Environmental Engineering Reference
In-Depth Information
Table 3
Linear regression results for the best attraction analysis (from Gonzalez-Feliu et al.
(
2012a
))
Model
R
2
Degrees of freedom
F Value
Significance of F
n
k
3C- CUA
0.77
25
4
27.29
7.4 9 10-8
3C-NP
0.75
20
4
20.90
5.4 9 10-6
3C-FP
0.84
19
3
58.01
1.7 9 10-8
the Student's test of each coefficient. The calibration method and the main anal-
yses are presented in Gonzalez-Feliu et al. (
2012a
). Taking into account the
degrees of freedom and the quantity of data used for the analysis, the values are
consistent and in all cases reject the hypothesis of the simultaneous nullity of all
the coefficients (Table
3
).
5.1.2 Catchment area Model
After the trips are generated as a function of their shopping destination, we pro-
pose to connect them to households using a catchment area model (Kubis and
Hartman
2007
). The approach proposed is a probabilistic retail gravity model
which can be formulated as follows:
E
a
1
j
E
a
2
i
P
ij
¼
T
ij
T
j
NrH
i
c
ij
¼
A
j
where: E
i
and E
j
are the number of employees in retailing activities for zones i and
j respectively; NrH the number of households of zone i, and c
ij
the cost of
transportation between i and j. Moreover, we define A
j
as follows, in order to
ensure the flow balance (Ortuzar and Willumsen
2001
):
1
A
j
¼
P
k
E
a
1
j
E
a
2
i
NrH
i
c
ij
a
1
, a
2
and b are parameters determined first by linear regression then readjusted
using Hyman's (
1969
) iterative procedure to minimize the error between surveyed
and estimated mean distances (a comprehensive description of the calibration
method can be found in Gonzalez-Feliu et al.
2010b
) (Table
4
).
Table 4
Estimation results (from Gonzalez-Feliu et al.
2010a
,
b
)
Surveyed
Estimated
Error (%)
Number of trips per day
301629
296230
1.7
Distance (Millions of km/day)
4.34
4.44
2.1
Mean distance c (in km)
14.39
14.99
4.2
Initialization value b
0
-
0.89
-
Best b value
-
0.91
-
Search WWH ::
Custom Search