Geoscience Reference
In-Depth Information
Fig. 18.2
The set of points
within 20 miles of three
disaster recovery centers
(DRCs)
50
40
30
20
10
0
-20
-10
0
10
20
30
40
50
-10
X COORDINATE (MILES)
Customers
Centers
explain to the sponsor, in part due to the figures. (6) The well-dispersed locations of
the covering model also had political and geographic redundancy advantages.
The three-location solution to the actual location model for
r
D
20, which covered
97.4 % of the parcels, was accepted by the sponsor. The following is a quote from a
letter the sponsor provided to the team.
The Florida Division of Emergency Management has requested that all county emergency
management offices provide at least three sites preidentified as potential DRCs. With
completion of this project, Alachua County is now able to comply with this request :::
Overall, this was an outstanding project which has provided the Office of Emergency
Management with tangible results. When DRCs must be opened in the future, it will be
based upon careful research and problem solving rather than guesses on which locations
would be best.
In closing, we remark that this approach easily generalizes to covering problems
using network distances, given adequate network data. The approach worked well,
and controls the covering error. We recommend its use for aggregating covering
location problems, as well as unweighted
p
-center problems.
18.4
Aggregation Error Measures
While there can be other types of error in location models, the one we focus on
is
demand point aggregation error
, which result from replacing DPs by ADPs.
Thus, instead of actual distances we obtain approximating ones. The use of these
approximating ADPs creates error. It is thus important for the location modeler
who does the aggregation to be aware of the aggregation error being created. The
modeler who does DP aggregation intentionally introduces error into the model.
TheuseofADPsisthe
cause
of the aggregation error, but there are error
effects
—
including inaccurate values of the objective function and of server locations, due
