Environmental Engineering Reference
In-Depth Information
50
45
40
35
30
25
20
15
10
0
20 10
30
10
40
9
8
50
7
70 60
6
5
4
80
3
Fig. 10.3 Response of the multivariable index developed with a fuzzy system to the variation of
the variables total number of fish species and percent cover of opportunistic macroalgae
Since it is based on seven different metrics, the index has a complex, multi-
dimensional behaviour. A simplified representation of the index variation in response
to input values variation is shown in a 3D graph (Fig. 10.3 ): two metrics, percent cover
of opportunistic macroalgal species and total number of fish species are plotted on the
horizontal axes, whereas all the other metrics are fixed in their “worst case” condition,
i.e. at the values corresponding to bad ecological status. For this reason, the maximum
index value obtained with this simulation is 50 instead of 100. The index increases in
response to the increase of fish and to the decrease of opportunistic macroalgae. The
general behaviour of the index is non-linear, but there are regions of linearity, due to
the selected shape of membership functions (triangular and trapezoidal).
10.4 Future of Fuzzy Ecological Models
In the scientific literature, many examples of ecological fuzzy models already exist,
but many more are expected to come, considering the wide range of solutions that
fuzzy logic offers to ecologists and environmental scientists. Future progress of the
modelling techniques is desirable, in particular relative to the fuzzification, infer-
ence and defuzzification strategies. Most of the currently published models make
use of the simplest available options (linear functions, default inference operators
and defuzzification methods), regardless of their semantic meaning. Furthermore,
Search WWH ::




Custom Search