Database Reference
In-Depth Information
Now the SICK disappear from the stock into a “cloud.” You are not explicitly
modeling where they go. Run the model again and find that the steady state number
of SICK has dropped to 77 after about 90 days.
Now let us find a way to run a more accurate form of the model by converting
our stock of SICK into a conveyor. A conveyor is really a symbol for a long series
of connected stocks, in our case each with a one-day residence time. A conveyor
must have the time spent in it, the Transit Time explicitly defined (here, 25 days).
Use the dynamite symbol in the STELLA menu to remove the connector from SICK
to GETTING WELL and the converter SICK TIME. When you dynamite a model
component, put the fuse of the dynamite precisely over the object to be removed.
For information arrow removal, put the fuse directly over the circle at the end of the
arrow.
Open the stock SICK and click on conveyor, and then set Transit Time equal to
25. The initial value of the conveyor is set at 10, and STELLA will distribute this
value evenly over its 25 time periods. If you want a more specific initial distribution,
you must enter a value for each of the time periods, separated by a comma. For
example, a conveyor with initial values of 2,1,3,.. would have a 2 as its initial value
for the first time interval, a 1 for the second time interval, a 3 for the third, and so on.
Here is how your model should look now (Figure 1.15):
SICK
GETTING WELL
GETTING SICK
~
CONTAGION RATE
AWARENESS LEVEL
Fig. 1.15
Run the model and note the surprising result: damped cycling of the number
of SICK. This is due to the explicit time delays introduced by the conveyor. Such
delays are ever-present in the real world and introduce great complexity into the
results of dynamic models. Here is a long-term picture of the results showing a
long-term steady state of about 79 SICK, slightly higher than our previous, less
accurate model (Figure 1.16):
The stock can also be turned into a queue or an oven. An oven allows entities
to remain in a stock for a predetermined duration before they get released into a
flow, and queues allows them to be queued up in their order of arrival. Each of
these state variable forms is explained in the STELLA help file that comes with the
full-implementation program.
You can of course continue to expand the model indefinitely, and the process is
so easy and rewarding that you will be inclined to do so. But the goal of model is to
