Information Technology Reference
In-Depth Information
As new container transport requests arrive over the day the decision maker
should apply online planning tools which provide a frequent and reactive plan
revision of the needed hinterland container movements. In order to improve plan-
ning, the purpose of this paper is to present a decision model for a snapshot of this
online planning problem. The model assumes a given set of transport requests
together with the current positions of the vehicles and containers. The proposed
model is an extension of the well-known less-than-truckload pickup-and-delivery
problem (PDP). Unlike previous models for drayage transport operations, the
model at hand is able to handle 20-foot and 40-foot containers simultaneously.
The model therefore contributes to reduce the gap between real-world drayage
planning and mathematical planning models used as core components of decision
support systems.
The remaining paper is organized as follows. Section 2 describes the container
pickup-and-delivery problem (CPDP) and Section 3 reviews related literature. A
model for the CPDP with less-than-truckload requests is presented in Section 4
and is validated by means of a worked numerical example in Section 5. Section 6
concludes the paper.
2 The Container Pickup-and-delivery Problem
The presented problem is denoted as container pickup-and-delivery problem
(CPDP). The CPDP has to be solved by a trucking company that organizes the
transport of containers in the hinterland of a port or a rail yard. Fig. 1 shows an
exemplary layout of such a hinterland network. There is at least one (e.g. sea-
side) container terminal, the set of terminal nodes is denoted as
N
TERMINAL
.
Through a terminal, containers enter or leave the hinterland. Furthermore, there
is a set
N
CUSTOMER
of customer locations. A customer may receive a loaded
container from a terminal (
import request
) or ship a loaded container to a termi-
nal (
export request
). At the customer locations there might be an imbalance of
inbound and outbound containers. Therefore, some customers have a surplus of
empty
containers while others are in need of empty containers. A customer with
a need for an empty container for his or her goods initiates a
provide request
.
Hence, the delivery location of a provide request is known but the empty con-
tainer may be picked up from any location with a surplus of empty containers,
e.g., from a container depot or from another customer. On the other hand, a
customer with a surplus of empty containers initiates a
storage request
which
requires a pick up of an empty container at the customer location. However, the
empty container may be delivered to a container depot or to another customer
in need of an empty container. Storage and provide requests are also denoted as
incomplete requests
, because either the pickup or the delivery location is initially
unknown.
We take the perspective of a trucking company that is responsible for perform-
ing all required road-based container movements in the hinterland system neces-
sary to fulfill the customer-specified requests. The considered trucking company
maintains a fleet
V
of homogenous vehicles that are used to fulfill the container
