 Dear participants, welcome to the course on supply chain digitization. It is jointly being taught by Professor Priyanka Burma, Professor Sushmita Narayana and Professor Devapratadas from Mumbai. So, in last class, we gave you a brief introduction about digital twin. What are various types of digital twin exist specifically in the context of supply chain and operations management? What are the various enablers of digital twin? How do you get the data? How do you connect physical system with the digital system and so on. So, in today's session, we will try to develop one small digital twin to do the green field analysis. So, before we go ahead, let us quickly understand what is supply chain digital twin and what is the framework which we discussed in the last class. So, it is a virtual system comprised of digital visualization of a physical supply chain and its elements like firms, flows and products. It has digital technologies which provides information about physical supply chain and its elements like ERP system, IoT sensors, cloud, blockchain. It also helps you to take better decision. It can do that by using descriptive analytics, predictive analytics and prescriptive analytics. So, this is the broad framework of supply chain digital twin. In the last class, we had detailed discussion on it. So, in today's session, we will focus on how to develop a digital twin to perform green field analysis. So, for that we have a case study, very small case study. The senior management of a pharmaceutical company is planning to set up a distribution center to cater the demand of western region of India. The estimated annual demand is shown in the table. So, now if you see, we have the demand data. So, it has 4 markets in the western region. Pune, let us say is my market 1. Mumbai is my market 2. Ahmedabad is my third market and Surat is my fourth market. And it has annual demand like this. In Pune, the annual demand is 157,680 units. In Mumbai, it is 179,580 units. In Ahmedabad, the demand is 155,490 units. And in Surat, it is 43,800. So, my objective is to set up the DC. So, the transportation cost is minimized. The question is where should I set up this DC? Should I set up it near to Pune? Should I set up it near to Mumbai? Should I set up it near to Ahmedabad? Should I set up it near to Surat? Or where should I set it up so that my overall transportation cost is minimized? So, then how do you approach this problem? So, for that, first I need to get the exact location of my customers. So, let us say I am giving this information to you. So, the Pune, the latitude and longitude of the customer location is 18.52 is the latitude, 73.85 is the longitude of Pune location. Then we have latitude and longitude of Mumbai location is 19.09 and 72.87. Ahmedabad is 23.02 and 72.57. For Surat, latitude is 21.18 and longitude is 72.83. So, now I have given you some more information. You not only know the demand at each of these 4 markets, you also know the exact latitude and longitude of these 4 markets. Now, the next question is where should I set up the distribution centers so that I can quickly reach to the customers and my overall transportation cost is minimized? So, where should you set up the DC? That is the question. So, how will you approach this problem? So, to do this first let us visualize it so that you will get a better understanding where the customers are located. So, my first customer which is Pune which is located over here, then second customer which is Mumbai is located over here, then Ahmedabad is my third customer which is located over here and Surat is located over here. So, I know their exact location. So, these are my latitude and longitude. So, let us say for Pune it would be 18.52, 73.85 and the demand is 1,57,680, that is my demand for this location. For Mumbai it is 19.09, 72.87, this is my location and demand at this location is 1,79,580 units. Then I have Ahmedabad which is located here and the exact latitude and longitude is 23.02,72.57 and the demand is 155490 and the fourth customer is located here. The latitude is 21.18, longitude is 72.83 and the demand is 43,800. So, now I know on the map where are they located. So, I have to find out where should I locate my DC so that overall transportation cost is minimized and transportation cost depends upon the demand as well as distance because I need to meet this demand that means I need to send the product from the DC to this demand locations and these many units I need to send. So, there are infinite possible options are there. I can place DC in any place closer to this region, but where exactly should I set up so that overall cost is minimized. By looking into data looking into this data I am tempted that if I locate it near Mumbai then huge volume of demand can be satisfied quickly and my transportation cost from DC to Mumbai will be minimized. Let us say if I locate close to Mumbai and transportation cost from DC to Mumbai will be very low, but however transportation cost from DC to Ahmedabad will go up transportation cost from DC to Sirat will go up from Pune it will be little higher. If I put it close to Pune then from DC to Pune cost will be lower, but DC to Mumbai cost will be higher DC to Ahmedabad cost will be further high DC to Sirat cost will be higher. If I place it near to Ahmedabad then DC to Ahmedabad the cost will be lower, but however DC to Mumbai DC to Pune and DC to Sirat cost will go up. So, therefore there are lot of possibilities are there I need to find an optimum solution which will minimize my overall cost. So, for that we need to develop a model so that I can get the exact location of the DC which will minimize the overall transportation cost. So, for that we need to develop an algorithm develop a optimization model. So, let us look into this symbols x i is the x coordinate of market i y i is the y coordinate of market i. So, let us say market 1 is Pune. So, this is x 1 y 1 x 2 y 2 x 3 y 3 x 4 y 4. So, I can put latitude and longitude in x and y coordinate and then I will get the exact location of this markets. Then we are mentioning demand i demand of market i. So, this will be demand 1 demand of Pune will be demand 1 demand of Mumbai will be demand 2. So, whatever demand of Mumbai will be as demand 2 then this will be demand 3 the demand of Ahmedabad will be denoted as demand 3 and demand of Surat will be denoted as demand 4. Then we have x DC and y DC. So, this is my variable this is my decision variable. So, I need to find out where should I locate the DC that means what should be latitude or should be longitude of the distribution center. So, these are my unknown I do not know I have to find out this value. So, let us say my DC is located over here somewhere. So, this will be my x DC y DC. So, if it is located over here then I need to find out the distance to Surat I need to find out the distance to Ahmedabad I need to find out the distance to Mumbai I need to find out the distance to Pune. So, first I need distance formula distance formula to find out the distance from DC to various location. Now, this is not the exact location this is a representative location the exact location will be determined by the optimization method. Suppose, this is the location then this is my distance to Ahmedabad this is my distance to Surat this is my distance to Mumbai this is my distance to Pune. So, now once I get all the distance from DC to each of this demand points I need to multiply this with the corresponding demand. So, if you see this distance x i y i comma x DC y DC is nothing, but distance from distribution center to each of this demand location. So, this value z I need to minimize it. So, I need to minimize the value of z and find out what should be optimum value of x DC what should be optimum value of y DC. So, this is the overall optimization model. So, now we need to solve this model to find out the optimum location of x DC y DC that is optimum location of the distribution center. So, for that one of the important point is distance. So, let us focus on what various distance formula the first one is direct line method suppose I have location A I have to go to location B I can go directly. So, what is the formula suppose this is my x i y i and this is let us say x DC y DC the distance between these two point is x i minus x DC square plus y i minus y DC square square root of this value. Now, unfortunately the distance between two points are not like straight line I cannot travel in the straight line I have to take left turn I have to take right turn and so on. So, there is another method called coronar method. So, if I have to go from point A to point B I will first go in this direction and then I will go up. So, this is a much better representation of real life scenario where you need to take turns to reach from one location to another location. So, what is the formula of distance between x i y i and x DC y DC is the formula mod of x i minus x DC plus mod of y i minus y DC. So, this coronar method formula gives a better representation of real life distance between two points because we need to take turns left right and so on. So, therefore, coronar method is a much better approximation of real life distance, but however what happens in reality the surface the earth surface is not flat. So, if I have to go from point A to B since it is not a flat surface I need to improvise the formula ok. So, if I use this particular formula then I would be able to capture the earth surface which is not flat and the distance between two points A and B. So, in this case since we are planning to replicate the real life scenario we will use this distance formula. So, that I can capture the earth surface properly and the distance will be much more accurate compared to the direct method or corner method distance ok. So, now this is this formula we will apply. So, my objective is to optimize the value of x DC and y DC that means, latitude of DC and longitude of DC ok. So, with this objective in mind we have put the data in an excel file. So, I have four markets Pune, Mumbai, Ahmedabad, Surat and I have the demand of each of this market. So, this is the demand data, this is the market. Now, we have assumed that cost per kilometer per unit is 1 dollar ok, 1 US dollar we have assumed. So, this value can be changed and put the actual value of your cost per kilometer per unit. Then I have latitude and longitude of each of these four markets ok. So, now using the formula which was shown in the previous slide we calculated deviation in latitude, deviation in longitude, value of A, value of C and distance between two points ok. So, one point is this market ok. So, I have four market 1, 2, 3 and 4. I need to find out the distance from DC to this market ok. So, initially to start the problem or to start the optimization model I need to take a random value of latitude and longitude ok. So, initially we are assuming that latitude of the distribution center is 20 and longitude of the distribution center is 72 ok. So, this is the random value, it is not the optimum value we are putting a random value then we will optimize. So, if let us assume 20 is the latitude of DC, 72 is the longitude of DC. If this is the correct information then what is the distance from DC to Pune is nothing, but 254.543 which I am getting from the formula shown in the previous slide. Similarly, I can find the distance from DC to Mumbai using that previous formula is 136.197 from distribution center to Ahmedabad. This is my distance 340 kilometer from distribution center to Surat distance is 157 kilometer. So, I got the distance value. Now, I need to multiply this distance with demand. So, I have to travel 254 kilometer and transport 157 680 units. So, my total travelling cost from DC to Pune will be from DC to Pune cost will be 157 680 units that is what I am transporting. How much kilometer I am travelling 254.543 and what is the cost per kilometer per unit is 1 ok. So, this is the cost of sending 157 680 units from DC to Pune. Similarly, I can also calculate the cost from DC to Surat let us say. What is the demand? Demand is 43800 and how many kilometers I have to travel 157.099 kilometer and what is the cost per kilometer per unit is 1. The same way I can calculate the distance from DC to Ahmedabad. How much is the demand in Ahmedabad? 155 490 and the distance is 340.945 into 1. Same way I can calculate DC to Mumbai. How much is the demand? Demand is 179 580 multiplied by distance. DC to Mumbai distance is 136.197 and the cost is 1 color per kilometer per unit. So, if I sum this up I will get the total cost of transporting all this demand to various markets from the DC and this value will be like this ok. So, you can do this calculation you will get the same value. Now, this is not the optimum location of DC. This is for this random latitude and longitude of the DC this is the cost. Now, I need to find out what should be this optimum location or should be this optimum location. So, that my overall cost is minimized ok. Now, for that I will get the formula once again for you. So, this is the formula which we have already built in the excel file and we have also explained this and got the cost 12489 119. I need to minimize this value. So, we will put the same thing in an excel solver. So, my objective is to minimize B 9. B 9 is this is B, this is 9. So, B 9 is this cell I need to minimize this value by changing variable B 7 to B 8. B 7 is latitude of DC, B 8 is longitude of DC. This is my decision variable I need to find out the optimum value of these two variables. So, therefore, I have put them over here and there is no constant and I have to solve it using a non-linear method. So, the formula which you have seen in the last slide it is a non-linear function. So, therefore, I need to use a non-linear optimizer and if I solve this after optimization I am getting this solution. Can you see my optimum value of latitude of DC is 19.09 and longitude of DC is 72.87 and the total cost has come down. Now, the cost is 97392135. So, in this case I have Pune, Mumbai, Ahmedabad and Surat four demand location I can easily solve it using an excel solver. Imagine a scenario where you will have 4000 market across the country. Sometimes you will have across the globe then how do you decide? How do you do the optimization? Excel will not be able to do it. So, can I do it systematically? Can I do it using some other technique? Yes you can. So, for that we will take the help of supply chain digital twin and in this case specifically I have taken the help of software called Nelogistics. Nelogistics helps us to develop a digital twin in the context of supply chain. So, now we have. So, first thing what we have to do? You can read more about Nelogistics in this website and they have shown very nicely like how to build a digital twin from the square specifically the queen fin analysis with the example they have explained. So, for a reference purpose you can do this. So, first thing we have to put the latitude and longitude of these four markets Pune, Mumbai, Surat, Ahmedabad. So, if you put the latitude and longitude automatically it will be plotted in the Nelogistics software. If I do not know latitude and longitude if I spell this CTs correctly CTs name let us Mumbai, Pune, Surat, Ahmedabad automatically Nelogistics will pull up the latitude and longitude of this location. So, just to avoid confusion you can put exact latitude longitude or if you do not want to put latitude and longitude on your own you can ask Nelogistics to retrieve it on your behalf. So, it has the facility it can retrieve it on your behalf. So, if I have 4000 demand points imagine I cannot do it manually for each of this I just need to put the name of this location Nelogistics will automatically retrieve the latitude and longitude of each of these locations and plotting. Now, after plotting it will apply the optimization algorithm which we have shown in the last 2, 3 slides back it will apply that algorithm and tell you where should I locate it. So, let us see after applying the algorithm it is telling me that my distribution center should be located near to Mumbai this is Mumbai you are not able to read it this is Pune, Surat and Ahmedabad. So, it is telling me that my optimal location of this is 19.09 latitude 72.87 is longitude and location is over here. So, it is not only optimizing it it is also plotting it on the map and connecting the demand point it is saying that the optimum location of DC 19.09 comma 72.87 and it is exactly Mumbai location. So, if you see the Mumbai locations latitude and longitude latitude is 19.09 longitude is 72.87. So, after optimizing it is telling me that the DC should be located in Mumbai we also got the same result the excel result suggest that my latitude of DC 19.09. So, this is Mumbai location and in any logistics also it is telling me that my optimal location of DC is Mumbai. So, GFA stands for green field analysis and DC is 19.09 latitude longitude is 72.87. So, for a small problem I can do it in excel I can visualize it is easy, but for a large scale problem imagine where I have thousands of customers, lakhs of customers if I have to find out where should I have my green field distribution center I can use supply chain digital twin I can use specifically the analogistic software and in a click up the button you will get the actual location that is optimum location of the distribution center. So, this is the advantage of supply chain digital twin you can visualize it in a better way you can get the optimum solution in a better way. So, overall it helps you to take better decision in a systematic way in analytical way at a faster rate that is at a quicker time. So, thank you so much I am sure you have understood how a digital twin is formed from the scarce specifically for green field analysis it was small example. So, in the next session we will do some more complex example we will take more data and form some more complex supply chain digital twin. So, that you can get better understanding. So, thank you so much look forward to seeing you in the next class.