A Metamodelling Approach to the Management of

Intermodal Transportation Networks

Valentina Boschian

, Mariagrazia Dotoli

, Maria Pia Fanti

Giorgio Iacobellis

and Walter Ukovich

Department of Electronics, Electrical, Engineering and Computer Science

University of Trieste, Trieste, Italy

Department of Electrical and Electronic Engineering 70125 Bari, Italy

Abstract. The paper develops a novel and broad metamodel of a generic Inter-

modal Transportation Network (ITN), devoted to provide a reference model for

the real time management and control of such systems. In order to take opera-

tional decisions, the presented model describes in detail the ITN structure and

dynamic evolution that is updated on the basis of the information obtained in

real time by modern information and communication technologies tools. The

proposed metamodelling approach consists in employing a top down procedure

and is based on the UML formalism, a graphic and textual modelling language

intended to describe systems from structural and dynamics viewpoints. Hence,

the paper models a generic ITN starting from the network description and

shows as an example the metamodel of one of the most important nodes that

compose it: the port subsystem. To this aim, we present the main UML dia-

grams describing the structure and dynamics of a case study.

1 Introduction

An Intermodal Transportation Network (ITN) can be defined as a logistically linked

system integrating different transportation modes (rail, ocean vessel, truck etc.) to

move freight or people from origin to destination in a timely manner [9]. The 21st

century will see a renewed focus on ITN, driven by the necessity of moving ever

growing quantities of goods and by the technological evolution each of the transpor-

tation modes has recently gone through. However, ITN decision making is a very

complex process, due to the dynamical and large scale nature of the ITN, the hierar-

chical structure of decisions, as well as the randomness of various inputs and opera-

tions. Typically, ITN management techniques are based on a three-level hierarchy:

strategic, tactical and operational ones. Strategic planning involves ITN design and

considers time horizons of a few years. Tactical planning basically refers to the opti-

mization of the flow of goods and services through a given ITN. Finally, operational

planning is short-range planning and involves transportation scheduling at all trans-

porters on an hour-to-hour basis, subject to the changing market conditions as well as

to unforeseen transportation requests and accidents.

Boschian V., Dotoli M., Fanti M., Iacobellis G. and Ukovich W. (2009).

A Metamodelling Approach to the Management of Intermodal Transportation Networks.

In Proceedings of the 3rd International Workshop on Intelligent Vehicle Controls & Intelligent Transportation Systems, pages 96-105

 SciTePress

The related literature has largely addressed the ITN modelling and management prob-

lems at strategic and tactical levels. Recently, entrepreneurs and researchers are being

attracted by the key problem of using effectively and efficiently the latest develop-

ments of ICT (Information and Communication Technologies) for ITN operational

management [6], [11], [12], [13]. In fact, since intermodal transportation is more

data-intensive than conventional transportation, ICT are considered a primary “ena-

bling tool” for the safe and efficient real time management and operation of ITN.

Indeed, today the effective use of the modern ICT tools has made it possible to know

the state of the system in real time and therefore manage and change on-line paths,

vehicle flows, deliveries and orders. In order to operate such choices, there is a need

of dynamic models that can track the state changes of the various system components

and determine operation indices typical of the real time management, such as utiliza-

tion, traffic indices and delivery delays [14]. In the domain of ITN models at the

operational level we recall the class of discrete event system models [1], [3], [4] and

of the simulation models [13], [14]. However, the above models are designed to de-

scribe a particular ITN and do not fully take into account the multiplicity of elements

that can influence the ITN dynamics and the corresponding information structure.

Since ITN are often complex and distributed systems, they have also been effectively

represented by multi-agent techniques. However, this promising approach to transpor-

tation and traffic management is still at its early stages [2].

This paper develops a novel metamodel of ITN at the operational level intended

for real time management and control of such systems. The metamodel has a general

and modular structure and is characterized by high information integration. In order

to take operational decisions, the presented model describes in adequate detail the

structure and dynamic evolution of the ITN and is updated on the basis of data ex-

changed by the players in the system and information obtained in real time by using

modern ICT tools [7]. The proposed approach consists in applying metamodelling by

a top down procedure based on the UML formalism [9], [10], a graphic and textual

modelling language intended to understand and describe systems from various view-

points. Hence, UML enables us to describe the structure and dynamics of a generic

ITN starting from the description of the network, until the model of the most impor-

tant entities that compose it (classes) and their corresponding activities. Moreover,

UML unifies the formalism by using appropriate and effective diagrams that can be

easily translated into a simulation software. Indeed, the approach to the management

and control of ITN is based on the construction of a reference model that simulates

the evolution and dynamics of ITN and provides to the control modules the knowl-

edge base necessary for decisions in real time. To this aim, the model reproduces the

behaviour of the ITN by storing the real events such as variations in demand and

orders, blockages, accidents, breaks and all those occurrences that interact with the

flow and management of materials and transporters. Hence, based on the knowledge

of the reference model state and the events, decision makers can make the appropriate

choices optimizing suitable performance indices. To the best of the authors’ knowl-

edge, such a UML based metamodel approach has never before been proposed for

ITN.

The paper is organized as follows. Section 2 describes the main steps of the ITN

metamodel approach. Subsequently, sections 3 and 4 respectively describe the static

and dynamic diagrams in the ITN metamodel. A conclusion section closes the paper.

2 The Metamodel of Intermodal Transportation Networks

We consider a generic ITN constituted by a set of terminals (ports, airports, railway

stations, etc.), together with the interconnections between them and land infrastruc-

tures. Metamodelling is a technique that applies to models [5]. A metamodel provides

an accurate description of the constructs and rules needed to obtain semantic models

and encapsulates all the concepts that are necessary to describe the structure and dy-

namics of a particular system.

The metamodelling approach presented in this paper is a top-down approach that

decomposes the system to gain insight into its compositional sub-systems. In the top-

down approach an overview of the ITN is formulated, specifying any first-level sub-

systems: ports, airports, railway stations, intermodal terminals, ground, sea and air

connections, information systems, carrier and freight forwarder. Subsequently, each

subsystem is refined in detail from the structural and dynamical points of view, using

suitable UML diagrams, such as class diagrams and activity diagrams. In the follow-

ing sections we detail the main steps of the proposed ITN metamodelling approach.

First, a procedure addressing static models is devised, defining the so-called package

diagram and the class diagrams. Subsequently, process flows are considered and

dynamic models are obtained, by referring to the so-called activity diagrams.

3 Static Modelling of Intermodal Transportation Networks

3.1 The Package Diagram

The first step of the presented metamodelling approach consists in identifying the

main subsystems composing an ITN. These can be divided into structural subsystems

(i.e., ports, airports, railway stations, intermodal terminals, ground, sea and air con-

nections, carrier and freight forwarder) and the information system. They represent

the generic concepts used within the metamodelling framework and are modelled in

UML by packages. However, such subsystems are complex nodes that can be consid-

ered composed by other generic objects (or classes). Hence, we represent the overall

ITN by the UML package diagram [9], [10]. Packages are groups of entities related to

each other. Figure 1 depicts the package diagram of a generic ITN. We identify the

following seven packages that form the ITN: the port, the airport, the railway station,

the ground, sea and air connection, the intermodal terminal, the information system

and the carrier and freight forwarder (see Fig. 1). Each package is composed by dif-

ferent classes representing structural basic objects interconnected with each other.

Arrows show the cases in which a class in one package needs to use a class in another

package. This causes a dependency between packages: for example, the information

system is updated on the basis of data obtained in real time using modern ICT tools.

We assume that each package includes an information class representing the informa-

tive structure devoted to manage the considered system. However, we consider also

the possibility of the existence of a centralized information system that can manage

and coordinate different packages. For example, the port package contains an infor-

mation class that manages the flow of trucks, trains, cranes, etc. On the other hand,

the external and higher level information system can control the interactions between

the port and the infrastructures, by receiving data from the port area and the ground,

sea, rail and air connections.

PortIntermodal terminalGround Sea AirRailway Station Airport

Information System

Carrier and freight forwarder

Fig. 1. The package diagram of the ITN: arrows show dependence among packages.

Port_AreaQuay

Loading_Area

Unloading_Area

Parking_Area

Truck_Parking_Area

Tow_Parking_Area

Container_Yard

Landside_Yard

Seaside_Yard

Crane_Area

Crane

Queue

Gate_Queue

Exit_Queue

Truck_Gate_Queue

Ship_Gate_Queue

Crane_Gate_Queue

Parking_Gate_Queue

Train_Gate_Queue

Parking_Exit_Queue

Truck_Exit_Queue

Ship_Exit_Queue

Train_Exit_Queue

Transportation_Means

Train

Truck

Ship

Tow

Customs_Authority

check

Intermodal_Transport_System

Container Pallet

Freight

Costs Manager

Management_system

Routing

Sequencing

Priority

Synchronization

Timing

Fig. 2. The port class diagram: connections show relationships among objects in the class.

3.2 The Class Diagrams

The subsequent step of the static modelling consists in setting up the so-called class

diagrams, specifying the configuration of the various packages defined in the previ-

ously discussed package diagram [9], [10]. A class diagram describes the types of

objects in the ITN and the various static relationships between them. These are repre-

sented by different pictorial connections and may be relationships of association

(solid line), aggregation (solid line with a clear diamond at one end), composition

(solid line with a filled diamond at one end), inheritance or generalization (solid line

with a clear triangle at one end), realization (dashed line with a clear triangle at one

end) and dependency (dashed line with an arrow at one end). Moreover, the class

diagram may show the features of a class, i.e. the name, attributes and operations of

the class.

As an example, we show in Fig. 2 the class diagram of one of the packages in Fig.

1, namely the Port: for the sake of brevity we omit the class diagrams of the remain-

ing packages in Fig.1, since they may be set-up similarly to the diagram described

here. The main classes included in the diagram in Fig. 2 are the Intermo-

dal_Transport_System, the Management_System and the Cost_Manager. In the se-

quel we briefly describe such classes together with their typical attributes and meth-

ods. Obviously, these features may vary depending on the ITN under study [14].

The Intermodal_Transport_System class represents the overall port terminal sys-

tem. The class attributes are: 1) the dynamic lists of ships, trains and trucks currently

in the terminal; 2) the dynamic lists of ships, trains and trucks already served by the

operators in the terminal or by the quay cranes, waiting for permission to exit from

the terminal; 3) the dynamic lists of ships, trains and trucks queued and waiting for

service; 4) the dynamic lists of ships, trains and trucks currently being served; 5) the

dynamic lists of ships, trains and trucks currently leaving the terminal; 6) the lists of

occupied quay cranes and available ones. The class methods are: 1) the registration of

ships, trains and trucks entering the terminal; 2) the extraction from the list of ships,

trains and trucks waiting for service; 3) the extraction from the list of available

cranes; 4) the assignment of a crane to a specific task of freight loading/unloading; 5)

the crane activation; 6) the extraction from the list of ships, trains and trucks exiting

from the terminal; 7) the update of the list of served ships, trains and trucks; 8) the

update of the list of waiting ships, trains and trucks; 9) the update of the list of ships,

trains and trucks exiting the terminal; 10) the update of the list of available cranes.

The Intermodal_Transport_System class aggregates the following classes: Port_Area,

Queue, Transportation_Means and Customs_Authority.

The Port_Area class represents the physical port area, modelled generalizing the

following classes: Quay, where the ship loading/unloading processes take place, Con-

tainer_Yard, representing the freight stock area, Crane_Area, describing the crane

pick up and delivery point and Parking_Area, representing the zones where vehicles

are parked. Hence, all these classes inherit the properties and basic services of their

parent class, i.e. the Port_Area. This has the following attributes: 1) dimensions; 2)

list of occupied locations; 3) list of unoccupied locations; 4) opening time; 5) closing

time; 6) number of operators. Its methods are: 1) the extraction from a list of locations

in the port area; 2) the access control; 3) the assignment of a location to an entity

(ship, train, truck); 4) the clearing of a location upon the leaving of an entity. The

Container_Yard class defines the storage area where freight is stored and waits for

being delivered to their destination. The Quay class models the quay where ships are

loaded/unloaded. Hence, the Quay class aggregates the Ship class, while it general-

izes the Loading_Area and Unloading_Area classes. The Crane class attributes are: 1)

the crane type (quay crane loading/unloading freight onto/from the ship, stacker crane

retrieving/storing freight from the Container_Yard, automated guided vehicle or trac-

tor moving freight from the quay area to the container yard area, etc.); 2) the crane

100

identification number. The Loading_Area class models the ship loading area, while,

the Unloading_Area class represents the ships unloading area. The Parking_Area

class is the area where trucks wait either to be loaded or to be embarked. The class

also includes means of transportation waiting for any reason. The Parking_Area class

obviously includes the Truck class and generalizes the Truck_Parking_Area and

Tow_Parking_Area classes. The Container_Yard class includes the Container class

and generalizes the Seaside_Yard and Landside_Yard classes. The Seaside_Yard

(Landside_Yard) class represents the warehouse area where containers, freight, tow

or pallets to load on trains or trucks destined to the inland (to be shipped) are stored.

The Queue class enables the management of queues and the computation of wait-

ing times and hence of costs. The attributes of this class are: 1) the maximum number

of entities a queue may have; 2) the queue management policy; 3) the velocity of

queue management; 4) the costs associated to the waiting time in queue of a user; 5)

the queue identification number; 6) the current number of users in a queue; 7) a Boo-

lean flag indicating whether the queue is full. Its methods are: 1) the inclusion of an

entity in the queue; 2) the cancellation of an entity from the queue; 3) the queue man-

agement; 4) the queue cost computation. Its children classes are called Gate_Queue

and Exit_Queue, respectively representing the input and output queues in all the port,

e.g. queues in parking area, in crane area, etc. The Gate_Queue and Exit_Queue both

have as children classes those representing the truck, train, ship, parking and crane

queues. For instance, the Parking_Gate_Queue and the Parking_Exit_Queue classes

respectively represent the input and output truck queues in the parking area.

The Transportation_Means class represents the transportation means circulating in

the port, i.e., trucks, trains and ships. Hence it generalizes the Truck, Train and Ship

classes and is associated to the Customs_Authority class. Indeed, customs have the

task of controlling the arriving transportation means and applying the corresponding

taxes. The Transportation_Means class exhibits the attributes: 1) the transportation

means identification number; 2) the transportation means dimensions; 3) the carrier

name; 4) the category of transported goods (e.g., hazmat); 5) the goods place of ori-

gin; 6) the goods destination; 7) the goods weight; 8) the identification number of the

carried tows, containers, freight or pallet; 9) a Boolean flag indicating full load; 10)

the current waiting time for the unload and load operation. The class methods are as

follows: 1) the load/unload operation; 2) the waiting time computation. Note that the

class diagram shows that the Truck class is included in the Train and Ship classes,

while it aggregates the Tow, Container, Freight and Pallet classes. The Container

class includes the Freight and Pallet classes, which in turn includes the Freight class.

Note that both the Tow and Container class have one attribute, the identification

number. On the contrary, the Pallet class attributes are the identification number and

its capacity, while the Freight class attributes are the identification number and di-

mension. The Train class represents trains moving freight entering or exiting from the

port. Such a class includes the Container class and makes use of the Intermo-

dal_Transport_System class, so that trains remain in the railway as minimum time as

possible. The Ship class models ships berthed at the port. This class includes the

Container and Truck classes.

The Customs_Authority class represents the customs and their tasks of controlling

all transportation means and their freight and applying taxes. It is therefore connected

to the Transportation_Means class by an association function named “check”. Its

101

attributes are: 1) the opening and closing times; 2) the control time of goods; 3) the

number of customs operators. Its methods are: 1) the freight control; 2) the tax as-

signment.

The Management_System class is devoted to managing the whole Intermo-

dal_Transport_System class. This class provides the rules to perform choices and

decisions in the ITN. Hence, it uses the database provided by the Intermo-

dal_Transport_System class to elaborate the decisions. It aggregates the following

classes: the Synchronization class, introduced to minimize operation delays in the

intermodal terminal so as to maximize synchronization of operations; the Priority

class, which deals with the assignment of priorities to the transportation means (e.g.,

trucks transporting hazmat are assigned priority or a FIFO logic may be considered);

the Routing class, which describes the operation path that has to be followed; the

Sequencing class, assigning the subsequent operation that the system has to execute;

the Timing class, which determines the timing of the next operation.

Finally, the diagram in Fig. 2 includes the Costs_Manager class that computes the

managing costs of the container terminal. Hence, this class uses data stored in the

Intermodal_Transport_System and exhibits the cost calculation operations.

4 Dynamic Modelling of Intermodal Transportation Networks

In this section we employ activity diagrams to provide a “dynamic view” of the sys-

tem. Activity diagrams aim to describe the logic of the involved processes and the

workflow [9], [10]. They are similar to flowcharts, but they allow representations of

parallel elaborations in order to explain the critical points in the processes and work-

flow of the whole system by pointing out all the possible paths, parallel activities and

their subdivisions. The main elements of these diagrams are: the initial activity (de-

noted by a solid circle); the final activity (denoted by a bull’s eye symbol); activities,

represented by a rectangle with rounded edges; arcs, representing flows, connecting

activities; forks and joins, depicted by a horizontal split, used for representing concur-

rent activities and actions respectively beginning and ending at the same time; deci-

sions, representing alternative flows and depicted by a diamond, with options written

on either sides of the arrows emerging from the diamond; swim lanes, highlighting

responsibilities; signals representing activities sending or receiving a message, which

can be of two types: input signals (message receiving activities), shown by a concave

polygon, and output signals (message sending activities), shown by a convex poly-

gon. Moreover, activities may involve different participants in a system. Hence, parti-

tions are used to show which actor is responsible for which actions and divide the

diagram into columns or swim lanes.

102

can be unhooked and successively it is loaded into the ship [1]. At this point, on the

basis of suitable priority rules that are established by the management system, either a

complete truck can be loaded into the ship or a tow is unhooked and waits for the

assignment of an idle crane. More precisely, trucks move to the parking area and can

subsequently be loaded onto the ship. Alternatively, after the unhooking process,

tows are loaded on a crane and are stored in the landside yard were they wait for the

ship loading. Finally, tows are loaded by the quay crane on the ship.

We remark that the described logic flows may be significantly improved by suit-

able arrangements employing modern ICT tools, e.g. setting up a truck tracking de-

vice, an electronic transportation document system, etc. Obviously, the corresponding

activity diagram can be obtained by suitably updating the ship loading process de-

picted in the diagram of Fig. 3a.

4.2 The Activity Diagram of the Ship Unloading Process

The logic flow associated to the ship unloading operations is similar to the previously

analyzed loading phase and is described as follows (see Fig. 3b). The ship enters the

port and the freight forwarder prepares the documents to unload vehicles and goods.

These documents are uploaded in electronic form and printed by the customs author-

ity. Data contained in the documents and freight are checked. Tariffs to unload the

ship are then paid by the freight forwarder, successively the payment is registered and

the ticket is printed. When the documents are checked and the tickets are ready, the

authorization is issued. Hence, the port area staff, on the basis of the priority assigned

by the management system class either enables the exit of trucks or assigns an idle

crane to tows. In the latter case the tows are parked in the seaside yard, where they

wait for trucks. Finally, trucks hook the tows and leave the port.

Similarly to the activity diagram of the ship loading process, also the activity dia-

gram of the unloading process may be significantly improved by modern ICT tools.

5 Conclusions

The paper presents a novel top down procedure to develop a metamodel of Intermo-

dal Transportation Networks (ITN), devoted to real time management and control at

the operational level. To this aim, the model describes in detail the structure and dy-

namic evolution of the ITN so that it can be updated in real time using information

from the real system obtained by modern information and communication technolo-

gies tools. The proposed metamodelling procedure is based on the UML formalism, a

graphic and textual language able to describe systems from structural and dynamics

viewpoints. To show the proposed model effectiveness, the paper focuses on a par-

ticular node of the ITN, i.e. the port, with particular reference to the port of Trieste

(Italy). The detailed description of the main system components and of two basic

processes of the port show how the UML diagrams can effectively depict the struc-

ture and activities of such complex and large systems. Hence, the proposed metamod-

elling approach and the used UML formalism provide a reference model that closely

104

simulates the evolution of the ITN. This feature of the reference model is crucial and

can be effectively employed in order to supply the control modules with the knowl-

edge base necessary for decisions in real time. Future research will address the model

of all the nodes of the ITN and the specification of the decision and control modules.

In addition, we plan to apply and validate the proposed approach to a real case study.

To this aim, preliminary studies are being carried out with several European ports and

authorities in the framework of a research project funded by the European Commis-

sion. Finally, further studies could be developed to support learning from data in the

provided model, e.g. adding decision support and control modules based for instance

on agent techniques.

References

1. Danielis, R., Dotoli, M., Fanti, M.P., Mangini, A.M., Pesenti, R., Stecco, G., Ukovich, W.:

Integrating ICT into Logistics Intermodal Systems: a Petri Net Model of the Trieste Port.

Proc. Eur. Control Conf., Budapest Hungary (2009).

2. Davidsson P., Henesey L.E., Ramstedt E., Törnquist J., Wernstedt F.: An Analysis of

Agent-Based Approaches to Transport Logistics. Transportation Research Part C: Emerg-

ing Technologies, Vol. 13 (2005), 255-271.

3. Di Febbraro, A., Porta, G., Sacco, N.: A Petri Net Modelling Approach of Intermodal

Terminals Based on Metrocargo System. Proc. Intelligent Transportation Systems Conf.

(2006), 1442–1447.

4. Fischer, M., Kemper, P.: Modeling and Analysis of a Freight Terminal with Stochastic Petri

Nets. Proc. 9th IFAC Symposium Control in Transportation Systems (2000).

5. Ghazel M., Toguyéni A., Bigand M.: An UML Approach for the Metamodelling of Auto-

mated Production Systems for Monitoring Purpose. Computers in Industry, Vol. 55 (2004),

283-299.

6. Giannopoulos, G.A.: The Application of Information and Communication Technologies in

Transport. Eur. J. Oper. Res., Vol. 152 (2004), 302-320.

7. Ikkai, Y., Oka, H., Komoda, N., Itsuki, R.: An Autonomous Distributed Information Sys-

tem for Logistics Control with Data Carriers. 2003 IEEE Int. Conf. Industrial Technology

(2003), 1079-1082.

8. Macharis, C., Bontekoning, Y.M.: Opportunities for OR in Intermodal Freight Transport

Research: a Review. Eur. J. Oper. Res., Vol. 153 (2004), 400-416.

9. Miles, R., Hamilton, K.: Learning UML 2.0. O’Reilly Media, Sabastopol CA USA (2006).

10. OMG: UML Resource Page. Available at http://www.uml.org/ (2009).

11. Ramstedt, L., Woxeni,us J.: Modelling Approaches to Operational Decision-Making in

Freight Transport Chains. Proc. 18th NOFOMA Conf. (2006).

12. Verbraeck, A., Versteegt, C.: Logistic Control for Fully Automated Large-Scale Freight

Transport Systems. Proc. IEEE Conf. Intelligent Transportation Systems (2001), 770-775.

13. Xu, J., Hancock, K.L.: Enterprise-Wide Freight Simulation in an Integrated Logistics and

Transportation System. IEEE Trans. Int. Transp. Sys., Vol. 5 (2004), 342-346.

14. Yun, W.Y., Choi, Y.S.: A Simulation Model for Container-Terminal Operation Analysis

Using an Object-Oriented Approach. Int. J. Prod. Econ., Vol. 59 (1999), 221-230.

105