The models we have looked at so far, use spreadsheet columns to indicate the passing of time, from January to February, from one year to the next. This way of representing passage of time and the change in variables over time, is referred to as changes in discrete time. It's possible to design a spreadsheet model to show changes in continuous time. To describe the difference between discrete and continuous time, my colleagues here at Worden uses the analogy of a digital clock versus an old fashioned analog clock. The digital clock moves forward in jumps. The analog clock, theoretically, could have an infinite number of moments between one and two o'clock. To implement variables that occur in continuous time, we will need to change the structure of our models, and the formulas we use as objective functions. Let's look at an example in a spreadsheet. As a reminder, models like this cash flow spreadsheet use columns for the passing of months or years. So, for example, column C here has data from January. Column D has data from February. Even if the columns in this model represented days or minutes, this approach is still considered treating time as discreet time. Continuous time will require a different layout for the model. I've copied here a tally of customers for a new start up company. The company has been in business for one year. Each month on the sheet is a cumulative count of all customers since the start in January and to the end of that month. The numbers are interesting, this doesn't look like linear growth, it looks like exponential growth. We'll talk more about that in another lecture but for now let's focus on the passage of time. How do we set things up, so that the passage of time is continuous? In this case, let's say we're trying to forecast the number of customers we might have next year. And not just at the end of each month, but at any point in the year. I did a quick scatter plot of last year's numbers. I highlighted the numbers from January through December. I chose to insert a chart of the type Scatter. And you see the result. Again, the increase in numbers looks pretty good. It's better than straight line growth. It has a curve to it. I added a formula in row seven to show the growth rate from period to period that occurred last year. I also added some forecast variables that we'll use for next year to calculate possible exponential growth. Those include the total number of customers at the base period from last year at 270. That's from December. Next, our forecast of the rate of increase per period. For now, let's say .3. And now, in cell A14, I'll put the number of periods to calculate. I'm setting up these numbers as variables that will be easy in the future for me to change. Now I've added in cell B18 an equation that calculates exponential growth. CF is the customers at a future period that I'm trying to calculate. CB Is an input in my formula that refers to customers in my base period, that is December of last year. The E in the formula is a mathematical constant associated with exponential growth calculations. It's roughly 2.718. R is the growth rate per period And T is the number of periods after the base that I'm looking for. So the formula in B18 includes cell references for total customers at the base period, the rate of increase per period that I'm forecasting, and the number of periods. It also uses the mathematical constant E. In Excel, that's the EXP function. Notice the I can now change the time period by any number including fractions of a period. For example, in this case, here's a forecast for the number of customer's I expect to have in a month and a half. I can change that to 7.25 or any other fractional number, and see the results in B18. That's the difference between a model that handles time as continuous and a model like the cash flow forecast we looked at earlier, that handles time as discreet time.