GBA 327: Module 7D AVP Transcript Title: The Monte Carlo Simulation Using Risk Solver Title Slide Narrator: Although the use of a data table illustrates how we can apply Monte Carlo simulation to a decision model, it is impractical to apply to more complex problems. Slide 2 (Screenshot of outsourcing decision model.) This file is available for download in Doc Sharing. Narrator: In most cases, we need to conduct thousands of trials in order to obtain good statistical results and to avoid sampling errors associated with a small number of trials. To perform a large-scale Monte Carlo simulation, we will use the Risk Solver Platform that you downloaded for this course. We will use the original Outsourcing Decision Model File to perform the simulation using Risk Solver, which is
available in Doc Sharing. Before we do the large-scale simulation, I will show you that we can do one trial by simply entering the Risk Solver Psi probability distribution functions in cells B10 and B12. Slide 3 (Formulas shown.) Narrator: In Cell B10, we will assume that production volume is normally distributed with a mean of 1,000 and a standard deviation of 100 as in the previous example. Slide 4
(Added new formula to Unit Cost.) Narrator: We will make the problem a little more complicated by assuming that the unit cost of purchasing from the supplier is also uncertain and has a triangular distribution with a minimum value of $160, a most likely value of $175, and a maximum value of $200. Slide 5
(Formulas hidden and results shown.) Pressing F9 produces a new trial. Narrator: Notice how pressing F9 produces another trial. Slide 6
(Analytic Solver Platform selected with Distributions highlighted.) Narrator: To run a simulation with 5,000 trials, we will use the Distributions button in the Risk Solver Platform ribbon. Slide 7
(Distributions, Common, and Normal ribbon menu items highlighted.) Narrator: Click in cell B12, the production volume. Then click the Analytic Solver tab. Then click the distributions button in the Simulation Model Group, select the Common category, and then click Normal. Slide 8
(Pop up window shown with bell curve graph and parameters highlighted on right menu bar.) Narrator: In the parameters area on the right of the dialog box that appears, enter 1000 (no comma) for the mean and 100 for the standard deviation. Then click save. Slide 9
(Back to spreadsheet and Distributions, Common, and Triangular ribbon menu items highlighted.) Narrator: Next select cell B10, the unit cost. From the Common distribution category, select Triangular. Slide 10
(Pop up window shown with triangular graph and parameters highlighted on right menu bar.) Narrator: Finally, enter 160, 175, and 200 for min, likely, and max in the parameters section on the right of the dialog box and click save. Note that you will see an error message if you do not enter three values. Slide 11
(Back to spreadsheet with Analytic Solver Platform ribbon active and Results, Output, and In Cell ribbon menu items selected. Narrator: Next we will define the cell we wish to use to predict and create a distribution of output values based on changes to the inputs we selected. For this model, we want to predict the cost difference between manufacturing in-house and outsourcing, so we click in cell B19. Then click the results button in the Simulation Model group, choose Output, and select In Cell. Slide 12
(Results shown and formula in formula bar highlighted.) Narrator: Risk Solver changes the formula in cell B19 to =B16-B17+PsiOutput(). You should note that the output cell must be numeric. Thus we couldn t choose B20 as an output cell. Slide 13
(Options ribbon menu item is highlighted.) Narrator: Now we run the simulation. To do this, click the Option button in the Option Group of the Risk Solver ribbon. Slide 14
(Analytic Solver Platform 2014 Options pop up window is shown.) Narrator: In the dialog box, we change the number of trials to 5000. Because Monte Carlo Simulation is essentially statistical sampling, the larger the number of trials, the more precise the result will be. The educational version of Risk Solver only allows for up to 10,000 trials. Next we select Monte Carlo as the simulation method in the dialog box and click OK. Slide 15
(Simulate ribbon menu item is highlighted.) Narrator: Finally, we run the simulation by clicking the Simulate button in the Solve Action group. Slide 16
(Simulation Results for cell $B$19 are shown in bar graph form.) Narrator: When the simulation finishes, a dialog box appears showing the results. Slide 17
(Simulation Results for cell $B$19 are shown in bar graph form and Mean is highlighted.) Narrator: The Frequency tab shows the frequency of cost difference for 5,000 trials of the Monte Carlo simulation. We see that the mean of the cost difference is approximately -$3,000, which suggests that, on average, it would be better to manufacture in house rather than outsource. Slide 18
(Simulation Results for cell $B$19 are shown in bar graph form and Minimum and Maximum are highlighted.) Narrator: We also can see that the minimum difference is approximately -$40,000, and the maximum approximately $23,000, lending further evidence that it might be better to manufacture in house. Slide 19
(Simulation Results for cell $B$19 are shown in bar graph form and Upper Cutoff is highlighted.) Narrator: If we enter zero in the Upper cut-off of the Chart Statistics section of the Statistic pane of the dialog box and then click save Slide 20
(Right half of the bar graph is in light blue color now and 58.9% in the right menu is highlighted. Narrator: we can see that 60% chance of a negative difference, suggesting further that we should manufacture in house. Slide 21
(Spreadsheet is shown with small image of the bar graph with a highlighted button stating Click here to open full chart. ) Narrator: If the dialog box disappears when you enter zero and save, we can open it again by clicking in cell B19 and choosing open full chart. Slide 22 End of presentation.