Geoscience Reference
In-Depth Information
equal revenue potential (to foster competition) which are compact over a geographic
region (to be easier to manage and more economical to maintain).
23.2.4
Distribution Districting
Another important field of applications is the design of pickup and delivery districts
in logistics. Typically, such problems are modeled and solved as vehicle routing
problems. However, if there exists considerable uncertainty in the demand of
customers, several authors propose a two-phase approach that first builds the pickup
and delivery districts and then does the routing on a day-to-day basis. This conforms
with the well-known “cluster first-route second” paradigm for vehicle routing
problems. Hence, basic units correspond to potential customers, given as points,
and the task is to partition the set of customers into districts, one for each driver,
such that the districts satisfy certain planning criteria. A first advantage of these
fixed customer assignments is that the driver becomes familiar with his district.
This, in turn, increases the driver's performance since he becomes quicker at finding
customer addresses, localizing offices within buildings as well as organizing his
routes (Zhong et al.
2007
). A second advantage is that customers become familiar
with their drivers, which increases customer satisfaction (Jarrah and Bard
2012
).
These advantages however have to be carefully weighed against flexible customer
assignments on a daily basis which enable the planner to maximize the driver
utilization and minimize the routing costs (Zhong et al.
2007
).
Concerning the criteria for the districting process, districts should be contiguous
and compact, and the workload should either be balanced or at least not exceed a
given upper bound, e.g., the driver working time. The workload includes the service
time at the customers and typically also an estimate of the average travel time within
the district and to a centralized depot (Galvão et al.
2006
; Haugland et al.
2007
;
Zhong et al.
2007
; Jarrah and Bard
2012
; Lei et al.
2012
).
A final application concerns the establishment of a distribution center which
involves a considerable level of risk due to its enormous start-up investment and
volatile customer demand patterns. One way of reducing this risk is to avoid both
overcrowding and, especially, underutilization of centers by balancing the allocation
of customers to them (Zhou et al.
2002
).
23.3
Notations
This section introduces notations for the main components of districting problems.
