Microsoft Office Tutorials and References
In Depth Information
So let us take what we have learned thus far and organize the steps that we need
to follow in order to perform effective and efﬁcient modeling:
1. A pre-modeling or design phase that contributes to our preliminary understand-
ing of the problem. This could, and often is, called the problem deﬁnition phase.
This step can take a considerable proportion of the entire modeling effort. After
all, if you deﬁne the problem poorly, no amount of clever analysis will be help-
ful. At this stage, the goal of the modeling effort should be made clear. What are
we expecting from the model? What questions will it help answer? How will it
be used and by whom?
2. A modeling phase where we build and implement a model that emerges from the
pre-modeling phase. Here we reﬁne our speciﬁcation of the problem sufﬁciently
to explore the model’s behavior. At this point the model will have to be populated
with very speciﬁc detail.
3. An analysis phase where we test the behavior of the model developed in steps (1)
and (2), and we analyze the results. In this phase, we collect data that the model
produces under controlled experimental conditions and analyze the results.
4. A ﬁnal acceptance phase where we reconsider the model speciﬁcation, if the
result of the analysis phase suggests the need to do so. At this point, we can
return to the earlier phases until the decision maker achieves desired results. It
is, of course, also possible to conclude that the desired results are not achievable.
7.4.1 Pre-Modeling or Design Phase
In the pre-modeling or design phase it is likely that we have not settled on a precise
deﬁnition of our problem, just as Fr. Eﬁa has not decided on the detailed design of
his event. I refer to this step as the pre-modeling phase, since the modeling is gener-
ally done on paper and does not involve the use of a computer based model. Fr. Eﬁa
will use this phase to make decisions about the activities that he will incorporate into
Vegas Night at OLPS ; thus, as we stated earlier, he is still deﬁning the event’s design.
Voitech used the interview exercise in Table 7.1 to begin this phase. The resulting
actions of Table 7.1 then led to the preliminary process ﬂow design in Exhibit 7.4.
If the problem is already well deﬁned, this phase may not be necessary. But more
often than not, the problem deﬁnition is not easy to determine without considerable
work. It is not unusual for this step to be the longest phase of the process. And why
not! There is nothing worse than realizing that you have developed an elegant model
that solves the wrong problem.
7.4.2 Modeling Phase
Now it is time to begin the second phase— modeling . At this point Fr. Eﬁa has
decided on the basic structure of the events—the games to be played and their odds,
the restrictions, if any, on the number of games played by attendees, whether an