Information Technology Reference
In-Depth Information
Table 1.
Comparison of different models from the literature
truck- incomplete
load
requests depots terminals TW
Jula et al. 2005 [5]
FT
no
1
m
no
Imai et al. 2007 [4]
FT
no
1
1
no
Zhang et al. 2010 [14]
FT
yes
m
m
yes
Braekers et al. 2011 [1]
FT
yes
1
m
yes
Zhang et al. 2011 [13]
FT
yes
1
1
yes
Vidovic et al. 2011 [11]
LTL
no
1
1
no
Reinhardt et al. 2012 [7]
FT
no
1
m
yes
Braekers et al. 2013 [2]
FT
yes
1
m
yes
Wang and Yun 2013 [12]
FT
yes
1
m
yes
Nossack and Pesch 2013 [6]
FT
yes
m
m
yes
CPDP, this paper
LTL
yes
1
1
no
FT = full truckload, LTL = less-than-truckload, m = multiple, TW = time windows
permanently assigned to a vehicle. That is, a vehicle which fulfills an import
request first and then an export request uses for both requests the same con-
tainer. This is not required in our formulation. Furthermore, an inhomogeneous
fleet regarding travel cost and maximum travel time, but not with respect to
the supported container sizes, is considered. In [3] the vehicle routing with full
containers (VRPFC) is extended by time windows and solved as a VRPTW. The
paper [7] extends the model proposed in [3] by multiple container terminals. The
approach is tested by means of real world data.
[13] model the problem with flexible tasks. A flexible task is a request with an
a priori unknown delivery or pickup location which is equal to our approach. Fur-
thermore, they consider a dynamic planning situation which allows that requests
may be added during the day.
In contrast to the previously mentioned approaches, empty containers are
explicitly considered as transportable resources in Zhang et al. [14]. The model
is denoted as Inland Container Transportation Problem (ICT) and takes into
account multiple terminals and multiple depots. The ICT is also studied in [9]
where a solution approach based on tabu search is proposed. In [10] a variant
of the ICT model is used to evaluate the potential of collaborations between
the several carriers in the hinterland. [12] extend the ICT by considering train
transportation in addition to truck transportation. The model of [6] also builds
upon the model of Zhang et al. [14]. Each truck serves exactly one container (i.e.
only 40-foot containers are regarded). Consolidation of two 20-foot containers
on a truck does not take place.
In a recent paper [2] also considers the situation where pickup or delivery loca-
tions of empty containers are undefined. An empty container allocation problem
is proposed which is based on a transportation problem before solving the rout-
ing problem. However, only homogenous container sizes are considered. The goal
is to minimize the travel distance of empty containers.
