THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE

(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

A SYSTEMATIC APPROACH TO CONTAINER TERMINAL

PLANNING AND OPERATIONAL DECISION MAKING
Amr B. Eltawil
Industrial Engineering and Systems Management, Egypt-Japan University of Science and
Technology, PO Box 179 New Borg Elarab City, Alexandria, Egypt. Email: [email protected]

ABSTRACT: This paper elaborates the different operations research methods used to solve
container terminal operational decisions including terminal design, berth allocation, quay crane
assignment, quay crane scheduling, container storage space allocation, container location
assignment, retrieval, and pre-marshalling, as well as resource scheduling and terminal logistics in
general. Solution methods including optimization using mixed integer linear programming,
heuristic methods and discrete event simulation will be described. Implementing these methods
will have dramatic impact on container terminal performance and operating costs.

Keywords: container Terminal, operations research, heuristics, discrete event simulation, berth
allocation, quay crane assignment, container storage, container pre-marshaling.

INTRODUCTION
Suez 21 "the Suez economic corridor" is going to provide Egypt a fully integrated system of
trade, industry, transport, logistics, and technology within a dynamic region, supported by rail, air,
manufacturing clusters and a cluster of container terminals and ports. Developing the Suez Canal
corridor will transform Egypt into one of the world's leading logistics hubs connecting European,
Asian, and African markets‎1.
The corridor with its planned state-of-the-art infrastructure including a "magnetic levitation"
high-speed rail system is expected to maximize trade and goods flow to and from the gateway
ports on the Suez canal, leading to an increase in the container traffic not only in the surrounding
container terminals but in all Egyptian container terminals as well.
The Egyptian ministry of transport recorded a total container traffic of 6.62 million TEU in
Egypt in 2011, this number is expected to dramatically increase with the introduction of the Suez
Canal corridor. In order to cope with the increasing trends of container traffic, local container
terminals operating in the Suez Canal and all over Egypt coasts must systematically work on
improving the efficiency of its operations.
The need for optimization using methods of operations research in container terminal
operation has become more and more important in recent years. This is because the logistics
especially of large container terminals has already reached a degree of complexity that further
improvements require scientific methods. The impact of concurrent methods of logistics and
optimization can no longer be judged by operations experts alone. Objective methods are
necessary to support decisions. Such decisions are nowadays unthinkable without the effective
and efficient use of information technology as well as optimization and operations research
methods (Steeken et al 20042‎ ).
Different measures of performance can be used to gauge the success of a container terminal, a
particular one is the vessel time in port. Therefore, a crucial competitive advantage is the rapid
turnover of the containers, which corresponds to a reduction of the time in port of the container
ships. One easy but expensive solution is to increase the logistics resources (quay cranes, trucks
and gantry cranes), but, this is not typically available. The challenge in modern container terminal
is how to optimize operations with the available resources, and they are usually scarce, especially,
given the accelerating containerization of today's economy.

MARLOG 2 1
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

THE SUEZ CANAL AND CURRENT CONTAINER TERMINALS ACTIVITY

The Suez Canal is considered to be the shortest link between the east and the west due to its
unique geographic location; it is an important international navigation canal linking between the
Mediterranean Sea at Port Said and the red sea at Suez. The unique geographical position of the
Suez Canal makes it of special importance to the world and to Egypt as well. This importance is
getting augmented with the evolution of maritime transport and world trade. The maritime
transport is the cheapest means of transport, whereas more than 80 % of the world trade volume is
transported via waterways (seaborne trade).

The number of vessels passing through the Suez Canal over the period 2006 -2012 is presented
in Figure 1. The main vessels to use the canal are container ships; accounting for 38% of the total
number of vessels passing through the canal in 2010 .There was a growth in the number of
Container and General Cargo ships between 2006 and 2008 by 15 % and 19 % respectively.
Although there was a decline in numbers for all types of ships passing through the canal due to the
economic crisis in 2009 , the numbers from 2010 show that there is a potential increase in the
number of vessels passing through Suez Canal in the next years.

9000
8000
7000
6000
Number of Vessels

5000
4000
3000
2000
1000
0
2006 2007 2008 2009 2010 2011 2012
Container Ships 6974 7718 8156 6080 6852 7,178 6332
General Cargo 1670 2101 2069 1862 1618 1,395 1415
Ro-Ro Ships 433 386 350 263 270 254 284

Figure 1: The types of vessels passing Suez Canal (2006- 2010)

CONTAINER TERMINAL OPERATIONAL PROBLEMS AND SOLUTION

METHODS
In container terminals three areas exist: the quay side, the land side, and the container yard. In
order to check the performance of a container terminal, performance indicators are used. One of
the most important indicators is the berthing time of vessels. It is desired to minimize the berthing
time of a vessel which does not only depend on the performance of the Quay Cranes (QC), but also
depends on the performance of Yard Cranes (YC) and the whole logistics system of the terminal as
depicted in Figure 2 (Steeken et al 2004‎12‎ ).

MARLOG 2 2
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

Figure 2: The different components of container terminal logistics system

Classification of Container Terminal Problems

Container terminal decisions are classified in three levels: terminal planning decisions,
operative planning decisions, and real-time control decisions. This paper will focus on terminal
design and operative planning decisions as illustrated in Figure 3.

Figure 3: Logistics planning and control issues in seaport container terminals

Typical solution methods are mixed integer programming models, heuristic based methods, and in
modeling full operation of the terminal simulation based methods are typically used. Steenken et al
2004‎12‎ and Stahlbock, and Voß (2008)‎3 provide comprehensive reviews of operations research
methods used in solving the different classes of container terminal problems, while, Bierwirth and
Meisel (2010)‎4 present a focused review of berth allocation and quay crane scheduling problems.

There are many different decisions involved in operating container terminals and all these
decisions affect each other. For example, decisions about the storage of containers in the yard
directly affect the workloads of the yard cranes in the blocks and the traveling distances of the
Internal Trucks (ITs) and indirectly affect the efficiency of QCs. All these decisions are also
related to the berth allocation of vessels. Given the multi-criterion nature, the complexity of
operations, and the size of the entire operations management problem, it is impossible to make the
optimal decisions that will achieve the overall objectives. Logically, the hierarchical approach is
adopted to break the whole problem into smaller sequential problems. The input to a problem is
actually the output of its immediate predecessor, and is treated as a known quantity after the
preceding problem is solved. Figure 4 gives a typical hierarchical structure of operational
decisions in a container terminal (Zhang et al. 2003)‎5.

MARLOG 2 3
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

Figure 4: Decision Hirerchacy in The Container Terminal

The Berth Allocation Problem (BAP)

Problem Description: The entire quay in a Multi-User Terminal is partitioned into several
berths, and the allocation of the ships to the proper quay locations is based on the berth. The
problem of allocating ships to quay locations (or berths) is referred as the Berth Allocation
Problem (BAP). The problem is to assign a berthing position and a berthing time to each vessel,
such that a given objective function is optimized. An example for the graphical representation of a
berth plan with five vessels is shown in Figure 5. Berth planning has been shown to be an NP-hard
problem.

Figure 5: The Berth Allocation Problem

A ship’s handling time depends on the quay location where the particular ship is handled. More
precisely, it is assumed that the handling time is defined by the physical relationship between
ship’s quay location and its container storage location in the yard.

MARLOG 2 4
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

The handling time may remain unaffected regardless of where a ship is moored and where its
relevant container storage location is, if every ship employs a sufficient number of yard trailers
resulting in no interruption or delay of the quay crane cycle. This is only possible if there is a very
large fleet of yard trailers available to cover simultaneously all the ships in the terminal which
turns out to be very costly as there is a considerable redundant fleet when the terminal is not so
busy or when the ships are moored properly nearby their container storage location even when the
terminal is at a busy state. Based on this consideration, it is assumed that due to the limited size of
trailer fleet, every ship does not necessarily engage a trailer fleet large enough to keep seamless
movement of cranes.
In practice, the handling time also depends on the number of quay cranes engaged in handling
the ship. Usually the particular number of quay cranes that are assigned to a ship depends on the
size of the ship and the container movements to be made. An exception to the rule is made in the
occurrence of a late arrival of a ship requiring a quick turnaround, in which case many more cranes
are assigned to it.
Imai et al. (2005)‎6 classify the BAP into the following cases: (a) Discrete layout: The quay is
partitioned into a number of sections, called berths. Only one vessel can be served at each single
berth at a time. The partitioning can either follow the construction of the quay (Figure 6a) or is
organizationally prescribed to ease the planning problem (Figure 6b). (b) Continuous layout: There
is no partitioning of the quay, i.e. vessels can berth at arbitrary positions within the boundaries of
the quay (Figure 6c). For a continuous layout, berth planning is more complicated than for a
discrete layout at the advantage of better utilizing quay space. (c) Hybrid layout: Like in the
discrete case, the quay is partitioned into berths, but large vessels may occupy more than one berth
(Figure 6d) while small vessels may share a berth (Figure 6e). An indented berth results if two
opposing berths exist, which can be used to serve a large vessel from both sides (Figure 6f).

Figure 6: Spatial Classification of Berth Allocation Problem (Bierwirth and Miesel 2010)
In case of draft restrictions further spatial constraints must ensure that vessels are berthed at
positions of sufficient water depth. The general goal of berth planning is to provide fast and
reliable services of vessels. This is reflected in the literature by various objective functions.
Models to minimize the sum of the waiting and handling times of vessels (i.e. the port stay times)
clearly prevail, as shown in equation (1).
∑
Where mi is the starting time of handling of vessel ; a i the arrival time of ; hi the total
handling time of vessel . Further objectives are, for example, the minimization of the

MARLOG 2 5
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

workload of terminal resources and the minimization of the number of vessels rejected to be
served at a terminal. The performance of a berth plan is often measured in terms of costs which
allows to combine different goals in an overall cost function. Typical solution methods include
mathematical programming and heuristic methods. The typical inputs of the problem are:
 Estimated arrival time of the ships
 Length of each set of ships
 Quay length
 handling time of ship
 quay location with the minimum handling time of ship

 Quay location of each ship

And the typical outputs are:

 Start time of handling for each ship

 Completion time of handling for each ship

The Quay Crane Assignment Problem (QCAP)

In the QCAP a feasible berth plan and a set of identical QCs, which are available for service are
given. For all the vessels included in the berth plan, the volume of containers to be loaded and/or
unloaded is known as well as the maximum number of cranes allowed to serve it simultaneously.
The cranes are supposed to be lined up alongside the quay. They can be moved to every vessel but
they are not able to pass each other. The problem is to assign cranes to vessels such that all
required transshipments of containers can be fulfilled. In Figure 7, QCs numbers 2, 3 and 4 are
assigned to vessel 3. In time period 5, two cranes are shifted from vessel 3 to start serving vessel 2.
The QCAP and BAP are basically interrelated, because solving the QCAP can have a strong
impact on the vessels' handling times. Only in case of a discrete berth layout, where each berth
holds a set of dedicated cranes, an explicit assignment of cranes to vessels is not necessary.

Figure 7: The Quay Crane Assignment Problem

The QCAP problem is typically not addressed individually but in an integrated manner with
the BAP.

The Quay Crane Scheduling Problem (QCSP)

In the QCSP we consider a set of tasks, representing transshipment operations for a vessel, and a
set of assigned QCs. Precedence relations among tasks can be given to ensure that unloading

MARLOG 2 6
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

precedes loading and to represent the stacking of containers as defined by a stowage plan. Every
task must be processed (usually without preemption) once by a QC while a QC can process at most
one task at a time. A solution to the problem, called a QC schedule, defines a starting time for
every task on a crane. Usually, the minimization of the makespan of the QC schedule is pursued
because it represents the handling time of the considered vessel. Tasks to be scheduled on a QC
describe the granularity in which the workload of a vessel is considered in a QCSP model. Tasks
can be defined on the basis of bay areas or single bays (Figure 8a), or on the basis of container
stacks, container groups, or individual containers (Figure 8b) (Bierwirth and Miesel 2010)4‎ .

Figure 8: Storage location structure of a vessel (a) and a cross-sectional view of a bay (b).

As the number of tasks is bounded by the size of the vessel, the problem complexity is still
moderate. Reducing the granularity further on allows improving crane schedules at the expense of
increasing the problem complexity. Next to task attributes, also crane attributes appear in QCSP
models to specify the crane operations in more detail:
(a) Ready times: An individual ready time is used to designate the earliest possible operation of a
crane.
(b) Time windows: For each crane one or more time windows can be used to specify time spans
where the crane is available to serve the considered vessel. These time windows are often an
outcome of variable-in-time crane assignments.
(c) Positions: Initial and final positions are prescribed for a crane.
(d) Travel times: The speed of the crane movement is given in terms of the time required to
travel between bays.

The BAP, QCAP, QCSP Integration schemes

Basically, berth allocation, QC assignment, and QC scheduling decisions can be made in a
sequential fashion as shown in Figure 9. This way the overall problem complexity of seaside
operations planning is broken down into a series of decisions. Nevertheless, existing interrelations
between the planning levels are almost completely ignored by sequential planning. Often, this leads
to plans of poor overall quality. Imai et al (2008)7‎ , presents an alternate approach to problems
integration. Imai illustrated the simultaneous berth and quay crane allocation problem that
minimizes the total service time and a genetic algorithm based heuristic based solution.

MARLOG 2 7
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

Figure 9: Integrated Sequential Solution of BA, QCA, QCS problems

In order to be able to solve the problem simultaneously, many assumptions had to be made to
simplify it, for berth allocation:
(a) Each berth can serve one ship at a time.
(b) There are no physical or technical restrictions such as ship draft and water depth.
(c) Ship handling time is dependent on the berth where it is assigned.
(d) Ship is served after its arrival.
(e) Ship handling tasks must be finished without interruption once they get started.
For crane scheduling, the following assumptions are made:
(a) Ship handling requires a specific number of cranes and it does not begin till that number of
cranes is available.
(b) Cranes cannot move from one berth to another via other berths if the other berths are engaged
in ship handling.
(c) Cranes get through an idle berth having some cranes present by the pushing-in and pulling-out
procedure.

The Storage Space Allocation Problem

Problem description: Containers to be handled in the yard can be classified into the
following four types according to their status at different handling stages.
(a) Vessel discharge (VSDS) containers: I/B and transit containers on vessels before they are
unloaded and allocated to the yard.
(b) Container yard pickup (CYPI) containers: I/B containers already in the yard waiting for
picking up by customers.
(c) Container yard grounding (CYGD) containers: O/B containers before they are brought in
and stored in the yard.
(d) Vessel loading (VSLD) containers: O/B and transit containers already in the yard waiting
for loading to vessels.
In the storage space allocation problem the objective is to decide in which blocks to place the
VSDS and CYGD containers of each vessel.

MARLOG 2 8
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

Three different types of containers are handled in a container terminal. They are inbound (import
containers), outbound (export containers), and transshipment containers. The storage space
allocation decision can be handled separately, or taking into consideration all the container types in
the same time as illustrated in Zhang et al (2003)‎5, one approach to solve this problem will be
further illustrated.
At the first level, to minimize vessel berthing times, we balance the workload of RTGCs and QCs
for vessels. With workloads of a vessel dispersing in different blocks, the yard cranes in the blocks
serve as parallel servers processing jobs for the vessel, and the deberthing time of the vessel is the
maximal processing time of these parallel servers. Balancing the workload of parallel servers
generally works well to minimize the completion times of vessels. Similar results on the RTGC
deployment problem confirm that balancing workloads of blocks reduces delay in container
handling. There are several aspects of balancing at the first level. It is natural to balance the total
number of containers handled among different blocks, which equalizes the workload of RTGCs.
However, purely doing so ignores the key that VSDS and VSLD containers are related to the on-
time departures of vessels. We have to balance them and also highlight their effect as compared to
that of the total workload. See Section 4 for our choice of an objective function that considers
these two types of balancing.
The second level determines the number of containers associated with each vessel that constitutes
the total number of containers in each block in each period, in order to minimize the total distance
to transport the containers between their storage blocks and the vessel berthing locations.
Assignment of total numbers of containers to blocks: The first-level problem is
formulated as an integer programming model. Basically, the numbers of VSDS and CYGD
containers stored in each block for each planning period should be determined. The objective
function is:

∑ { [ {} ] [ {} ] [ {}
] [ {} ]} (2)
Where,
Git the total number of CYGD containers stored in block i that arrive at the terminal in period t,
1< = i< = B, 1< = t< = T ;
Dit the total number of VSDS (inbound and transit) containers stored in block i that are discharged
from vessels during period t, 1< = i< = B, 1< = t< = T ;
Lit the total number of VSLD (outbound and transit) containers stored in block i that are loaded
onto vessels in period t, 1< = i< = B, 1< = t< = T ;
P it the total number of CYPI containers stored in block i that are picked up by customers in period
t, 1< = i< = B, 1< = t< = T ;
Gitk the number of CYGD containers with full information stored in block i such that they arrive at
the terminal in period t and to be loaded onto vessels in period t + ‎k, 1< = i< = B, 1< = t< = T;
Ditk the number of I/B VSDS containers with full information stored in block i that are discharged
from vessels in period t and to be picked up in period t + ‎k, 1< = i< = B, 1< = t< = T;
Ritk the number of transit containers with full information stored in block i that are discharged from
vessels in period t and to be loaded onto other vessels in period t + ‎k, 1< = i< = B, 1< = t< = T;

In the objective function, (Dit + Lit) is the expected total number of vessel related containers that
need to be handled in block i during period t and (Dit ‎+ Lit + ‎ Git ‎+ Pit) is the expected total
number of containers to be handled in block i during period t. Therefore the two terms of (1)
measure the imbalances of the vessel related containers and of the total number of containers in the

MARLOG 2 9
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

blocks in each planning period, respectively. w1 and w2, the weights of the two terms in (1), are
adjusted according to the relative importance of the vessel related containers within the total
number of containers as interpreted by a terminal. Theoretically, it is possible to set (w1, w2) = (1,
0) or (0, 1), dependent on whether the vessel related containers or the total number of containers
are of utmost importance. In general, both w1 and w2 are strictly positive in practice, and are tuned
according to the needs of a container terminal.
Allocation of containers of each vessel to blocks: In the first level the total number of VSDS
and the total number of CYGD containers that can be assigned to each block in each planning
period has been determined. The second level determines the number of containers associated with
each vessel that constitutes the total number of containers in each block. The objective is to
minimize the total container moving cost. The moving cost is measured by the total distance
traveled by ITs between the berthing places of vessels and the storage blocks. After solving the
problem at the first level, the numbers of VSDS and CYGD containers to be placed to each block
in each planning period, Ditk ‫ ‏‬Ritk and Gitk, are fixed. Therefore, the second level decisions can be
made for VSDS and CYGD containers separately in each planning period to decide vessel
identifications of containers. The problem can be formulated as a general transportation problem
with the objective function of minimizing the total travelled distance by transportation trucks as
illustrated by Figure 10.

Figure 10: Second Level Problem Representation

After the storage block of the container has been assigned, the exact stacking location of the
container should be now determined as the row, tier and bay location in the target stack.
The Container Stacking Location assignment problem
Problem description: When storing or retrieving a container at the storage yard, if some
containers are on the moving path of inbound or outbound container, the obstructive containers
should be first retrieved from the storage yard in order to provide the inbound or outbound
container with the free moving path. This problem can be defined as the assignment of the inbound

MARLOG 2 10
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

and outbound containers to the storage yard with aim of minimizing the number of obstructive
container moves. Allowing dynamic reallocations often leads to a significant improvement in
space utilization.
The focus is to utilize the storage area in a more optimal manner thus reducing the time
required for the yard machines to transfer the containers from the storage area to the marshalling
area for loading onto the ships. The location assignment problem aims to assign each
import/export container to its slot. The place of that slot ensures to minimize the number of
obstructive container moves to locate a container. Also the assignment puts into consideration the
utilization of the storage area as possible. The typical inputs that may be considered are:

 Inbound and outbound time of each container.

 Depth and width of the storage yard.
 The travel time of each container between locations.
 The travel time of each container between the arrival positions.
 The travel time of each container between location and the departure position.

 Container location: the index of block, bay, row, and tier of each assigned container.
And the typical outputs are:

 Set-up time: time required to access the desired container at the storage area.

The Container Pre-Marshalling Problem

Problem Description: Export containers arrive at the terminal in random order. Typically,
containers start arriving at the terminal as early as seven days in advance and containers scheduled
for earlier ships are likely to arrive earlier than those that are scheduled for later ships. Therefore,
by the time loading starts for a particular containership, it is often the case that many of the
containers for this current ship are buried in the container stacks beneath other containers waiting
to be loaded to a later ship. Containers can also be stacked in the wrong order due to lack of
accurate information or other reasons. Because containers in the yard can be accessed only from
the top of a stack, re-handles will be needed if the target container is not located at the top.
Generally, the container stacking problem is classified into three main types: the pre-
marshalling problem, the remarshalling problem and the container retrieval problem (Steeken et al
2004)‎2. The pre-marshaling problem, is the problem of converting an initial layout of a bay into a
desired final layout within which containers are stacked above each other with the priority of
stacking the containers that will be served first at the top of the stack as shown in Figure 11
(Geith et al. 2013)‎8. This will minimize or eliminate future additional reshuffles by the YC. In the
case of the container retrieval problem, it is desired to remove a container from the bay with
minimum number of reshuffles, and then remove another container and so on till the bay is empty.
The three classes of the problem are of prime importance knowing that in large container
terminals, the average number of movements made by yard cranes is 15,000 movements per day,
which means that the reduction of such moves will dramatically improve operations and
efficiency.
One way to reduce the number of re-handles while loading is through pre-marshalling. In a
container yard, pre-marshalling means to re-position the export containers before the loading
process starts, so that the containers can be loaded with few or no re-handles. Pre-marshalling
requires additional cost, but is executed when the ship has to be loaded as fast as possible.

MARLOG 2 11
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

Figure 11: The container Pre-marshalling problem

The optimization goal is to minimize the number of movements required to transform the
container yard from the initial layout to the desired final layout. The final layout is determined
from the vessel stowage plan. The typical problem inputs are:

 Total number of stacks in the blocks and their maximum stacking height.
 The number of container types, size, and weights.
 Movement cost of yard equipment per unit distance within the marshalling yard.

 The optimum Sequence of containers movements to reach the desired layout.

And the typical outputs are:

 The number of relocations.

Typical solution methods include integer programming models, branch and bound, and
heuristics. The most recent work is by Geith et al (2012)‎8 that illustrates a labeling and sorting
heuristic that achieved improved results in medium sized problems than other heuristics in the
literature.
Scheduling of Material Handling Equipment in Container Yard
Problem Description: Scheduling yard equipment is concerned with the operation order of
quay cranes, dispatching yard trucks to containers, and dispatching yard cranes to yard trucks
in storage yard. These problems are interrelated, and the efficiency of container terminal
operations depends on the coordination of different types of equipment.
The scheduling problem of a container handling system is formulated as a Hybrid Flow Shop
Scheduling (HFSS) problem. A hybrid flow shop consists of a series of production stages, each of
which has several machines operating in parallel. Some stages may have only one machine, but at
least one stage must have multiple machines. The flow of jobs through the shop is unidirectional.
Each job is processed by one machine at each stage and it must go through one or more stages.

The elements of the problem are:

(i) Jobs: Each container must go through several handling operations, and can be associated
with a job. A job is defined as a complete loading/unloading process for a container.
(ii) Machines: There are three different sets of machines: quay cranes, yard cranes, and yard
trucks.
(iii) Operations: Each job consists of three operations: a transfer operation of a container
from/ onto the ship, a transfer operation within the storage yard, a transfer operation
between quay cranes and yard cranes by yard trucks.

MARLOG 2 12
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

While, comparing the classical HFSS problem with the container yard scheduling problem , the

 Job precedence constraints - As for loading operation, containers in the hold must precede the
latter problem has several unique characteristics as follows:

 Blocking - There is limited or no buffer between two successive machines, therefore,

containers on the deck of the same vessel.

blocking happens when the buffer is full. E.g. when the yard truck carries a container to a

 Setup times - In container terminals, there is empty movement when a crane or yard truck
quay crane that is handling another container, the yard truck has to wait for the quay crane.

moves between two jobs. For example, once a yard truck carries an outbound container to a
quay crane, it has to make an empty trip to the storage yard in order to proceed next
container.

The objective of the scheduling problem is to minimize the make span (total service time) with
highest equipements utilization. Some researches aim for minimizing the total operating cost.

 The number of containers, and the number of yard equipment.

The typical problem inputs are:

 The set of all the jobs.

 The set of containers that might be processed on specific machine.
 The precedence relationship between containers.
 Processing time of each container at each stage.
 Setup times between containers for every stage.
 Number of vessels to be served, and the number of tasks on the vessel.
 Travel time of the machines from location to another.
 Variable and fixed cost of using each type of machines.
 Tardiness cost of the vessel, and earliness income of the vessel.
 Set of pairs of tasks that cannot be performed simultaneously.

Scheduling problems are solved using several methods in order to achieve the objective of
minimizing the make span to serve a set of loading and unloading ships in a given time horizon

 The assignment of each operation of the containers for every equipment in every stage.
and minimize the total cost. The following are typical outputs required from solving the problem:

 The precedence of operations on specific equipment.

 The starting time of each container at every stage.
 The number of operations assigned to the same machine.
 The completion time of the last container.
In order to solve the scheduling problems in yard planning there is several methods used to get
the optimum solution. Some of these methods give exact solutions and some give approximate
solutions. Exact algorithms include mixed integer linear programming models, and branch and
bound models. While, approximate algorithms include heuristics, genetic algorithms and discrete
event simulation.

Simulation: Many complex systems such as manufacturing, supply chain, and container
terminals are too complex to be modeled analytically. Discrete event simulation has been a useful
tool for evaluating the performance of such systems. However, simulation can only evaluate a
given design, not providing optimization function. Therefore, the integration of simulation and
optimization is needed. Simulation optimization is the process of finding the best values of some
decision variables for a system where the performance is evaluated based on the output of a
simulation model of this system.

MARLOG 2 13
 THE INTERNATIONAL MARITIME TRANSPORT & LOGISTICS CONFERENCE
(MARLOG 2)
SUSTAINABLE DEVELOPMENT OF SUEZ CANAL REGION
17 - 19 MARCH 2013

The simulation optimization for scheduling loading containers consists of two relative
unattached modules, namely, optimization module and simulation module. This integrating
method realizes the separation of optimization algorithm from model, which cannot only improve
the optimization algorithm, but also help for the software integration.

CONCLUSIONS
As the container transport system is capital intensive, the turnaround time of ships at container
terminals is an important factor for liner shipping companies to decrease their cost. The turnaround
time includes berthing, unloading, loading and departure, therefore, optimization of every
operation is critical to the overall performance of the container terminal.
This paper presented a systematic approach to container terminal problems decision making,
especially in terminal operative planning decisions.

REFERENCES
1. Egyptian National Competitiveness Council (ENCC), " A Sustainable Competitiveness
Strategy for Egypt ", http://www.encc.org.eg, 2012 [accessed February 2013]
2. Dirk Steenken, Stefan Voß, and Robert Stahlbock, " Container terminal operation and
operations research – a classification and literature review", OR Spectrum (2004) 26: 3–49
3. Robert Stahlbock, and Stefan Voß, "Operations research at container terminals: a literature
update", OR Spectrum (2008) 30:1–52, DOI 10.1007/s00291-007-0100-9,
4. Christian Bierwirth, Frank Meisel, " A survey of berth allocation and quay crane scheduling
problems in container terminals", European Journal of Operational Research 202 (2010)
615–627, doi:10.1016/j.ejor.2009.05.031
5. Chuqian Zhang, Jiyin Liu, Yat-wah Wan, Katta G. Murty, Richard J. Linn, "Storage space
allocation in container terminals", Transportation Research Part B 37 (2003) 883–903,
doi:10.1016/S0191-2615(02)00089-9.
6. Akio Imai, Xin Sun, Etsuko Nishimura, and Stratos Papadimitriou, "Berth allocation in a
container port: using a continuous location space approach", Transportation Research Part B
39 (2005) 199–221.
7. Akio Imai, Hsieh Chia Chen, Etsuko Nishimura, Stratos Papadimitriou, " The simultaneous
berth and quay crane allocation problem", Transportation Research Part E 44 (2008) 900–
920, doi:10.1016/j.tre.2007.03.003
8. M. S. Gheith, A. B. El-Tawil, N. A. Harraz, "Container Pre-marshalling Problem: A Review
and Solution Framework", Proceedings of the IEEE International conference on Industrial
Engineering and Management, Hong Kong, December 2012.

MARLOG 2 14