Environmental Engineering Reference
In-Depth Information
Table 7. Example of artificial utility scale (Tsamboulas & Kopsacheili, 2003)
Type of impact
Score
Strong negative impact
-1
Large negative impact
-0.75
Moderate negative impact
-0.5
Small negative impact
-0.25
No impact
0
Small positive impact
+0.25
Moderate positive impact
+0.5
Large positive impact
+0.75
Strong negative impact
+1
absence of impact and +1 the best scenario. The
utility function of a crooked linear form adapted
from Tsamboulas and Kopsacheili is as follow:
sion in the final decision matrix is done similarly
as with the crooked linear utility function. Both
forms of utility function produce artificial scores
that will be aggregated with qualitative weighting
or no weighting at all.
+
P A
if P
if P
if P
>
=
<
/
0
0
0
ij
ij
U
=
0
ij
ij
Weighting Using Cost Functions
−
P B
/
ij
ij
We propose an evaluation concept of certain
quantitative criteria using cost functions in an
attempt to overcome the objectivity limitations of
utility-based and qualitative methods. In contrast
to utility functions that occur at the scoring stage,
cost functions apply to the development of weights
in the next stage. They stem from the assumption
that all negative impacts incurred by a selected
alternative should be mitigated at a certain cost. It
is of course the case that any form of development
demands resource consumption and results in the
degradation of the natural environment.
The cost function is applied at the weighting
step. Quantitative criteria are based on specific
indicators with dedicated units. Scoring may thus
be done purely quantitatively using procedures
similar to those in EIAs. The cost function con-
cept developed here stems from the assumption
that negative environmental impacts that cannot
be avoided will have to be mitigated at a certain
cost. Airport developers would therefore choose
the alternative with the lowest mitigation cost
where
C
j
= criterion j,
P
ij
= physical and real performance of criterion j
measured as a change and not as an absolute value,
U
ij
= artificial performance of criterion j,
A,B
= constant variables that either depend on
measurement thresholds or are set by relevant
decision makers.
2. Utility Function with Defined
Artificial Impact Scores
The physical performance
P
ij
of purely qualitative
criteria is expressed verbally. There is no need for
a mathematical utility function as the informa-
tion reported by the indicator is not quantitative.
Tsamboulas and Kopsacheili (1999) proposed an
artificial scale of utility scores that correspond
to verbal description as set forth below in Table
7. Note that the scale also converts performance
values to utility values ranging from -1 to +1. Inclu-
