 Let us download this delay info forecast.mdl model and let us see what kind of graphs we get when the actual sales rate is 100 plus type of 10 comma 10, 100 plus 15 to pulse of 10 comma 1 and what if sales is normally distributed around a constant mean. See these are typical scenarios happen. The first two cases are just to understand how the dynamics work, but in reality we expect that the demand keeps fluctuating around a constant mean, so assuming normal, so in that case let us see how to simulate that scenario also. Let me quickly show the forecasted demand is 100, the actual sales is 100 plus type of 10 comma 10. Step of 10 comma 10 means at time 10 there is additional unit of 10. So from 0 to 10, 0 to time 10 the actual sales and the forecasted sales is exactly the same, so we do not expect any dynamics to happen, so system is started dynamic equilibrium. Smoothing constant is 0.2, so at every time period 20 percent of the difference is going to be added to the stock, so stock has to keep increasing, so that is what you expect. So if you simulate the system, the forecasted demand follows a simple exponential goal seeking process because the system that you just saw is nothing but a goal seeking model where instead of having a desired stock we called it as a reported value or the actual sales and as a goal changed I am going to approach the new value of the goal and saturate at 110. Here note that the material is not conserved. There is let me show forecasted and actual sales if you plot them, the actual sales increased to 110 and yield constant there but here you can see the forecasted demand slowly approaching 110 and after some point it will be the same but whatever happened in this transient period is lost like many no material is conserved there. The information was used to adjust my forecasted demand but beyond that I am not using that information again, so that is what we mean by flows are not conserved. Maybe it will be more apparent when we do the second option what was it 100 plus 50 into let me just copy this. So now I am just going to the actual sales rate and replacing with 100 plus 50 into pulse 10 comma 1. So, what has happened is at time 10 there is a pulse of 50 units there is it only one period there is abnormal suddenly demand peaked additional 50 units. So, the actual sales there is a red line went from 100 to 150 for just one period and came back but since we are reacting the next period I added 20 percent of difference so it went to 110 and next period I saw it was still only 100 then I slowly started to revise it downwards until I hit my goal. So, I never hit 150 I only go little way up and then I just slowly come back to the old goal. Again this is just you know simulation right, so it does not know what is going to happen cannot estimate until it actually sees the value. Now let us go to the third one in third case we want to do something called random normal where we do that so whenever new things come I expect you to I am just showing it because you will also learn how to do now there are various functions like the one if you want to simulate random distribution or Poisson triangle various distributions can also be given here. So, in demand typically we model as normal distribution or exponential distribution or various distribution some of it it can handle. So, let us see do I get random normal is what we want random normal though mathematically we just give two parameters for simulation purpose it looks like it wants 1, 2, 3, 4, 5 different parameters by default it gives normal distribution means 0 and variance 1 h is the mean r is the standard deviation in this third and fourth parameter mean a standard deviation what is m? m looks like the number of samples m is the minimum value it will return x is the maximum ok. So, m and x gives the minimum and maximum for a truncated normal distribution what it gives so you cannot give it minus infinity plus infinity if you really want it then you give a really large number positive and really large number negative. So, minimum value maximum value mean standard deviation and s I think stands for seed number so that when you run it multiple times you generate the same random number that is enough for now. So, let us go back minimum is 0 maximum 500 we have kept mean is 100 standard deviation 25 we have been given they just put a stream as default value as 1. So, let us simulate only let us plot the sales we will get a curve like this it is a randomly distributing the mean is approximately 100 as per this which is what we expect right mean is approximately 100 because of standard deviation that is the noise we are going to get fluctuating values of sales it looks more realistic. Now, let us do forecast and this together the green line in the middle is the forecast as you can see it is doing exponential smoothing so it is not going to go to the extremes because it is only 0.2 you get a green line like this it does not react so much to the demand pattern. If you change the alpha value to something larger let us see that let us see what happens when we change the demand value or smoothing constant to something larger instead of 0.2 let us consider 0.8 simulates over right now current 0.8 you can see 3 lines the red line this one where are the mouses the red line is the actual demand there is a I do not choose the color. So, the orange line seems like this is 0.2 and the blue line is with 0.8 0.8 more faithfully follows the same pattern because you are going to weightage of 80 percent so it is going to follow the same pattern if you do not want it to react so much to the deviations then we try to hover around the lines. Now, let us take up another example availability of job openings influence people to migrate to city and as people migrate to city they fill the available job openings. The simplistic SFD is shown here people then job job openings migration adjustment time do you think that this represents the description given the first question the availability of job openings influence people to migrate to city as people migrate in city they fill the available job openings as migration happen people come in then we are set of jobs. As per this model what should be the units for people as per whatever is given as stock flow diagram people unit could be person job units could be job so now what will be units of job openings. So, I am just taking the difference the structure is exactly the same so I have to differentiate difference between as per this it is job is minus people so there it will get a dimension mismatch and job opening even if it is converted into jobs then migration as per this it should be person per time correct so migration should be person per time but job opening somewhere I am just using the jobs so I need to actually have a another variable called as jobs per person or something like that or person per jobs either way. So, whenever there is a model here you see we will do three things we open the model check the model settings check the start time check finish time takes time step very easy then you check the value of constants and equations and then check whether it is dimensionally consistent. So, I am going to show all the three steps what is the name of the file again jobs if you go to model settings we will find the initial time is 0 the final time 50 times it was 0.25 units for a time is 1 fine people about 800 people are there initially units is persons then job with 1000 people are 1000 jobs are there but then I came to job openings I just put the unit as job so jobs minus people then when they went to migration I just did it as job openings divided by adjustment time very simple model and the way it is structured is exactly same as the previous two examples one we drew on the paper or one we just saw it is exactly same format. But to check the units if you do model units check it will say there are two unit errors discovered the first one is error in units for following job openings equation other is error in units of following migration equation because job openings had job minus people that means job unit and person unit has taken difference and then final unit you have said as jobs it says it does not match that is one. So, another thing while you are here I can go to the question so in your you can say now when some model there is some button here called as document all so if you click that button it will give the entire model along with the settings in equations form it shows a variable adjustment time to what is units what is the final time what is the initial time what is the time units then the equations job openings jobs migration people in this how can you figure out which is the stock from this which is able to see which one has the inter function is the stock and whatever is integrating it over it must be the flow right. So, that is how you find it ok now let us so this is a model we have let us see what we have to do to change it and then I will show how to change it. So now we have to make the model dimensional consistent let us introduce a variable called as persons to job ratio unit is persons by job let us introduce a new variable and connect it to person or connected to job openings and connected to migration so that now I can make it dimensionally consistent. So, I am going to introduce a person to jobs ratio of people to jobs let us just people to job ratio let me connected to migration as well as this is a two equation which requires changing so I am just connecting it to them as soon as you click f of x all three is going to go black person to job ratio let us call it person by job let us just say one person equal to one job for simplicity sake. Now, how will I change the job openings equation I need to convert it into units of jobs. So, I will say job minus people divided by people to job ratio and here I can check syntax into the next equation and then I go to the next equation migration. So, migration I just need to multiply jobs to people ratio because job openings are in jobs. So, when you multiply it will come back into people let us do model units check all units are ok. So, I just made it dimensionally consistent again when I get a one example I try to use a couple of other ideas so that both the jobs are done. So, now if I simulated ok what is the initial values the people initial value was 800 and job value was 1000 since people to job ratio is one it is kind of equivalent I just simulated it. So, let me select people and jobs and make a so people jobs jobs remain constant at 1000 people migrate and go from 800 to 1000 this is a very simple model of a basic delay. Now, let us make that example little more interesting any questions on this for all I just made a dimensional consistent, but other than this the structure does not did not change from the previous demand example. One example looks like a very constructed and structured example seen other example other causes also, but this one looks like we are talking about migration other stuff, but finally the basic structure is very similar.