn how many people it takes. And then you have to start guessing, okay, how many implementations do I think I'll have, and do they differ by different companies? We found that you have different times to implement based on whether it's a consumer packaged good manufacturer versus a heavy equipment manufacturer. So as you get more data, and you can start looking at correlations, you can start building a more accurate model out of this to see how many people I should staff. And it's going to be a function of the type of engagement, the number of engagements, all these other factors. And so I've used regression before for that, that makes sense. We actually use regression because one of the tools we're developing, saved company's money. And so we do it by running regression over what their current situation would be versus the new solution. We see what that gap would be. And we actually found in one company that they wouldn't save any money. And that forced us to go back and look at our regression better. And we found that we weren't really capturing it that correctly. So you can use these tools iteratively to see how well you're doing. Work very well. Great. So we have another question from Hima. She said that I'm about to start working in Japan's firm raw materials order, come from Japan to USMIC, and USMIC the product and the distributes. Of course, I have a lot of historical data. How should I approach to start making things better since the beginning, in terms of the historical data and the amount of data they have? So hopefully you can answer your own question, I hope. But how I would do it, so I have a bunch of them to make sure I understand. Raw materials coming from Japan, probably by ocean to the US. And then in the US, it's getting shipped to some manufacturing plant, I assume, from a port. And then it's being moved into distribution channels, whether that's into separate distribution centers, BCs, or to retailers or something like that. So you'll learn a little bit about, there's many different ways you can make it better. The first thing you might want to do is something called stapling yourself to an order. And so you can look at the transactional data and follow one transaction. Follow all the way from when an order is sent to the raw materials, how that product moves, and follow how it goes through the system, all the way through to final and delivery to customer. That does a couple things that helps you understand the breadth and all the touch points. The other things you can do is look at the data and try to understand certain key characteristics. For example, if you know the time, how long does it take for an order to come through? You understand where the problems are. Because I think the first thing you need to do is understand the problem. The second is to see, okay, where are the pain points? Is it that it's taking too long? Is it that it's expensive? Is it damaging things? What is the problem? Or is there a problem? Where can it be improved? Then you focus in on, okay, which of those things do I want to improve? The methodologies that we talked about, optimization, probability and statistics, simulation regression, come in on that third step when I want to look at some of these things. So one thing that you'll learn SC1X is looking at the transportation of that multi-segment move. And maybe you can find out ways to minimize the transit time. So maybe you can find ways of moving things in a more efficient manner or where to focus in. You might have bad stocking levels. So you can see what the impact of, if you go by ocean, for example, let's say it's a 28-day transit time from Japan into the West Coast. And then it gets moved in. Maybe you want to find a faster way to move things in. So we'll give you some techniques of how to make mode choice. But each of the segments that you identify, you want to identify where the pain point is, and then your methodology will be used to mimic that or to help you address that. Does that make sense? Yeah. So another question regarding that magic number we used for using normal distribution and T distribution, which was 30. So when the cycle size function, you know. And the forker is saying that it was used to use 32. So I agree. When I learned it was 32, for some people it was 40 and 30. So it doesn't actually matter if, you know, when the number of samples increase, you can use both. So the T distribution and normal distribution are actually the same. So in practice, you can always use T distribution because the less than 30, it's better to use T, and above 30, they are the same. So you can always use the T distribution. But for the learning purpose, we try to, you know, use T and the distribution just because we need the distribution. Yeah, I know. That's the first thing. Conor, one of the people working on the courses, was very adamant that you only need T. Never use normal. And that's correct. But the error, and you can do this, set this up in a spreadsheet, give it distribution, do a T and a normal, and just do it with a number of observations and you'll see what happens. And right, it starts mimicking the normal curve above 30, 32, 40. I think I learned a 32 as well. And I looked at most books now that it's around 30, because it might be 32.5 or 33, but it really doesn't matter. Once you get above around three dozen, the error goes away. Okay, so we are pushing to the end of the thumbs up. So for SC1X, SC1X starts just today, so you have still time to enroll to SC1X. And as Chris said, SC1X is supply chain fundamentals, you learn forecasting, inventory management, and transportation, which are the... Cornerstones, yeah. And warehousing, we also threw in warehousing. Yes, yes, warehousing is also there. So, yeah. All right, well, hopefully you enjoyed zero, good luck with the exam. Hope you all do well, and hopefully we'll see you in SC1X. And if you have any comments or suggestions for improving the course, please, you can reach us through the surveys or directly. And so, thank you a lot for running such a great course. Appreciate it. Thanks for everyone involved, hope you had fun. And good luck with the final exam in SC0X. All right.