Introduction - Environmental Modeling - page 5

and explains its strong competitiveness. Therefore Sect. 1.2 gives a brief reminder
of basic matrix operations.
The topic is accompanied by software containing advanced and final versions of
the program files described in the text. The Mathworks logo

appears where MATLAB files of the accompanying software are referenced.

The terms 'modeling' and 'simulation' are synonymously in the concerned scien-
tific and technical literature. However, the term 'model' appears to be more general,
encompassing all types of attempts to capture one or more aspects of a real system,
and is therefore preferred in this topic. The term 'simulation' also fits to the
presented approach, as it suggests that processes which are relevant for the behavior
of a system are included in the computer simulation. In the sequel the term is used
for time-dependent dynamics.
The topic contains relatively simple models throughout. It is not the case that
complex models constructed by MATLAB ® don't exist, but they are not appropri-
ate for an introduction into modeling techniques. For such an aim models should be
as simple as possible, even more, when novice modelers are addressed.
Usually the extensive work with a model leads to renewed extensions, which
turn simple models into complex ones almost as a rule. Not all models are improved
by doing this. Jørgensen ( 1994 ) envisages the connection between model complex-
ity and knowledge, gained by the model, as shown in Fig. 1.1 . Simple models can be
improved by extensions, but there is a certain peak position after which further
extensions do not add to the knowledge - rather quite the contrary. An improved
model design increases the quality of the model (lets take gained knowledge as
a quality measure), but further extensions of the improved model may finally lead to
a situation in which the increase of model complexity is counter-productive.
The model evaluation study of Constanza and Sklar ( 1985 ) provides a plot
similar to Fig. 1.1 , but with 'articulation' on the x -axis and 'effectiveness' on the

®

reference
improved

knowledge

Fig. 1.1 Model evaluation:
knowledge gained vs.
complexity

complexity

0

0