Page 22 - DCAP601_SIMULATION_AND_MODELING

Simulation and Modelling

Notes Nature of a Model

A model is a specification for behaviour generation and the modelling process is concerned
with the development of this specification. It is often suggested that the task is to ensure that the
behaviour of the model is as indistinguishable as possible from the behaviour of the SUI. This
assertion is only partially correct. A more appropriate statement of the task at hand is to develop
the specification so that it captures the behaviour properties.
The challenge is to capture all relevant detail and to avoid superfluous features. (One might
recall here the quotation from Albert Einstein, ‘Everything should be made as simple as possible,
but not simpler.’) For example, consider a project concerned with evaluating strategies for
improving the operating efficiency of a fast-food restaurant. Within this context it would likely
be meaningless (and indeed, nonsensical) to incorporate into the model information about the
sequence in which a server prepares the hot and cold drinks when both are included in a
customer’s order.
The notion of ‘behaviour’ is clearly one that is fundamental to these discussions and in particular,
we have suggested that there is usually a need to evaluate behaviour. But what does this mean
and how is it done? At this point we have to defer addressing these important questions until a
more detailed exploration of the features of models has been completed.
Modelling is a constructive activity and this raises the natural question of whether the product
(i.e., the model) is ‘good enough.’ This question can be answered only if there is an identified
context and as we show in the discussions to follow, there are many facets to this key issue. One
other words, a key question is always whether the model is good enough from the point of view
of the project goals. The corollary of this assertion is that it is not meaningful to undertake any
modelling study without a clear understanding of the purpose for which the model will be used.
Perhaps the most fundamental implication of the above discussion is that it is never meaningful
to undertake a study whose goal is simply ‘to develop a model of . . . .’
There is a variety of ways in which the specification of behaviour can be formulated. Included
here are: natural language, mathematical formalisms, rule-based formalisms, symbolic/graphical
descriptions, and combinations of these. It is typical for several distinct formulations of the
model (or perhaps only portions of it) to evolve over the course of the study. These alternatives
are generally created in formats that are best suited to capturing subtleties or providing
clarification.
A particular format that plays a very special role is a specification formulated as a computer
program. The importance of such a specification arises because that computer program provides
the means for actually carrying out the experiments that are central to the modelling and
simulation approach. This illustrates, furthermore, the important fact that some realisations of
the specification (which, after all, is the model) are actually executable and produce the behaviour
we seek to observe. This legitimises the implications in our frequent use of the phrase ‘the
model’s behaviour.

Example (Full-Service Gas Station)
To illustrate some facets of the discussion above, we consider a modelling and simulation
project whose system context (SUI) is a ‘full-service’ gas station with two islands and four service
lanes see Figure 1.4. A significant portion of the customers at this station drive small trucks and
vans which typically have gas tank capacities that are larger than those of most passenger
vehicles. Often the drivers of passenger cars find themselves queued behind these large-tank
vehicles which introduce substantially longer wait times when they arrive at the gas pumps.
This can cause aggravation and complaints. The station management is considering restricting
these large-tank vehicles to two designated lanes. The goal of the modelling and simulation

16 LOVELY PROFESSIONAL UNIVERSITY