In our last video we built a simulation by adding uncertainty
to a flight overbooking model.
The model has two sources of uncertainty, the demand and the show-up rate.
Let's review the simulation model that we created.
So locate and open the Excel file, Flight Overbooking Simulation.
This workbook has two tabs.
Click on the Overbooking Simulation tab.
The top table has the problem data and the bottom table has a model.
The model starts with a policy decision of the overbooking limit.
This is the yellow cell, B13.
The two green cells, B14 and B17, are the ones with the certainty.
In B14, we use the PsiPoisson function from the analytic solver platform to
model the actual demand based on the average demand of 150 passengers.
In cell B17, we use PsiBinomial to model the show up process.
The binomial has two parameters, the number of passengers that were booked,
which is in cell B17, and the probability that they show up, which is in cell B6.
The model calculates the number of no shows and
then it determines how many passengers will actually board the plane,
versus those who are bumped to a later flight.
The last part of the model calculates revenue, cost and net revenue.
If we recalculate the spreadsheet by pressing F9,
we can see that the values in the green cells change.
And that causes changes in both the number of bumped passengers and
in the net revenue.
We have highlighted these two cells with light orange to indicate these
are outputs we're interested in tracking.
Every time that we recalculate the spreadsheet, 10,000 values for
the actual demand and the ready to board cells are generated.
Each set of values is what in simulation is called a trial.
As we mentioned in our previous video,
you can change the number of trials in the tools group of the ASP tab.
We need to tell the simulator that we would like to track the number
of bumped passengers and the net revenue.
This is done by selecting cells B20 and B24, and
clicking on Results in the simulation model group of the ASP tab.
Once we do that, we can check the cells to see what has changed.
By clicking on cell B20,
we can see that there is an additional function called PsiOutput.
This function does not modify the cell calculation.
It is there to tell the simulator that the cell is a simulation output and
it that it should collect data on it when the simulation is run.
The same function was added to cell B24.