Microsoft Office Tutorials and References
In Depth Information
8.6 Summary
As we have seen in both examples, a relatively few sources of uncertainty are suf-
ﬁcient to produce results that are not easily predicted. Thus, the interactions of the
uncertainties of a model are often unexpected and difﬁcult to forecast. This is pre-
cisely why simulation is such a valuable business tool. It provides a systematic
approach to understanding and revealing the complex interactions of models.
The value of a carefully conceptualized and implemented simulation can be great.
Beyond the obvious beneﬁts of providing insight into the risks associated with the
problem, the process of creating a model can lead to far greater understanding of
the problem. The process of creating a simulation requires a rigorous approach and
commitment to very speciﬁc steps, not least of which is problem deﬁnition. Even
the most clever simulation and sophisticated analysis is worthless if you are solving
the “wrong” problem.
It is important to keep in mind that the goal of simulation is to determine the
risk associated with a problem or decision; that is, what is the range of possible
outcomes under the model conditions? One of the steps of simulation that we did
not perform was sensitivity analysis . By changing the parameters of the problem, we
can note the reaction of the model to changes. For example, suppose you ﬁnd that a
relatively small decrease in the set-up times used in Autohaus can lead to signiﬁcant
improvements in service stock-outs. You would be wise to investigate carefully the
nature of step-ups and how you might be able to reduce them to take advantage of
this leverage in service improvement.
In the next chapter we cover a number of tools that are extremely powerful and
useful. They are available in the Data Ribbon—Solver, Scenarios, and Goal Seek.
Some of these tools can be used in conjunction with simulation. The modeling and
simulation we performed in Chaps. 7 and 8 has been descriptive modeling —it has
provided a method by which we can describe behavior. In Chap. 9 we will intro-
duce prescriptive modeling , where we are interested in prescribing what decisions
should be made to achieve some goal. Both are important and often work together
in decision making.
Key Terms
Model
Simulation
Point Estimates
Monte Carlo Simulation
Rapid Prototyping
Continuous Event Simulation
Discrete Event Simulation
Events
Deterministic Values
Risk Proﬁle
RAND()
Random Sampling
Resolution of Uncertain Events
Uniform Distribution
Census
Replications
Discrete Distributions
Continuous Distributions
Search JabSto ::

Custom Search