A model for optimal feeder vessel management in

container terminals
Abderrahmen Belfkih, Ibrahima Diarrassouba, Adnan Yassine, Cédric Joncour
Normandie Univ, LMAH, ULH, FR CNRS 3335
25 rue Philippe Lebon 76600 Le Havre, France
Email: FirstName.LastName@univ-lehavre.fr

Abstract—The maritime port logistical performances remains Results have shown that more than 50% of 10.000+ TEUs
a challenge for the socio-economic development. It is related to vessels arrivals were delayed more than 12 hours and 24%
different activities in the port. The feeder vessel management were delayed over 24 hours. Figure 1 and Figure 2 shown
problem is one of the different problems in maritime port. In
this work, we present the feeder vessel management problem that vessels with a capacity over than 10.000 TEUs have more
and the related works in the literature. In particular, we study delays at all of the ports except Bremerhaven, which can affect
the case of Le Havre sea port. We present the steps to resolve ports performances.
the feeder vessel management problem, based on an analysis of
the activities in this port. The steps include the proposition of a
model for improving the feeder vessel management. The objective
of this model is to define an optimal berth allocation planning of
feeder vessels, which minimizes the waiting time of ships (both
feeder and mother vessels) in Le Havre sea port.
Index Terms—Feedering management, Berth allocation prob-
lem, Mathematical model

I. I NTRODUCTION
In the last decade, container ports have shown a remarkable
evolution in all sides transport, infrastructure and logistic.
Container ships can be classified into seven types: small feeder,
Fig. 1: Percentage of vessel delays (> 1 day) at four European
feeder, feeder-max, Panamax, Post-Panamax, New Panamax
ports from 22 carriers from April 15 to May 15 [2]
and ultra-large Panamax. The container ships capacity have
increased from 8.000 TEU in 2000 to 21.000 TEU in 2017
[1]. Today, we find an ultra-large container vessel such as:
OOCL Hong Kong with 21,413 TEUs, Madrid Maersk with
20,568 TEUs and MOL Triumph with 20,170 TEUs. The
new megaships can reduce the maritime transportation costs.
However, it can generate a port congestion given their sizes.
They can also spend a long waiting time before loading or
unloading the goods. For this reason, container ports require
changes to improve terminals capacities and infrastructure.
The number and the capacity of feeder vessels will be
increased consecutively in order to serve mega-ships. This
evolution leads today that the mother vessels become feeder
vessels. Fig. 2: Average delays at four European ports from 22 carriers
The increase in container ship size will also have an impact from April 15 to May 15 [2]
on feeder management and consequently, on the port infras-
tructure and services. This means that ports require significant The feeder vessel management in container terminals con-
efforts to deal these challenges. For example, a delay in the sists in the selection of preferred berthing positions in order
arrival of a ship can lead considered reshuffles in the port. to improve both feeder and mother vessels service time. The
CargoSmart Company [2] has focused to study the impact traffic of feeder vessel increases from one year to another. This
of the vessels arrival delays, having a capacity greater than flow will have a significant impact on logistics performance
10,000 TEUs, and which concern 22 ocean carriers at several in maritime port. In this paper, we analyze the feeder vessels
ports in Europe. It has analyzed 2017 vessel arrivals at four traffic in Le Havre sea port. Then, we propose a berth
ports (Antwerp, Bremerhaven, Hamburg, and Rotterdam), in allocation planning of feeder vessels, which aims at optimizing
which 129 vessels have a capacity over than 10.000 TEUs. the service time for both feeder and mother vessels.

Authorized licensed use limited to: Anelis Plus consortium. Downloaded on December 13,2020 at 13:35:54 UTC from IEEE Xplore. Restrictions apply.
 This paper is organized as follows. In Section 2, we review based on genetic algorithm in order to solve the problems.
the related work. In Section 3, we present our contribution The numerical results have shown that the designed algorithm
and steps to achieve this goal. In Section 4, we give a vessels can obtain near optimal solutions within a reasonable time.
analyze of the port of Le Havre. Then, in section 5, we In addition, the proposed model has exposed best results
presents the problem and we gives the mathematical model. compared to the results of a model considering only a BAP.
Finally, in Section 6, we conclude the paper and we give some The authors confirmed, with the results obtained, that the direct
perspectives to this work. transshipment model can significantly reduce total operating
costs.
II. R ELATED W ORK Tsai et al. [6] have presented a new wharf-based genetic
In the literature, the feeder vessel management problem has algorithm for berth allocation planning. Their objective is to
attracted researcher’s attention. It has been discussed according reduce reliance on communication and to minimize the waiting
to several problematic (berth allocation with feeder and deep- time of vessels. Authors have presented the chromosomes
sea vessels, direct transshipment between feeder and mother by matching the vessels and wharves attributes. They aim to
vessels, feeder vessels and deep-sea vessels waiting times minimize the computation time associated with the correction
optimization, etc.). of chromosomes. Their genetic algorithm is able to adjust the
Lee et al. [3] have studied the feeder vessel management wharves use in order to increase the speed of convergence. It
in transshipment terminal with three tactical decision prob- facilitates the population initialization and mutation processes.
lems: berth template, schedule template and yard template Experimental results have shown the algorithm performance in
design. They have proposed three solutions to resolve this convergence speed. It is able to assign vessels to appropriate
problem: (1) a new feeder schedule template design, (2) a new berths as they arrive. It has demonstrate that it is faster to
feeder berth template design and (3) a new planning storage found optimal solutions compared to the three other existing
yard template for transshipment flows. The feeder schedule algorithms [7], cheong algorithm [8], Gkolias algorithm [9]).
template is designed to reducing the workload congestion of Monios [10] has focused to study the role of small container
quayside. It aims to balancing the temporal distribution of ports and their capacity of utilization with feeder shipping.
quayside workload. This solution can help the port operators to It has analyzed and discussed the accessibility constraints of
negotiate the vessel arrival time with shipping companies. The small container ports to serve vessels sub-1000 TEU, based
berth template is designed to fix the favorite berth positions on port calls. For this purpose, it has proposed a methodology
of feeders related to the yard storage locations. It helps based on three steps: the first step consists to analyze the port
also to control the arrival and departure berth positions of calls of small container vessels (sub-1000 TEU), to identify
the transshipment flows. Finally, they have proposed a yard their sea-ports. The second step consists to identify the vessels
template to direct the yard storage flows. sub-1000 TEU dimensions. Finally, the third step consists to
Authors have presented a mixed integer quadratic model, analyze the marine traffic flow and order book in relation
which has linearized thereafter. They have developed a with vessel sizes, to determine small vessels behaviors. Data
memetic heuristic including genetic search and tabu search analyses have shown that 436 ports of 707 container ports in
algorithms. The numerical results have shown that the memetic the world in November 2014 are served by sub-1000 TEU
heuristic gives the best solutions compared to best solutions vessels. It have also shown that the length of vessel increases
found by CPLEX to solve the feeder vessel management significantly with the capacity, from an average of 115 m at
problem. 500 TEU to 149 m at 1000 TEU. Author has found that 90
Emde et al. [4] have focused to solve the berth allocation container ports (21% of ports which have been studied) do not
problem (BAP) between feeders and deep-sea vessels berthing meet the berth depth criteria for large feeder vessels fixed by
at the same terminal. Their objective is to minimize the deep- least 8.7m, which can affect the viability of these ports.
sea vessels waiting times and the number of delayed containers In [11], Hsu has proposed a hybrid particle swarm optimiza-
from the feeders. Authors have adapted the Tabu Search (TS) tion (HPSO) approach, based on an improved (PSO) with an
procedure with the BAP problem. Then, they have suggested a event-based heuristic. He aims to resolve the dynamic and
new Simulated Annealing (SA) based on local search scheme discrete BAP (DDBAP) and the dynamic quay crane assign-
to solve the BAP with Feeder (BAPWF). The numerical results ment problem (DQCAP). Author have compared HPSO results
have shown that the simulated annealing procedure is able to to two Genetic Algorithms (GA) based approach (GA1) and
find feasible and near to optimal solutions in a reasonable hybrid GA (HGA). The GA1 uses the Two-Point Crossover
computational time, compared to the solutions found by TS (TPX) operation and Swap Mutation (SM) operation and only
procedure and CPLEX. supports time-invariant QC assignment. The HGA combines
Zeng et al. [5] have developed an optimizing model which GA1 with the same event-based heuristic used in the HPSO
integrates a Direct Transshipment Berth Allocation Problem so that it also supports variable in time QC assignment. The
(DTBAP) and a Storage Space Assignment Problem (SSAP). results have shown that the HPSO out performs the HGA and
Authors aim to minimize the mother and feeder vessels delay GA1 in terms of fitness value.
cost and reduce the trucks and yard cranes operational cost. Pani et al. [12] have illustrated a data mining approach
They have designed a nearest searching heuristic algorithm for predicting the daily level of mother and feeder vessels

Authorized licensed use limited to: Anelis Plus consortium. Downloaded on December 13,2020 at 13:35:54 UTC from IEEE Xplore. Restrictions apply.
which have exceeded their Estimated Time of Arrival (ETA). 3) the optimization of both feeder and mother vessels
Authors have defined three alarm daily levels which rank the For each policy and optimization objective, we propose
arrivals delay. They have used three Machine Learning models a mathematical model and discuss the related algorithmic
(Decision Trees, Naive Bayes and Random Forests) in order issues. Then, we solve each problem for a set of realistic
to estimate the alarm level. Then, they have compared their instances. Finally, based on our numerical results, we discuss
predictive powerful on a set of collected data. The results the relevancy of each policy from an operational point of view
of tests have shown that the Random Forest algorithm gives (cf. Figure 3).
the best results relative absolute error of 29% compared to
the Decision Trees and the Naive Bayes algorithms. Authors IV. P ORT OF L E H AVRE VESSELS ANALYSE
have emphasized on refining the model with new variables like In this section, we analyze the characteristics of terminals
weather conditions. For this purpose, they recommend to use and vessels in the port of Le Havre based on the AIS data
the past status of port activities to predict the present status in 2016. The port of Le Havre represents one of the most
using time series models. important port for trade and passenger services in France and
In the table I, we give a comparison between different in Europe. It was classified in 2016 as the first French port
literatures on feeder management problem. for container trafic. Located in the Normandy, it ensures more
than 6000 port calls each year. It also supplies the ports of
III. O UR CONTRIBUTION
the Seine Valley (Rouen, Caen and Paris). It includes many
Our work consist of developing the berthing of feeders services: containers management, terminal managements, ship
vessels in the port of Le Havre. We analyze and describe the and river convoy managements, etc.
port of Le Havre vessels operations. The analysis is done using The table II shows the length and depth of different berths
an AIS (Automatic identification Signal) database. The AIS and the number of vessels by container terminal.
represents an advanced technology in maritime navigation. We note in the table II that the TNMSC is the largest
They allow each ship to send different data related to their terminal in the port with 1500 meters of length and 23.92%
navigation and operation. They help for instance the port of container vessels. The France terminal receives the largest
authorities to manage the maritime traffic and to avoid and number of container vessels (32,71%). Then we found the
prevent accidents. They can also be used to analyze the Atlantic and the Ocean terminals. The ”Port 2000” includes the
workload and even congestions inside a container terminal. TNMC terminal, the France terminal and Ocean terminal. It
AIS data analyze can help to understand the port operations. manages the highest number of ships by year. However, other
It can determine information related to the container ships terminals are underserved. For this purpose, it’s important to
classification, the vessel waiting times, the vessel handling manage feeder vessels and affected them to free wharfs in
times, the number of ships assigned to quays, which allows us order to facilitate mother vessels operations.
studding the workload congestion of quayside and fixing the
favorite berth positions of feeder ships. A. Vessels categories distribution according to the quays
Based on the above analysis, we propose different vessel In this section, we study the vessels categories distribution
management policies which take into account the operational according to the berth terminals. We have classified vessels to
constraints of Le Havre sea port and the increase of the number 5 categories according to theirs TEUs (cf. Table III).
of feeders. In particular, we discuss the cases where the priority
is given to either:


"#
$!%$   &

 $   &
' &
Sending AIS AIS Message Decode
 () %'
messages  %'


!!!


AIS Database
AIS Hub


Feeder vessel


data model
Predictive data model
for feeder vessel berth

allocation planning
Mother vessel
Container ship

data model
Database


A model for optimal
      
feeder vessel  
 !
management
Fig. 4: Vessels categories distribution in 2016
Fig. 3: Our contribution
We note in the Figure 4 that the ”Port 2000” receives most
1) the optimization of the waiting of mother vessels or of Ultra Large Container Ship (ULCS) 22% of total vessels in
2) the optimization of the feeder vessels regardless the which TNMSC terminal 52%, Ocean tarminal 43% and France
mother vessels or tarminal 37%. In the second place, we find the Post-Panamax

Authorized licensed use limited to: Anelis Plus consortium. Downloaded on December 13,2020 at 13:35:54 UTC from IEEE Xplore. Restrictions apply.
 TABLE I: Comparison between different literatures on feeder management problem

Problematics Objective Approach Papers

Berth allocation planning Reducing reliance on communica- Genetic algorithm Tsai et al. [6], Frojan et al.
tion and minimizing the vessels [13], Yan et al. [14]
waiting time
Berth allocation problem with Minimize the deep-sea vessels Tabu Search and Simu- Emde et al. [4], Bierwirth
feeder and deep-sea vessels waiting times and the number of lated Annealing based local and Meisel [15]
delayed containers from the feeders search
Direct transshipment between Minimize vessels delay, trucks and Nearest searching heuristic Zeng et al. [5], Liang et
feeder and mother vessels yard cranes operational costs based on genetic algorithm al.[16], Mathur [17]
Feeder management Reducing the workload congestion Genetic algorithm and tabu Lee et al. [3][18], Cordeau
of quayside, fixing the favorite search. et al.[19], Monaco et al.
berth positions of feeders related to [20]
the yard storage locations, control
the yard storage flows
Continuous Berth Allocation Resolve the dynamic and discrete Particle swarm optimiza- Hsu [11], Imai et al. [21],
Problem BAP and the dynamic quay crane tion with an event-based Zhou and Kang [22]
assignment problem heuristic, genetic algorithm
based heuristic.

TABLE II: Length and depth of different berths of terminals to berth because its waiting time is expensive. For this purpose,
Terminal Berths Length/m Depth/m Nb.vessels their berth positions are known in advance. The berthing
(2016) positions of feeder vessels depend on the berthing positions of
France 4 1400 15 32.71% mother vessels. The time duration of the ships operations is
TNMSC 4 1500 15 23.92% known in advance. A feeder vessel may berth on the same quay
Atlantic 3 1308 13.50 17.75%
with a mother vessel to load or unload containers. We propose
a policy which consists to move the feeders on other quay to
Ocean 2 700 15 16.07%
leave the berthing place for another mother ship waiting, and
Others - - - 25.45%
also to move containers from one quay to another using shuttle
vessels.
TABLE III: Vessels categories The assignment model for optimal feeder vessel manage-
Vessels categories TEUs ment is based on following hypotheses:
Containership-Small Feeder ≤ 1000
• The berthing positions of vessels depend on the number
Containership-Regional Feeder 1000 - 1999
of busy docks and the work schedules of quays.
Containership-Feedermax 2000 - 2999
• The depth and length of berth can satisfy the requirement
Containership-Panamax 3000 - 4999
of vessel berthing.
Containership-Post-Panamax 5000 - 9999 • The quay can accommodate feeder and mother vessels.
Containership-ULCS ≥ 10000 • A berthing position can accommodate one vessel at a
time.
• The mother vessels have priority over the feeder vessels.
container ship with 21,86% of total vessels. We can observe • Approaching and manœuvring times are ignored.
that the TNMSC and Ocean terminals are served by ULCS and • Clearance time is integrated with the processing time.
Post-Panamax container ships. France terminal manages all • Each ship installed at the quay will be served without
container ships categories more specifically Panamax category. waiting time and taking on account of shifting.
Atlantic terminal manages Post-Panamax and Panamax cate- • The Estimated Arriving time correspond to the start
gories. The feeder vessels represent 16,82% of total container berthing time of vessel.
ships. Atlantic and France terminals manages most of it. • The times of docking and undocking are included in the
handling time of vessel.
V. M ATHEMATICAL MODEL • No interruption is authorized after handling operations
Our first objective is to minimize the waiting time cost of are started.
the mother ships in the sea. We have classified the container
vessels in three categories: mother’s vessels, feeder vessels The notations of the mathematical formulation are defined
and shuttle vessels. Mother vessels have the highest priority as follows:

Authorized licensed use limited to: Anelis Plus consortium. Downloaded on December 13,2020 at 13:35:54 UTC from IEEE Xplore. Restrictions apply.
A. Input data (σij + σji ) + (δij + δji ) ≥ Yiq + Yjq − 1, (9)
• M the set of mother vessels. ∀i, j ∈ V, i = j, q ∈ Q,

• F the set of feeder vessels. pi + l i ≤ lq Yiq , ∀i ∈ V, (10)
• S the set of shuttle vessels. q∈Q
• V := M ∪ F ∪ S the set of vessels. Yiq ∈ {0, 1} , ∀i ∈ V, q ∈ Q, (11)
• Q the set of berths.
σij , σji ∈ {0, 1} , ∀i, j ∈ V, i = j, (12)
• lq the length of quay q ∈ Q.
• Lmax = max{lq , q ∈ Q}. ti ≥ 0, ∀i ∈ V. (13)
• Tmax is an upper bound on the total time required to The objectives functions (1) and (2) aim to minimize the
service all the vessels. mother and feeder waiting time cost and containers transfert
• Fm the set of feeder vessels related to mother vessel m ∈ cost from a quay to another. The constraints (3) ensure that the
M. start berthing time of any feeder or mother vessel is greater
• li the length of vessel. than its ETA. The constraints (4) ensure that the handling
• ai the estimated time of arrival (ETA). of any feeder or mother vessel finishes before its ETD. The
• si the estimated time of departure (ETD). constraints (5) ensure that each feeder and mother vessel is
• hqi the estimated handling duration of ship i on quay q. assigned to a quay. The constraints (6) state that a ship cannot
• Tqq the cost for transferring containers between quay q be assigned to a quay whose length is shorter than that of
and q  . the ship. The constraints (7) to (10) prevent overlapping, in
both time and space, of vessels assigned to the same quay.
B. Decisional variables Constraints (11) to (13) define the variable types.
• ti the berthing start time of vessel i ∈ V . VI. C ONCLUSION
• pi the assigned berth position of vessel i ∈ V .
In this paper we have studied the feeder vessels management
• σij = 1, if in the space-time diagram, vessel i is
problem in a container terminal. We have detailed the prob-
completely to the left of vessel j, that is, vessel i is
lematic mainly related to container ports evolution regardless
completely processed before vessel j; 0, otherwise.
of the ship capacity and the evolution of maritime traffic. We
• δij = 1, if in the space-time diagram vessel i is com-
have discussed some related works, which have focused to
pletely below vessel j, that is, vessel i is positioned
study feeder vessel according to different problematics. Then,
completely to the right of vessel j (looking from the
we have presented the steps and the mathematical model that
quay); 0, otherwise.
corresponds to the first policy (optimization of the waiting of
• Yiq = 1, if a vessel i ∈ V is assigned to quay q ∈ Q; 0,
mother vessels).
otherwise.
We plan to implement the different vessel management poli-
Our objectives are to minimize the waiting time cost of cies and conduct the numerical experiment to find the optimal
mothers and feeder vessels and to minimize shuttle vessels solutions.
transportation cost
 ACKNOWLEDGMENT
min ci (ti − ai ) (1)
This work is co-financed by the European Union with the
i∈F ∪M
      European Normandy Regional Development Fund (ERDF) and
min Ymq Yfq Tqq (2) by the Normandy Regional Council.
q∈Q m∈M q  ∈Q f ∈F m

Subject to R EFERENCES
[1] A. Gharehgozli, J. P. Mileski, and O. Duru, “Heuristic estimation of
ti ≥ a i , ∀i ∈ F ∪ M, (3) container stacking and reshuffling operations under the containership
 q q delay factor and mega-ship challenge,” Maritime Policy & Management,
ti + hi Y i ≤ s i , ∀i ∈ F ∪ M, (4) vol. 44, no. 3, pp. 373–391, 2017.
q∈Q [2] CargoSmart Limited, “Shipment success through intelligent visibility,”
 CargoSmart’s Innovating, pp. 1–8, 2014.
Yiq = 1, ∀i ∈ F ∪ M, (5) [3] D.-H. Lee and J. G. Jin, “Feeder vessel management at container
q∈Q transshipment terminals,” Transportation Research Part E: Logistics and
Yiq = 0, ∀i ∈ F, q ∈ Q and li ≤ lq , (6) Transportation Review, vol. 49, no. 1, pp. 201 – 216, 2013.
 [4] S. Emde and N. Boysen, “Berth allocation in container terminals that
tj − (ti + hqi Yiq ) + Tmax (1 − σij ) ≥ 0, (7) service feeder ships and deep-sea vessels,” Journal of the Operational
Research Society, pp. 551–563, 2016.
q∈Q [5] Q. Zeng, Y. Feng, and Z. Chen, “Optimizing berth allocation and
∀i, j ∈ F ∪ M ∪ S, i = j, storage space in direct transshipment operations at container terminals,”
Maritime Economics & Logistics, vol. 19, no. 3, pp. 474–503, 2017.
pj − (pi + li ) + Lmax (1 − δij ) ≥ 0, (8) [6] A. Tsai, C. Lee, J. Wu, and F. Chang, “Novel wharf-based genetic
algorithm for berth allocation planning,” Soft Computing, vol. 21, no. 11,
∀i, j ∈ F ∪ M ∪ S, i = j, pp. 2897–2910, 2017.

Authorized licensed use limited to: Anelis Plus consortium. Downloaded on December 13,2020 at 13:35:54 UTC from IEEE Xplore. Restrictions apply.
 [7] E. Nishimura, A. Imai, and S. Papadimitriou, “Berth allocation planning
in the public berth system by genetic algorithms,” European Journal of
Operational Research, vol. 131, no. 2, pp. 282 – 292, 2001, artificial
Intelligence on Transportation Systems and Science.
[8] C. Y. Cheong, C. J. Lim, K. C. Tan, and D. K. Liu, “A multi-objective
evolutionary algorithm for berth allocation in a container port,” in 2007
IEEE Congress on Evolutionary Computation, Sept 2007, pp. 927–934.
[9] M. D. GKOLIAS, “The discrete and continuous berth allocation prob-
lem: models and algorithms,” Ph.D. dissertation, Rutgers University,
2007.
[10] J. Monios, “Cascading feeder vessels and the rationalisation of small
container ports,” Journal of Transport Geography, vol. 59, no. Supple-
ment C, pp. 88 – 99, 2017.
[11] H. Hsu, “A HPSO for solving dynamic and discrete berth allocation
problem and dynamic quay crane assignment problem simultaneously,”
Swarm and Evolutionary Computation, vol. 27, pp. 156–168, 2016.
[Online]. Available: https://doi.org/10.1016/j.swevo.2015.11.002
[12] C. Pani, M. Cannas, P. Fadda, G. Fancello, L. Frigau, and F. Mola,
“Delay prediction in container terminals: A comparison of machine
learning methods,” pp. 1–13, 07 2013.
[13] P. Frojan, J. F. Correcher, R. Alvarez-Valdes, G. Koulouris, and J. M.
Tamarit, “The continuous berth allocation problem in a container ter-
minal with multiple quays,” Expert Systems with Applications, vol. 42,
no. 21, pp. 7356 – 7366, 2015.
[14] S. Yan, C.-C. Lu, J.-H. Hsieh, and H.-C. Lin, “A network flow model
for the dynamic and flexible berth allocation problem,” Computers and
Industrial Engineering, vol. 81, no. Supplement C, pp. 65 – 77, 2015.
[15] C. Bierwirth and F. Meisel, “A survey of berth allocation and quay
crane scheduling problems in container terminals,” European Journal of
Operational Research, vol. 202, no. 3, pp. 615 – 627, 2010.
[16] C. Liang, H. Hwang, and M. Gen, “A berth allocation planning problem
with direct transshipment consideration,” Journal of Intelligent Manu-
facturing, vol. 23, no. 6, pp. 2207–2214, Dec 2012.
[17] C. PSanjeev Mathur, “Floating transfer stations: Simplicity in diversity,”
Port Technology International, Logistics S.p.A., Milan, Italy, Tech.
Rep. 32, 2011.
[18] D.-H. Lee, J. G. Jin, and J. H. Chen, “Terminal and yard allocation
problem for a container transshipment hub with multiple terminals,”
Transportation Research Part E: Logistics and Transportation Review,
vol. 48, no. 2, pp. 516 – 528, 2012.
[19] J.-F. Cordeau, G. Laporte, P. Legato, and L. Moccia, “Models and
tabu search heuristics for the berth-allocation problem,” Transportation
Science, vol. 39, no. 4, pp. 526–538, Nov. 2005.
[20] M. F. Monaco, L. Moccia, and M. Sammarra, “Operations research for
the management of a transhipment container terminal: The gioia tauro
case,” Maritime Economics & Logistics, vol. 11, no. 1, pp. 7–35, Mar
2009.
[21] “The simultaneous berth and quay crane allocation problem,” Trans-
portation Research Part E: Logistics and Transportation Review, vol. 44,
no. 5, pp. 900 – 920, 2008.
[22] P. fei ZHOU and H. gui KANG, “Study on berth and quay-crane
allocation under stochastic environments in container terminal,” Systems
Engineering - Theory & Practice, vol. 28, no. 1, pp. 161 – 169, 2008.

Authorized licensed use limited to: Anelis Plus consortium. Downloaded on December 13,2020 at 13:35:54 UTC from IEEE Xplore. Restrictions apply.