business process project management

Evaluating the impact of smart technologies on harbor’s logistics via BPMN modeling and simulation

Mario G. C. A. Cimino1 • Filippo Palumbo2,3 • Gigliola Vaglini1 • Erina Ferro2 •

Nedo Celandroni2 • Davide La Rosa2

Published online: 1 October 2016

� Springer Science+Business Media New York 2016

Abstract A smart Information and Communication

Technology (ICT) enables a synchronized interplay of

different key factors, aligning infrastructures, consumers,

and governmental policy-making needs. In the harbor’s

logistics context, smart ICT has been driving a multi-year

wave of growth. Although there is a standalone value in the

technological innovation of a task, the impact of a new

smart technology is unknown without quantitative analysis

methods on the end-to-end process. In this paper, we first

present a review of the smart ICT for marine container

terminals, and then we propose to evaluate the impact of

such smart ICT via business process model and notation

(BPMN) modeling and simulation. The proposed approach

is discussed in a real-world modeling and simulation

analysis, made on a pilot terminal of the Port of Leghorn

(Italy).

Keywords Smart harbors � Wireless sensor network � RFID � BPMN � Workflow modeling � Workflow simulation

1 Introduction and motivation

Logistics and freight transport are nowadays key factors of

competitive advantage, due to undergoing significant inno-

vation in the smart Information and Communication Tech-

nologies (ICT). These developments made possible the

emergence of new design paradigms in logistic systems

based on the integration of different aspects, such as opera-

tion, energy consumption, environmental performance, and

so on [3, 30, 60]. An example is given by the European

project SuperGreen [51], aimed at supporting the definition

and benchmarking of green freight corridors through Europe

with respect to environmental, technical, economic, social,

and spatial planning aspects. This has been achieved by

applying methodologies for the assessment and bench-

marking of corridors with smart ICT [28]. With the large-

scale integration of smart ICT, new models of organization,

planning, and management become possible. In essence, a

modern port is characterized by its containers traffic and its

incorporation in a logistics network, where land and sea

segments are integrated. The efficiency of a port is strongly

influenced by its ability to forge links with the hinterland in

order to let the goods quickly arrive at destination. Moreover,

the European Union encourages with several initiatives the

speeding, the safety, and the streamlining of the maritime

transport with the other transport modes.

A new phase in ICT integration started in the 2000s,

with the emergence of Web Services based on Extensible

Markup Language (XML), together with the Service-Ori-

ented Architecture (SOA) paradigm [21]. In parallel, the

concept of networked smart devices evolved, due to a

convergence of multiple technologies, ranging from wire-

less communications to the Internet and from embedded

systems to sensing systems, yielding the vision of the

Internet of Things (IoT) [5].

& Mario G. C. A. Cimino [email protected]

1 Department of Information Engineering, University of Pisa,

Largo L. Lazzarino 1, 56122 Pisa, Italy

2 Institute of Information Science and Technologies, National

Research Council, via G. Moruzzi 1, 56124 Pisa, Italy

3 Department of Computer Science, University of Pisa, Largo

B. Pontecorvo 3, 56127 Pisa, Italy

123

Inf Technol Manag (2017) 18:223–239

DOI 10.1007/s10799-016-0266-4

In the last decade, the integration logic has become an

important issue in the development of smart ICT enabled

systems, due to the growing complexity of the ICT in

logistics and the increasing demand for adapting systems to

new requirements [22]. To assess the impact of smart ICT

in the port logistics, it is necessary to firstly define the

specific processes of the considered port and then to dis-

cuss how ICT is charging the tasks in the chain. In a tra-

ditional design paradigm [26]: (i) each stage of the chain is

considered as an independent activity; (ii) economies of

scale are key competitive differentiators; (iii) horizontal

integration is the main strategic option; (iv) efficiency

optimization is fragmented; (v) a considerable amount of

uncertainty exists in supply chain performance of other

parties; (vi) ICT is mainly used for single task operations.

In contrast, a smart ICT-based paradigm is characterized

by [12, 43]: (i) business process viewed as an integrated

chain of value-adding activities; (ii) reduction in the costs

for all parties; (iii) vertical cooperation versus adversarial

relationships between parties; (iv) reduction in the uncer-

tainty in supply chain performance; (v) ICT used also at the

service level (i..e, integration level) and at the handoff level

(i.e., interoperability level).

In essence, the design and implementation of the inte-

gration logic can be considered as a high-level organiza-

tion, where tasks supported by different smart ICT systems

represent the building blocks organized by the integration

logic itself. Indeed, the goal of a smart ICT is to support the

work organization and the collaboration among tasks and

resources (humans and machineries) for a number of rep-

resentative cases. For this purpose, it is not sufficient to

equip workers with adequate smart ICT for their individual

workplaces, but also to consider the relationships among

work activities that are performed by different workers and

to provide support for their collaboration, as well as to take

into account the responsibility for executing each step in

the flow of work. Process models provide the conceptual

basis for defining when and under which conditions tasks

are actually carried out in the context of an integration

scenario [61]. In this context, the Business Process Model

and Notation (BPMN) [42] is a standard of the Object

Management Group (OMG), with the primary goal of

providing a notation that is readily understandable by all

business stakeholders. BPMN provides support to represent

the most common control flow modeling requirements.

Other language proposals in the literature get an abstract

representation of business processes [20], but the key

aspect is that BPMN is supported by an executable model

to enact instances of processes on ICT platforms [44].

Since a BPMN model can automatically be translated into

an executable model, it can also be computer-simulated.

The main advantage of simulation-based analysis is that it

can predict process performance by using a number of

qualitative and quantitative measures [22]. As such, it

provides a way to evaluate the execution of the business

process over a number of cases in order to determine

inefficient or inconsistent behavior. Thus, BPMN modeling

and simulation can be interchangeably used as a basis for

making managerial or technical decisions. From one side,

models, when simulated, can even be more realistic than

traditional experiments as they allow the efficient obser-

vation of a number of cases. From the other side, simula-

tions, when made via BPMN models, allow combining and

capturing a number of workflow patterns, thus setting up a

coherent environment for integration of complex interac-

tion behavior [25].

In this paper, we present an approach to measure the

improvements made possible by smart ICT on maritime

container terminals. A literature review of smart ICT for

harbors logistics is first provided. Second, the adoption of

the BPMN standard to model and simulate a terminal

process is discussed. Third, a marine container terminal of

the Port of Leghorn (Italy) is modeled and simulated via

the BPMN, in order to measure the impact of the appli-

cation of RFID and Wireless Sensor Network (WSN) in the

terminal. Simulation results show that significant saving on

both processing time and resources can be achieved. The

proposed approach is independent of the terminal and the

ICT solutions considered. This study comprises significant

outcomes of a Research Project of National Interest (PRIN)

founded by the Italian Ministry of Education University

and Research (MIUR), in which we carried out a feasibility

study.

The paper is structured as follows. Section 2 covers the

integration and sensing infrastructures available to the

smart harbor’s logistics, giving an insight of the state-of-

the-art in the application of ICTs to logistics and moni-

toring of containers. Core concepts of BPMN modeling,

together with the pilot case study are presented in Sect. 3.

Section 4 is devoted to fundamentals of BPMN simulation.

Experimental studies are described in Sect. 5. Finally,

Sect. 6 draws some conclusions and future works.

2 Integration infrastructures available to smart harbor’s logistics: literature review

The application of smart ICT helps to address several

issues in the management of maritime and harbor’s logis-

tics. The use of vertical solutions, involving the use of

sensors and actuators to control part of the port infras-

tructure, is acknowledged since many years. An example is

the Dover’s Smart Bridge [58], dated back to the ’90,

where lift bridges are enhanced with an automatic control

system to sense the vessel movements and to adjust the

ship-to-shore bridge shape in order to maintain the rail

224 Inf Technol Manag (2017) 18:223–239

123

stock transit. The work in [57] shows examples in which

port logistics has experienced benefits from investing in

smart yard handling technologies, such as: Double Rail

Mounted Gantry cranes or DRMG (Hamburger Hafen und

Logistik AG), semi-automated DRMG (Ningbo Beilun

Container Port), automated crane handling with remote

controlling back-office (Shanghai Waigaoqiao Port) and

fully automated straddle carrier system (Brisbane). Auto-

mated Guided Vehicles (AGVs), as well as Automated

Stacking Cranes (ASCs), can be used to increase the con-

tainer movements’ efficiency. AGVs are robotic vehicles

that travel along a predefined path defined by electric wires

embedded in the ground or a grid of transponders. ASCs

move on rails and are controlled by a central operating

system. Due to their minimal footprint, they can provide

high density container storage. Some terminals of the

Rotterdam Port use both AGVs and ASCs technolo-

gies [59]. Sensors have also been deployed for increasing

the security of critical port areas such as gas and oil ter-

minals or anchored naval vessels [36], both on the surface

or underwater. Waterside port security includes a range of

activities with preventive purposes, ultimately aiming at

controlling who and what enters into the port area from the

water side. It includes crews, passengers, and cargo

entering on-board of large announced vessels, on-board of

small unannounced surface crafts, swimmers, divers or

even small submersibles. The most challenging function-

ality of an autonomous surveillance system resides in the

automatic detection of ‘‘abnormal’’ events to call for

prompt specific operator attention. This requires features

extraction, recognition, and correlation functionalities [27].

When we focus on smart ICT-based horizontal solu-

tions, involving the use of different technologies in diverse

parts of the port environment, we can find works focusing

on: optimizing the design and operation of container ter-

minals [38, 53, 57, 59]; enhancing inland and maritime

transportation systems [11, 18, 24]; providing real time

locating systems [4, 39]. In order to lower the shipment

time and to enhance the productivity of container port

logistics, also the port management is now advancing to

smart ICT-based integration infrastructures focusing, in

particular, on cloud-based solutions [31, 32]. This provides

critical functionalities by employing ubiquitous computing

technologies that allow achieving real-time synchroniza-

tion of port logistics.

These functionalities can be summarized in two groups:

(1) localization, tracking, and identification of objects, and

(2) management of cargo freight, such as sea containers

transporting goods to be monitored in order to minimize

possible economic losses. For this reason, we identified two

main areas where smart ICT solutions can boost the per-

formance of harbor’s logistics: (1) localization, tracking,

and identification, and (2) smart containers.

These two key application scenarios will be investigated

in the next subsections in terms of hardware and commu-

nication technologies involved.

2.1 Localization, tracking, and identification

In the field of outdoor localization, existing solutions based

on GPS, sometimes with the support of traditional wireless

networks, are de-facto consolidated as standards. Never-

theless, when dealing with harbor’s scenarios, several

additional issues must be addressed. For example, with

GPS-based solutions it is possible to track the position of

cranes moving containers by means of centralized geo-

databases. This solution is generally effective but there are

limits due to the fact that containers are not always moved

by means of cranes, but also by trucks and tractors. Fur-

thermore, GPS systems alone are not enough reliable

solutions for tracking yard trailers because of a lot of dead

zones caused by huge quay cranes and container stacks. An

alternative solution to GPS-based systems is the use of

Real-Time Locating Systems (RTLSs), which give in real

time the exact position of containers when a RTLS tag is

attached to them [17]. These systems are characterized by

the use of techniques typical for indoor localization sce-

narios. Usually, the indoor localization process starts from

measuring distances between anchors, whose location is

predetermined, and mobile nodes. There are attractive

solutions for estimating the distance between nodes; among

them, the Received Signal Strength Indicator (RSSI) is the

predominant approach [49]. It is based on the radio path

loss model, in which radio signals exponentially attenuate

during transmission. In this field, different technologies can

be used, spanning from WiFi [8, 41], Bluetooth [45], and

WSN [33] to ultrasonic sounds [50], Chirp Spread Spec-

trum (CSS) [54], and Ultra Wide Band (UWB) [19].

A well-established technology for RTLS in the harbor’s

logistic scenario is the Radio-Frequency IDentification

(RFID) [15, 17]. RFID systems can provide a less accurate

positioning information, usually regarding the relative

positioning of objects, but they offer the added possibility

to number, identify, catalog, and track objects. These

characteristics enable container terminals to be managed in

a more efficient way thanks to a quick identification of the

containers, but they are less useful to determine their

position. Furthermore, RFID systems require a fixed or

mobile infrastructure to read the tags, and the process, in

many cases, includes human-driven or semi-automated

operations. Thus, with current solutions, real-time identi-

fication and localization of containers are error-prone

activities that still require human intervention to manage

anomalous situations (for example, by physically searching

the containers that are out of place). In order to overcome

these limitations, several works have been proposed to add

Inf Technol Manag (2017) 18:223–239 225

123

tracking and tracing capabilities to localization techniques

by using wireless sensor networks [14, 35, 62]. WSNs have

proven to be useful in different scenarios, from human

activity recognition [47, 48] to ambient assisted living [46]

as well as for general purpose indoor localization sys-

tems [6, 9]. In [2], authors propose an innovative approach

where the position of containers can be continuously

determined by means of a wireless sensor network. Each

container is equipped with a number of nodes that use

wireless communication to detect neighbor containers. At

the base station, geometrical constraints and proximity data

are combined together to determine the relative positions of

containers. An interesting hybrid RFID-WSN approach is

recently emerging, which puts together the key aspects of

both the technologies. In [10], authors propose a new

hardware platform that enables an active RFID tag, thought

for monitoring temperature in goods, to communicate as a

wireless sensor node with single- and multi-hop routing. In

[16], authors enhance RFID tags with extended commu-

nication capabilities over a ZigBee network. In [63], a way

to integrate RFID with WSNs is presented in order to add

to the typical identification functionality of RFID the

localization possibilities enabled by a WSN.

Following the clear indications provided by the litera-

ture review, we choose as booster factors in our simulation

model the presence of RFID- and WSN-based RTLSs for

localizing, tracking, and identifying objects like containers

in the considered scenario.

2.2 Smart containers

The progress of machine-to-machine (M2M) communica-

tion technologies and wireless sensor networks offers a

differentiating factor for logistics companies. As discussed

in the previous section, not only it is possible to locate and

track a package from origin to destination, but, thanks to

WSNs, companies and port authorities can take advantages

from monitoring the transportation conditions throughout

the container’s journey. If there was an excess of moisture

in the container, if goods were opened or inspected along

the line, if there were temperature fluctuations, it is pos-

sible to know when, where, and how these events occurred.

This can be made by embedding wireless sensors nodes in

the container, providing useful different sensing capabili-

ties. Their wireless communication capacities, autonomous

power, and small sizes allow the remote monitoring of

goods through the Internet to be maintained with less

human effort. Moreover, in cases of container falls, fires,

exposure to floods or other risks, sensors (e.g. Waspmotes 1 )

can send SMS alerts to the customer, the transportation

company, or the law enforcement to call for immediate

assistance. This makes a container ‘‘smart’’.

We consider as main goals for a smart container three

important features enabled by the presence of a WSN: (i) to

detect unexpected container openings, (ii) to monitor

transport conditions, and (iii) to identify storage incom-

patibilities. Sensors (i.e. light, magnetic contacts, temper-

ature) can be placed within a container to determine when

it was opened [23]. They can be programmed to

acknowledge estimated opening hours and to check if

opening times correspond to scheduled inspections, gen-

erating alerts by GPRS/3G. In some cases, containers carry

humidity and temperature-sensitive items such as food,

pharmaceuticals, or artworks. Adding sensors to measure

these environmental variables can be essential to ensure

that goods are managed and unspoiled during the trans-

portation process [34]. If goods are fragile, registering

shock and vibration impacts can assist in identifying

responsible authorities in the case of insurance claims. In

this case, 3-axis accelerometers can be embedded in sensor

nodes to detect such vibrations. Finally, the motes can act

as smart tags. Beyond the passive behavior of identifying

the content of what is being freighted, sensors can actively

exchange information with other pallets or containers

stored around by using RFID and Near Field Communi-

cation (NFC) technologies. This way, warning messages

can be generated if, for example, a pallet of dangerous

goods is placed side by side with flammable materials [52].

3 BPMN modeling: core concepts and pilot case study

The BPMN language has been developed with a solid

mathematical foundation provided by the process calculus

theory, which is an essential requirement to automate

execution and to easily provide proofs of general consis-

tency properties. To describe a workflow, BPMN offers the

business process diagram, with a rich set of elements and

attributes. For the sake of significance, in this paper we

report on the basic elements shown in Fig. 1.

The interested reader may refer to [42] for a detailed

study of the language. More precisely, events (represented

as circles) model something that can happen during the

process. A workflow is activated by a start event (a circle

with a single thin border) and terminated by an end event (a

circle with a single thick border), while intermediate events

(circles with double border) can occur anywhere within the

flow. Tasks (rounded-corner rectangles) are atomic activi-

ties of the workflow, whereas gateways (diamonds) are

decision points to control the flow of work. The exclusive

gateway routes the incoming flow to one of the mutually1 http://www.libelium.com/products/waspmote.

226 Inf Technol Manag (2017) 18:223–239

123

exclusive outcoming flows, on the basis of a logic condi-

tion. The sequence flow is represented by a solid arrow, and

it models the order of execution of activities in the work-

flow. Finally, pools and lanes are represented by rectangles

and they model different responsible subjects/areas.

Given the above elements, an essential BPMN model of

a marine container terminal system is presented in the

following subsection. The model is related to a terminal

located in the Port of Leghorn (Italy) and it takes into

account some scenarios.

3.1 Workflow modeling of the pilot marine

container terminal

In this section, the emphasis is put on internal logistics,

since the aim of the study is to assess the extent to which

smart ICT solutions can improve the overall efficiency of

the process [55, 57]. A marine container terminal is the

place where containers arriving by sea vessels are trans-

ferred to inland carriers, such as trucks, trains, and vice

versa. Each marine container terminal performs four basic

functions: receiving, storage, staging, and loading for both

import and export. Receiving involves container arrival at

the terminal, either as an import or export, recording its

arrival, retrieving relevant logistics data and adding it to

the current inventory. Storage is the function of placing the

container in a known and recorded location in order to

retrieve it when needed. Staging is the function of

preparing a container to leave the terminal. The containers

that are to be exported are identified and organized so as to

optimize the loading process. Import containers follow

similar processes, although staging is not always per-

formed. An exception is a group of containers leaving the

terminal via rail. Finally, the loading function involves

placing the correct container on the ship, truck or other

mode of transportation.

Figure 2 represents the layout and the resources of a

container terminal system. More precisely, the berth (a

space for a vessel to anchor) is equipped with quay cranes

to unload containers. Unloaded containers are first trans-

ported to yard positions (the storage area), usually struc-

tured into stacks and differentiated into sub-areas for

export, import, special, and empty containers. The trans-

port between quay and yard can be performed either by

trucks with trailers, straddle carriers (SC), and automatic

guided vehicles (AGV). The formers can also serve the

landside operation, where containers departing or arriving

by road or railway are handled within the truck-and-train

areas.

Figure 3 shows a BPMN model of the container termi-

nal system. The model is based on 8 main lanes, 24 tasks, 7

gateways and 6 types of resources (different types of

machinery). After the arrival at the roadstead, vessels are

berthed according to a priority assigned via commercial,

security, and traffic management policies. Non-priority

vessels enter the roadstead and lie at anchor there, whereas

priority vessels directly enter into the harbor. The vessel is

then assigned to a berth, moved via berthing tugs and

finally moored. Unloaded containers are transported to the

quay. Here, a container can be placed on the top of a stack

(accessible location) or under other containers of the stack

(inaccessible location). The former is usually performed for

short storage, whereas the latter for medium-long storage.

Since there is no sufficient information to exactly establish

the storage duration, sometimes a number of movements

are required for a container before picking-up it. Each

movement may place the container to a next location (ac-

cessible or inaccessible). Once picked-up, the container is

moved, if needed, to a de-consolidation area (where mul-

tiple shipments from various suppliers are unpacked for

delivery) via a trailer, where it is consolidated. If consol-

idation is not needed, the container is directly moved to the

train or truck area, where it is loaded and checked out in the

train or truck gate out, respectively.

The handling machinery employed by the terminal

systems are: (i) Portainer (PT), a large dockside gantry

crane for loading and unloading containers from ships; (ii)

Rubber Tyred Gantry (RTG), a mobile gantry crane run-

ning on rubber tires to ground or stack containers; (iii) Rail

Mounted Grantry (RMG), a mobile gantry crane running

on rails; (iv) Reach Stacker (RS), to pile the containers;

(v) Trailer (TR), to move containers from a place to

another one; (vi) Berthing Tugs (BT), to move the vessel

into the harbor.

For each activity, Table 1 shows the area, the needed

resources, the duration interval, and the estimated duration,

derived by a number of interviews and measurements. In

this paper, the focus is on the activities that can be

improved by using smart ICT.

Start Intermediate End Task

N am

e N am

e N am

e

Pool with two lanes

X Exclusive Gateway Sequence Flow

Fig. 1 BPMN basic elements

Inf Technol Manag (2017) 18:223–239 227

123

4 BPMN simulation

In this Section, simulation functionality is first defined and

illustrated on a basic BPMN model. Then, simulation is

used to evaluate the impact of selected smart ICT on the

BPMN model of Fig. 3. Simulations have been carried out

using a specific simulation tool, namely Visual Paradigm

Logizian. 2

However, it is worth noting that the presented

approach is independent of the BPMN simulation tool. The

next subsection is devoted to the analysis of the BPMN

simulation tools. The interested reader may refer to [22] for

a comprehensive study on BPMN simulators.

4.1 State of the art of BPMN simulation tools

Since to conduct simulative experiments is an interactive

activity, an important choice of our approach is the simu-

lation tool. Today, most Business Process Management

(BPM) systems provide simulation facilities. A modern

simulation tool should provide building blocks for a certain

application area, to support the composition of a simulation

model via a visual notation, as well as a scripting language

to model complex behavior. However, scripting languages

RESOURCES

LAYOUT

SEASIDE

QUAY CRANE

VEHICLES STACK WITH RMG

VEHICLES

TRAIN TRUCKS

BERTH

V E

S S

E L

LANDSIDE

Y A

R D

(CONSOLIDATION AREA)

TRUCK GATE OUTTRAIN

GATE OUT

(TRAIN AREA)

(TRUCK AREA)

Q U

A Y

(R O

A D

S TE

A D

)

Fig. 2 Resources and layout of the pilot container terminal system

2 www.visual-paradigm.com/features/process-simulation/.

228 Inf Technol Manag (2017) 18:223–239

123

force to chart the situation in terms of a programming

language, make modeling time-consuming and the simu-

lation program itself provides no insights. The best tool

combines a visual design environment and a scripting

language, to offer graphical analysis capabilities and ani-

mation. The interested reader is referred to [22] for a

summary of the available business process simulation

tools. A negative feature of a simulation tool is the use of

proprietary building blocks, which makes it hard to inter-

change simulation models between packages. Simulation

tools based on more widely used languages, such Petri Nets

or BPMN, are more open and can exchange process models

M A

R IN

E C

O N

T A

IN E

R T

E R

M IN

A L

S E

A S

ID E

VESSEL ARRIVAL

PRIORITY VESSEL? VESSEL

LYING

NONO

TUG VESSEL

YESYES

MOOR VESSEL

UNLOAD CONTAINER

TRANSFER CONTAINER

PLACE CONTAINER

[ACCESSIBLE] YESYES

PLACE CONTAINER

[INACCESSIBLE]

NONO SHORT

STORAGE

STORAGE

REQUIRED MOVEMENT? REQUIRED

MOVEMENT?

LOCATE CONTAINER

ESYES

PICK UP

NONO

LOAD ON TRAILER

REQUIRED CONSOLIDATION?

TRANSFER CONTAINER

YESYES

PUT ON THE GROUND

(DE-) CONSOLIDATION

PICK-UP FROM THE GROUND

LOAD ON TRAILER

TRANSFER CONTAINER

BY TRAIN/TRUCK?

LA N

D S

ID E

T R

A IN

A R

E A

T R

A IN

G A

T E

O U

T

T R

U C

K A

R E

A T

R U

C K

G A

T E

O U

T

TRANSFER CONTAINER

TRAINTRAIN

TRANSFER CONTAINER

TRUCKTRUCK

LOAD ON TRAIN

CHECK OUT

LOAD ON TRUCK

CHECK OUT

NONO

Y A

R D

R O

A D

S T

E A

D Q

U A

Y

PLACE ON ACCESSIBLE POSITION?

(D E

-) C

O N

S O

LI D

A T

IO N

A R

E A

Fig. 3 BPMN model of a container terminal system

Inf Technol Manag (2017) 18:223–239 229

123

with different analysis tools [1]. BPMN, EPC (Event-dri-

ven Process Chain), UML AD (UML Activity Diagrams),

and other business process modeling notations have in

common that they all use token-based semantics. There-

fore, there are many techniques and tools to convert Petri

Nets to such languages, and vice versa. As a result, the core

concepts of Petri nets are often used indirectly, to enable

analysis, to enact models, and to clarify semantics. How-

ever, Petri Nets imply a severe representational bias, which

is relevant for the understandability of the results and vital

to guide process modeling and simulation via the majority

of the involved actors. In contrast, BPMN 2.0 is the de

facto standard notation for modeling business processes

understandable by a wide audience of people. An absolute

majority of freeware and commercial BPM tools and

Business Suites, like Oracle BPM Suite, IBM Business

Process Manager, jBPM, Activiti, Appian BPM Suite,

Bizagi BPM Suite, MagicDraw Enterprise Architect

(Sparx), Mega Process (MEGA), Signavio Process Editor

and others, either natively support BPMN or provide con-

version in order to stay compatible and up to date. In the

literature, the modern simulation studies in the context of

port logistics are based on BPMN [7, 13, 37, 40, 56].

4.2 Fundamentals of BPMN simulation

In order to define the BPMN simulation, we introduce the

concept of token traversing the sequence flow and passing

through the elements in the process [22]. Figure 4a shows

a token as a gray circle, in a basic BPMN model. Every

time a new process occurs, the start event creates a token

(1). The exclusive gateway takes the incoming token and,

according to a given condition, decides to which sequence

flow it would be routed (2). An activity receives a token

and forwards it after completion (3). The join exclusive

gateway just takes an incoming token (4) and moves it to

the outgoing sequence flow (5). Finally, the end event

removes the token (6). It should be noted that the token

does not carry any information, other than a unique iden-

tifier of the process occurrence. Figure 4b and c show the

two scenarios of the BPMN model, respectively.

Since many process instances may be generated, many

tokens can be created during time. When the needed

resources are assigned to a task, the task is said to be

started. If the task gets completed without interruptions,

the task is said to be completed. Tokens can also be pro-

cessed in parallel by different tasks when sufficient

Table 1 Resources employed and duration of each activity in the pilot scenario

Activity Area Needed resources Duration interval Estimated duration

Tug vessel Quay Berthing tugs [20 m, 40 m] 30 m

Moor vessel Quay Berthing tugs 10 m 10 m

Unload container Quay Portrainer [3 m, 5 m] 4 m

Transfer container Quay Trailer [5 m, 8 m] 6.5 m

Place container (accessible) Yard Rubber tyred gantry 2 m 2 m

Place container (inaccessible) Yard Rubber tyred gantry [4 m, 14 m] 9 m

Short storage Yard – 1 d 12 h

Storage Yard – [2 d, 3 d] 2.5 d

Locate container Yard Yard staff [10 m, 20 m] 15 m

Pick up Yard Rubber tyred gantry 2 m 2 m

Load on trailer Yard Rubber tyred gantry 2 m 2 m

Transfer container (De-)consolidation Trailer [5 m, 8 m] 6.5 m

(de-)consolidation

Put on the ground (De-)consolidation Reach stacker trailer 2 m 2 m

(de-)consolidation

(De-)consolidation (De-)consolidation Consolidation staff 20 m 20 m

Pick-up from the ground (De-)consolidation RTG on consolidation 2 m 2 m

Load on trailer (De-)consolidation RTG on consolidation 2 m 2 m

Transfer container Yard RTG on consolidation [5 m, 8 m] 6.5 m

Transfer container from Yard

to train/truck area

Train/truck area Trailer 8 m 8 m

Load on train/truck Train/truck area Reach stacker (truck) 4 m 4 m

Rail mounted grantry (train)

Check out Train/truck gate out Gate out staff 15 m 15 m

230 Inf Technol Manag (2017) 18:223–239

123

resources are available. Otherwise, tokens will be pro-

cessed in series and queued at the input of tasks. The

workflow status is determined by the position of all

available tokens.

A simulation is generally made of a number of com-

bined scenarios, whose tokens are competing for resources.

Each scenario is characterized by a well-defined path (from

a start event to an end event), a number of process

occurrences (tokens) and their arrival rate. An example of

typical simulation question is ‘‘How long will it take to

process?’’. To answer this question, several variables need

to be declared: the duration of each task, the branching

proportion of each outgoing flow of each gateway, the

resources needed by each task, and the available resources.

Cost and other quality parameters can also be defined. Cost

can include the variable cost related to the duration (e.g.

hourly wages of the involved human resources) as well as a

fixed additional cost (e.g. shipping cost). During simula-

tion, the simulator keeps track of the time each process

instance spends in an activity and the time each resource

assigns to that activity. Hence, it provides a realistic way

for measuring and analyzing the actual costs of the

activities.

Basically, a workflow simulator takes the workflow

model and the above mentioned additional information as

an input, and provides the Key Performance Indicators

(KPIs) as an output. The simulation also provides anima-

tion, showing the tokens’ position, the task status, and the

queues’ size, when simulation takes place, to help under-

standing relevant phenomena in an interactive manner. As

an example, Fig. 5 shows some qualitative and quantitative

results of a simulation of the basic BPMN model of Fig. 4,

with the following additional simulation information: task1

duration: 60 min; task2 duration: 30 min; pool available

instances: 4; number of arriving tokens: 100; inter-arrival

time: 0 min; branching proportions: 40 % (scenario 1),

60 % (scenario 2). More precisely, Fig. 5a shows the status

of the workflow after 11 h of simulated time. Here, the

number and the position of tokens waiting at the input of

tasks are represented by an overturned triangle: 10 and 31

waiting tokens for task1 and task2, respectively. The

number of processing tokens for each task is represented by

a gear: 3 and 1 for task1 and task2, respectively. Indeed,

since the overall number of available instances of the pool

is 4, the workflow can only process 4 tokens in parallel.

The other 41 tokens are then queued. In this basic model,

the pool is the unique resource type. In general, other

resource types can be defined and associated to a pool, with

a quantity for each type. In this way, each task can be

associated to a needed quantity of the available resource

types. In Fig. 5a, the workflow is overall handling 13 ? 32

tokens, belonging to scenario 1 ? scenario 2, respectively.

Since scenario 1 and scenario 2 are supplied with 40 and 60

total tokens (due to the branching proportion), 27 and 28

tokens were already handled by the respective scenarios.

task1

task2

1

2 3 4

5 6

(a)

(b) (c)

Fig. 4 Simple simulation scenarios of a basic BPMN model. a Routing of a token in a basic BPMN model. b Scenario 1. c Scenario 2

Fig. 5 Qualitative and quantitative result of a simulation run. a Ongoing simulation. b Completed tokens and resource usage against time. c Queue time

Inf Technol Manag (2017) 18:223–239 231

123

All the tokens were processed in 17 days and 30 h of

simulated time. Figure 5b shows the number of tokens

completed in time, for each scenario. Here, the dark and

light gray curves are related to the scenario 1 and 2,

respectively. Both scenarios are characterized by the same

linear trend, which means that the number of completed

tokens per time unit is the same. For this purpose, scenario

1 used, on average, more resources than scenario 2, since

task 1 duration is higher than task 2 duration. More pre-

cisely, after 11 and 15.5 h each scenario completed about

28 and 40 tokens, respectively. The former instant of time

corresponds to the status illustrated in Fig. 5a, whereas the

second instant of time corresponds to the end of the sce-

nario 1. At the end of scenario 1, all resources were

available to scenario 2, and then its linear trend suddenly

increased, thus handling 20 tokens in about two hours.

Figure 5b also shows the resource usage against time. It

can be observed that the four available pools are fully used

all the time (100 %), because it is a general-purpose

resource. In other circumstances, different types of

resources are constrained to a subset of tasks; then, their

use is determined by the availability of tokens at the entry

of such tasks. Finally, Fig. 5c shows the queue time at each

task, which is important to determine the internal efficiency

of the workflow. The queuing on task 2 is higher because,

as discussed above, scenario 2 had fewer resources than

scenario 1, although task 1 duration is twice as long as task

2 duration.

The discussed example is representative of the features

of BPMN simulation function, although it is numerically

simple and then non-representative of its complexity. In

general, the behavior of BPMN models is highly non-lin-

ear, due to their structural complexity. Although the BPMN

has been founded on a mathematical model, analytic

solutions of a workflow are often impossible, too compli-

cated, and extremely expensive to validate. Moreover, it is

often impossible or extremely expensive to observe the

occurrence of different scenarios in the real world. For

these reasons, BPMN modeling and simulation are a valid

and fundamental approach to study the operation of a

workflow and to infer properties concerning the behavior of

an actual system.

With regard to the illustrated BPMN model of a con-

tainer terminal system, some additional simulation infor-

mation has been already presented in Table 1. In the

following subsection, other useful information for the

simulation is derived.

4.3 Considered scenarios and smart ICTs used

Our analysis is focused on the assessment of an important

KPI of a port’s performance from the point of view of the

exporter/importer: the dwell time of cargo in port,

measured in terms of the number of days that a given

amount of cargo remains in port after a peak in demand.

Considering the pilot container terminal with its available

resources, in order to determine a typical peak situation we

exploited the data log available in the Tuscan Port Com-

munity System 3

(tpcs.tpcs.eu). More precisely, the peak

scenario consists in the arrival of 5 vessels of 750 con-

tainers, almost at the same time. It is considered a critical

situation because the terminals layout does not allow the

processing of 5 vessels at the same time.

Figure 6 shows the smart ICT technologies experi-

mented, together with the application context. More pre-

cisely, the purpose of the simulation is to evaluate the

impact of RFID and WSN technologies on the overall flow

of work, when used for speeding the truck check-out and

the locate container tasks. The optimized version of a

business process (called to-be process) is carried out

starting from the current version (called as-is process).

Table 2 shows the duration of the two considered activities,

considering an as-is view (as in Table 1) and the to-be

version of the processes. Such data have been derived from

the data log of the information systems available at the

Truck Gate Out and at the Yard. More precisely, it can be

observed that the use of RFID sensibly speeds up the

check-out operations at the Truck gate-out area. Indeed, the

as-is version implies a number of manuals steps, such as

stopping the truck, delivering the hard copy of documents

to the gate officer, and waiting for the pass before restarting

the truck. In contrast, by using RFID technology the

activities are almost totally automated, since, when the

truck approaches the access gate, the RFID allows regis-

tering and controlling the container, thus allowing opening

the gate. Sometimes additional documents are needed, and

for this reason the average time spent is about 30 seconds.

Currently, this method is successfully working at different

gates of the Port of Leghorn. The use of the WSN tech-

nology for container localization has been previously pro-

posed by [2]. In essence, each container is equipped with a

number of nodes that use a WSN to detect neighbor con-

tainers. At the base station, geometrical constraints and

proximity data are combined together to determine the

relative positions of containers, thus speeding up their

localization. It can be observed that the adoption of WSNs

reduces the duration of the activity from 15 to about 1

minute. Indeed, the existing solution for localization and

identification is based on GPS and RFID technologies. GPS

enables the tracking of the position of the crane moving the

container. This solution is generally effective, but there are

3 The Tuscan Port Community System is a web-services based

information hub for the procedures of import and export of goods. It

enables the efficient exchange of relevant logistics information and

ensures the smooth flow of shipments from cargo origin to

destination.

232 Inf Technol Manag (2017) 18:223–239

123

also some limits, mainly due to the fact that containers are

not always moved by means of cranes, but also by trucks

and trailers. RFID solutions enable a quick identification of

containers, but they are less useful to determine their

position. Moreover, RFID systems require a fixed or

mobile infrastructure to read the tags, but the process is

usually human-driven. Thus, with current solutions, real-

time identification and localization of containers are error-

prone activities, requiring human intervention.

Table 3 shows the five considered simulation scenarios.

They basically differ in the number of movements needed

before going to the truck area.

5 Experimental studies

Simulation is handled over specific scenarios and by

incrementally changing the model parameters. More

specifically, the methodology is called what-if analysis, and

it consists in a data-intensive simulation activity whose

goal is to inspect the behavior of a part of the enterprise

business model under some given hypotheses called sce-

narios. In practice, the what-if analysis measures how

changes in a set of parameters impact on the process

performance with reference to the simulation model

offering an abstract representation of the significant fea-

tures of the business, and tuned according to the historical

enterprise data [29].

This Section is devoted to the simulations of the as-is

and to-be systems. Two types of business process

improvements have been carried out: (a) to reduce the

number of resources, keeping the process duration con-

stant, in order to establish whether or not some machines

can be removed, thus reducing the cost and the environ-

mental impact of the process; (b) to reduce the duration of

the process by adopting new smart ICT, for a better effi-

ciency of the process. In the following, the simulation

details and the obtained results for the two types of

improvements are described.

5.1 Simulation of the as-is model with full resources

Table 4 shows the quantity available of each resource in

the pilot container terminal. The first simulation is made by

using all the available resources. As a result, the total

duration for processing the 5 vessels of 750 containers is

7 days, 15 h and 53 min. Since this temporal duration is

calculated without any improvement initiative, it will be

used as a baseline for measuring progresses. Moreover, the

resource usage against time has revealed that the resources

PT, TR, and RTG are fully used in some intervals of the

simulation. Thus, reducing them implies an increase of the

total simulation duration. In contrast, the RS resource has a

maximum usage of 1 unit, which means that it is possible

to load truck one-by-one, since the flow of tokens is suf-

ficiently sparse at the end of the workflow. Thus, the next

(a) (b)

Fig. 6 The smart ICTs used during the experimentation (The devices shown in figure are IRIS Motes. They are 2.4 GHz Mote modules

used for enabling low-power, wireless sensor networks. The technol-

ogy underlying the sensor network works on IEEE 802.15.4

compliant RF transceivers using 2.4 to 2.48 GHz band, a globally

compatible Industrial, Scientific, and Medical (ISM) band. It involves

the use of direct sequence spread spectrum radio which is resistant to

RF interference and provides inherent data security at a 250 Kbps data

rate.). Photography supplied courtesy of Leghorn Port Authority and

Paolo Barsocchi. a RFID installed on the truck at the gate-out. b WSN installation on the container

Table 2 Duration of the considered activities when using smart ICTs

Activity Duration

(as-is) (min)

Duration after

using smart ICT

(to-be) (min)

smart

ICT used

Truck check-out 15 0.5 RFID

Locate container 15 1 WSN

Inf Technol Manag (2017) 18:223–239 233

123

improvement will be based on reducing underexploited

resources of the as-is process.

5.2 Simulation of the as-is model with reduced

resources

Table 5 summarizes the simulation results of the as-is

model with reduced resources. In particular, simulation S01

shows the starting point with no resource reduction at all,

calculated in the previous section. Starting from simulation

S01 with the maximum of RS availability, i.e. 14, the total

duration is not affected by the progressive reduction in the

RS units up to 1, as shown by simulation S02. Here, the

down-arrow at the right of RS means that the type of

resource has been reduced with respect to the previous

simulation, whereas the left-right-arrow in the Variation

column means that no variation occurred in the total

duration. Significant simulations are shown in bold-

face. From the queue time generated by simulation S2 it is

possible to discover that (i) the locate container and (ii) the

truck check out activities are bottlenecks. Indeed, (i) all

yard staff is involved for a relevant amount of the total

duration; (ii) all check-out staff and the gate-out areas are

busy for a lot of time. By reducing the number of trailers,

as shown by simulations S03-S07, it can be observed that

the minimum number of needed trailers is 13. Indeed, in

simulations S05-S07 in the face of a reduction of TR, there

is a positive variation in the total duration. Moreover, by

reducing the number of RTG, it can also be observed by

simulations S08-S11 that the best RTG number is 3.

Finally, by reducing the PT resource in S12-S14, it results

that the best number is 6. Indeed, from animation it can be

Table 3 Considered simulation scenarios

Scenario Sequence of gateway conditions Sequence of choices

Zero movement 1) Priority vessel? Yes

2) Place on accessible position? No

3) Required movement? No

4) Required consolidation? No

5) By Train/truck? Truck

One movement 1) Priority vessel? Yes

2) Place on accessible position? Yes

3) Required movement? Yes

4) Place on accessible position? Yes

5) Required movement? No

6) Required consolidation? No

7) By Train/truck? Truck

Two movements 1) Priority vessel? Yes

2) Place on accessible position? Yes

3) Required movement? Yes

4) Place on accessible position? Yes

5) Required movement? Yes

6) Place on accessible position? Yes

7) Required movement? No

8) Required consolidation? No

9) By train/truck? Truck

Three movements 1) Priority vessel? Yes

2) Place on accessible position? Yes

3) Required movement? Yes

4) Place on accessible position? Yes

5) Required movement? Yes

6) Place on accessible position? Yes

7) Required movement? Yes

8) Place on accessible position? Yes

9) Required movement? No

10) Required consolidation? No

11) By train/truck? Truck

234 Inf Technol Manag (2017) 18:223–239

123

observed that in the first phase the bottleneck is at the

locate container task, and then varying RTG does not

produce any significant variation. It is worth noting that to

reduce the other machinery or the staff is not useful, since

they are already fully used. As a result, Table 6 summa-

rizes the resources and correspondent saved quantities in

the simulated scenarios with the as-is model with reduced

resources.

5.3 Simulations of the to-be model with full

resources

The first simulation is made by using the resources in

Table 4, and removing the gate out staff of the Truck area

caused by the use of RFID. It can be observed by

simulation S15 in Table 7 that the total time duration is

7 days, 6 h, and 50 min, which is lower than in the cases of

the as-is model (7 days 15 h and 53 min). This temporal

duration will be used as a baseline for measuring pro-

gresses. Similarly to the approach adopted in the as-is

model, the resource usage against time allows to choose the

next resource to reduce, thus providing a to-be model with

reduced resources.

5.4 Simulations of the to-be model with reduced

resources

The first step is to reduce the RS resource usage to 1 unit

(simulation S16 in Table 7). Subsequently, the resource

usage against time has shown that the number of TR can

also be reduced to 13 units (simulation S17). Surprisingly,

with respect to the as-is model, the number of RTG can be

further reduced to 2 units (simulations S18-S20). The rea-

son is that the locate container activity is sensibly faster,

and then there is a larger queue on the place container

(accessible) activity, which acts as a more powerful buffer.

Furthermore, by reducing the number of PT, the resulting

optimal value is 2 (simulations S21-S23), i.e., lower than in

the as-is model. We can also sensibly reduce the Yard staff

from 15 to 1 (simulations S24,S25) thanks to the faster

Table 4 Resources and quantity available for the pilot scenario

Resource Quantity

Portrainer (PT) 10

Rubber tyred gantry (RTG) on Yard 10

RTG on consolidation area 5

Reach stacker (RS) 14

Trailer (TR) waterside, train/truck areas 30

Trailer (TR) (de-)consolidation area 6

Rail mounted grantry (RMG) 3

Berthing tugs (BT) 12

(De-)consolidation staff 10

Train gate out staff 5

Truck gate out staff 6

Yard ctaff 15

Train gate out 3

Truck gate out 3

Table 5 Total durations of the as-is model with reduced

resources

Simulation Type of resource Availability Total duration (as-is) Variation

S01 RS 14 7 days 15 h 53 min

S02 RS # 1 7 days 15 h 53 min $ S03 TR # 27 7 days 15 h 53 min $ S04 TR # 13 7 days 15 h 53 min $ S05 TR # 12 7 days 16 h 07 min " S06 TR # 11 7 days 18 h 06 min " S07 TR # 10 7 days 19 h 35 min " S08 RTG # 5 7 days 15 h 53 min $ S09 RTG # 3 7 days 15 h 53 min $ S10 RTG # 2 7 days 15 h 55 min " S11 RTG # 1 7 D 16 h 10 min " S12 PT # 9 7 days 15 h 53 m $ S13 PT # 6 7 days 15 h 53 m $ S14 PT # 5 7 days 16 h 14 min "

Significant values are in boldface

# reduction, " increase, $ constant

Table 6 Resources and quantity saved in the simulated scenarios

Resource Quantity

Portrainer (PT) 10 ! 6 Rubber tyred gantry (RTG) on yard 10 ! 3 Reach stacker (RS) 14 ! 1 Trailer (TR) waterside, train/truck areas 30 ! 13

Inf Technol Manag (2017) 18:223–239 235

123

locate container activity made by means of WSNs. Finally,

since the truck gate out is very fast with RFID, we can

reduce the number of gate out from 2 to 1. This also

reduces the implementation costs, because only 1 RFID

reader is needed, in place of 3.

5.5 Simulations of the to-be model with reduced

time

In order to further reduce the total time duration, let us

consider the highest bottleneck, where some resources can

be added. The queue time reveals that the bottleneck on the

truck area is the most relevant. Here, we can increase up to

4 (simulations S28-S34 in Table 8) the number of truck

areas and RS available. This way, there is a significant

reduction in the total time duration. A further study of the

queue time reveals that the major queues are still located at

the truck area. However, there is no significant improve-

ment by increasing the size and the resources of the area

(simulation S35).

As summarized in Table 9, relevant economies result

by adopting RFIDs and WSNs. More specifically, con-

sidering Table 7: (i) the truck gate-out staff can be

removed, due to results of Simulation S15; (ii) the number

of RTG and PT can be reduced from the initial 10

(Table 4) to 5 and 2, respectively (Table 7, Simulation

S18 and S22) ; (iii) the number of RS and TR can be

diminished from 14 to 4 and from 30 to 13, respectively

(Table 7 Simulations S34 and S17). The number of Yard

staff can be reduced from 15 to 1, due to Simulations S24

and S25 of Table 7. Finally, the number of Truck gate out

and Truck are modified from 3 to 1 and from 1 to 4, due

to Simulations S27 and S30-S34.

6 Conclusions and future work

In this paper, an approach for evaluating the impact of

smart ICT technologies in the logistics has been discussed.

The approach is based on workflow modeling and simu-

lation. The main motivation comes from the intrinsic nat-

ure of smart ICT as enabler of synchronized interplay of

different key factors operating at workflow level, whose

integration logic is often impossible to be tackled and

validated with analytic solutions. The effectiveness of the

approach is founded on the BPMN language, which offers

a comprehensive technology for modeling patterns, as well

as standardized simulation engines. The executable char-

acter of BPMN also represents a bridge between the

workflow design/simulation and its implementation on a

service-oriented environment. The approach has been dis-

cussed and applied to a real-world analysis on a marine

container terminal of the Port of Leghorn. For this purpose,

a BPMN model of the terminal system has been provided,

together with a comprehensive survey of smart ICT for

harbor’s logistics. Finally, the impact of the application of

RFID and WSN has been measured by simulating the

operation of the modeled workflow, revealing significant

properties on the behavior of the actual system. This study

has been carried out in the framework of a national

research project founded by the Italian Ministry of the

Universities and the Research (MIUR), and has been cur-

rently focused on the hardware ICT technologies for har-

bor’s logistics. The adoption of the approach to software

ICT in the same application context is considered a key

investigation activity for future work.

Simulation results show that the adoption of the con-

sidered smart ICT allows achieving relevant potential

savings: (i) 53.2 % of saving on processing time; (ii) 67.7

Table 7 Total duration with the to-be model with resources

reduction

Simulation Type of resource Availability Total duration Variation

S15 Truck gate out staff # 0 7 days 06 h 50 min # S16 RS # 1 7 days 06 h 50 min $ S17 TR # 13 7 days 06 h 50 min $ S18 RTG # 5 7 days 06 h 50 min $ S19 RTG # 3 7 days 06 h 50 min " S20 RTG # 2 7 days 09 h 13 min " S21 PT # 6 7 days 06 h 50 min $ S22 PT # 2 7 days 06 h 50 min $ S23 PT # 1 7 days 12 h 55 min " S24 Yard staff # 15 7 days 06 h 50 min $ S25 Yard staff # 1 7 days 06 h 50 min $ S26 Truck gate out # 2 7 days 06 h 50 min $ S27 Truck gate out # 1 7 days 06 h 50 min $

Significant values are in boldface

# reduction, " increase, $ constant

236 Inf Technol Manag (2017) 18:223–239

123

% of saving on machinery (PT, RG, RS); (iii) 95.24 % of

saving on staff involved at the Truck Gate Out and Yard. In

terms of potential implementation hurdles: (i) the adoption

of RFID at the Truck Gates achieved a patent license

owned by a company participating to the project

(www.itpass.eu), which started to effectively implement a

number of installations on various gateways; (ii) the

adoption of WSN is currently handled within the overall

standardization of the Smart Container technology. Indeed,

since maritime containers move around the world on dif-

ferent ports, it is more convenient to include other sensors

in order to increase scalability, robustness, and

interoperability.

In our analysis, we do not measure the performance

variability because it is not significant with respect to the

average in a peak scenario. In other scenarios, an indication

of both the average and variability of the performance of

the process might be provided. Hence, task execution times

and process arrival rates can be defined by an average value

plus some distribution information. BPMN simulators can

incorporate statistical distributions to model non-deter-

ministic decision flows. However, the use of broad spec-

trum statistical simulation introduces relevant complexity,

increasing the chance of meaningless results for the busi-

ness analyst [21]. An interesting approach is to adopt the

interval-valued simulation in place of statistical-based

simulation: it allows an easier understanding of the process

by means of multi-valued representations [21]. For this

purpose, future work will investigate the use of interval-

valued simulation in the same context.

Acknowledgments Work carried out in the framework of the Italian PRIN Project ‘‘Eguaglianza Nei Diritti Fondamentali Nella Crisi

Dello Stato E Delle Finanze Pubbliche:Una Proposta Per Un Nuovo

Modello Di Coesione SocialeCon Specifico Riguardo Alla Liberal-

izzazione E RegolazioneDei Trasporti’’, activity ‘‘Sistema Di Moni-

toraggio A Distan-za Ed Automatizzato Delle Merci E Di Eventuali

PersonePreposte Alla Sorveglianza Delle Stesse’’.

References

1. van der Aalst WM (2015) Business process simulation survival

guide. In: Handbook on business process management, vol. 1.

Springer, pp 337–370

2. Abbate S, Avvenuti M, Corsini P, Vecchio A (2009) Localization

of shipping containers in ports and terminals using wireless

sensor networks. In: proceedings of international conference on

computational science and engineering, CSE’09, vol. 2. IEEE,

pp 587–592

3. Andalusian Institute of Technology: SMART-PORT: Action plan

towards the smart port concept in the Mediterranean area. http://

medmaritimeprojects.eu.dev10.tildecms.com/section/smartport.

Accessed 18 May 2015

4. Asosheh A, Afshinfar A, Kharrat M, Ramezani N (2008) A

network model for the intelligent marine container tracking. In:

8th WSEAS International conference on applied informatics and

communications (AIC08) Rhodes

5. Atzori L, Iera A, Morabito G (2010) The internet of things: a

survey. Comput Netw 54(15):2787–2805

6. Awad A, Frunzke T, Dressler F (2007) Adaptive distance esti-

mation and localization in WSN using RSSI measures. In: Pro-

ceedings of 10th euromicro conference on digital system design

architectures, methods and tools, DSD 2007. IEEE, pp 471–478

7. Badura D (2014) Modelling business processes in logistics with

the use of diagrams BPMN and UML. Forum Sci Oecon 2:35–50

8. Bahl P, Padmanabhan VN (2000) RADAR: an in-building rf-based

user location and tracking system. In: Proceedings of INFOCOM

2000. Nineteenth annual joint conference of the IEEE computer

and communications societies, vol 2. IEEE, pp 775–784

9. Barsocchi P, Lenzi S, Chessa S, Furfari F (2012) Automatic

virtual calibration of range-based indoor localization systems.

Wirel Commun Mob Comput 12(17):1546–1557

Table 8 Total duration with the to-be model with time reduction

Simulation Type of resource Availability Total duration Variation

S28 Truck area ", RS " 2, 1 4 days 22 h 31 min # S29 Truck area ", RS " 2, 2 4 days 15 h 12 min # S30 Truck area ", RS " 3, 2 3 days 23 h 38 min # S31 Truck area ", RS " 3, 3 3 days 22 h 28 min # S32 Truck area ", RS " 4, 2 3 days 14 h 35 min # S33 Truck area ", RS " 4, 3 3 days 14 h 16 min # S34 Truck area ", RS " 4, 4 3 days 14 h 08 min # S35 Truck area ", RS ", Gate out " 4, 5, 2 3 days 14 h 07 min #

Significant values are in boldface

# reduction, " increase, $ constant

Table 9 Resources and quantity saved in the simulated scenarios

Resource Quantity

Truck gate out staff 6 ! 0 Portrainer (PT) 10 ! 2 Rubber tyred gantry (RTG) on yard 10 ! 5 Reach stacker (RS) 14 ! 4 Trailer (TR) waterside, train/truck areas 30 ! 13 Yard staff 15 ! 1 Truck gate out 3 ! 1 Truck area 1 ! 4

Inf Technol Manag (2017) 18:223–239 237

123

10. Bilstrup U, Wiberg PA (2004) An architecture comparison

between a wireless sensor network and an active RFID system.

In: The 29th Annual IEEE international conference on local

computer networks, 2004. IEEE, pp 583–584

11. Bontekoning Y, Macharis C, Trip J (2004) Is a new applied

transportation research field emerging? A review of intermodal

rail-truck freight transport literature. Transport Res Part A: Policy

Pract 38(1):1–34

12. Boschian V, Fanti MP, Iacobellis G, Ukovich W (2012) Analysis

of impact of ICT solutions in international freight management.

Eur TransportnTrasporti Europei (51) 13. Caceres R, Mendoza H, Tuñón G, Rabelo LC, Pastrana J (2015)

Modeling and simulation of berthing processes for a Panamanian

container terminal using BPMN and discrete event simulation. In:

Proceedings of the 2015 international conference on operations

excellence and service engineering

14. Chantzis K, Chatzigiannakis I, Rolim J (2014) Design and eval-

uation of a real-time locating system for wireless sensor net-

works. J Locat Based Serv 8(2):97–122

15. Cho H, Choi H, Lee W, Jung Y, Baek Y (2006) LITeTag: design

and implementation of an RFID system for IT-based port logis-

tics. J Commun 1(4):48–57

16. Cho H, Kim J, Baek Y (2011) Large-scale active RFID system

utilizing ZigBee networks. IEEE Trans Consum Electron

57(2):379–385

17. Cho H, Kim T, Park Y, Baek Y (2012) Enhanced trajectory

estimation method for RTLS in port logistics environment. In:

proceedings of 2012 IEEE 14th international conference on high

performance computing and communication and 2012 IEEE 9th

international conference on embedded software and systems

(HPCC-ICESS). IEEE, pp 1555–1562

18. Christiansen M, Fagerholt K, Nygreen B, Ronen D (2007) Mar-

itime transportation. Handb Oper Res Manag Sci 14:189–284

19. Chung WC, Ha DS (2003) An accurate ultra wideband (UWB)

ranging for precision asset location. In: Proceedings of 2003

IEEE conference on ultra wideband systems and technologies.

IEEE, pp 389–393

20. Ciaramella A, Cimino MG, Lazzerini B, Marcelloni F (2009)

Using BPMN and tracing for rapid business process prototyping

environments. ICEIS 3:206–212

21. Cimino MG, Marcelloni F (2011) Autonomic tracing of pro-

duction processes with mobile and agent-based computing. Inf

Sci 181(5):935–953

22. Cimino MG, Vaglini G (2014) An interval-valued approach to

business process simulation based on genetic algorithms and the

BPMN. Information 5(2):319–356

23. Craddock R, Stansfield E (2005) Sensor fusion for smart con-

tainers. In: Signal processing solutions for homeland security,

2005. The IEE Seminar on (Ref. No. 2005/11108). IET, p 12

24. Crainic TG, Kim KH et al (2006) Intermodal transportation.

Transportation 14:467–537

25. Dumas M, Van der Aalst WM, Ter Hofstede AH (2005) Process-

aware information systems: bridging people and software through

process technology. Wiley, Hoboken

26. Evangelista P (2002) The role of ICT in the logistics integration

process of shipping lines. Pomorski zbornik 40(1):61–78

27. Garnier B, Andritsos F (2010) A port waterside security systemic

analysis. In: Waterside security conference (WSS), 2010 Inter-

national. IEEE, pp 1–6

28. Georgopoulou C, Kakalis NM, Psaraftis HN, Recagno V, Fozza

S, Zacharioudakis P, Eiband A (2014) Green technologies and

Smart ICT for sustainable freight transport. In: Efficiency and

innovation in logistics. Springer, pp 15–33

29. Golfarelli M, Rizzi S (2009) What-if simulation modeling in

business intelligence. Int J Data Warehous Min (IJDWM)

5(4):24–43

30. Grajek M (2012) ICT for growth: a targeted approach. Technical

Report, Bruegel Policy Contribution

31. Heilig L, Negenborn RR, Voß S (2015) Cloud-based intelligent

transportation systems using model predictive control. In: Com-

putational logistics. Springer, pp 464–477

32. Heilig L, Voß S (2014) A cloud-based SOA for enhancing

information exchange and decision support in ITT operations. In:

Computational logistics. Springer, pp 112–131

33. Hong SH, Kim BK, Eom DS (2009) Localization algorithm in

wireless sensor networks with network mobility. IEEE Trans

Consum Electron 55(4):1921–1928

34. Jedermann R, Pötsch T, Lang W (2014) Smart sensors for the

intelligent container. ITG-Fachbericht-Smart SysTech 2014

35. Jiang J, Guo, Y, Liao W, Li S, Xie X, Yuan L, Nian L (2014)

Research on RTLS-based coordinate guided vehicle (CGV) for

material distribution in discrete manufacturing workshop. In:

Proceedings of 2014 IEEE international conference on internet of

things (iThings), and green computing and communications

(GreenCom), IEEE and cyber, physical and social computing

(CPSCom). IEEE, pp 1–8

36. Kastek M, Dulski R, Zyczkowski M, Szustakowski M, Trza-

skawka P, Ciurapinski W, Grelowska G, Gloza I, Milewski S,

Listewnik K (2012) Multisensor system for the protection of

critical infrastructure of a seaport. In: SPIE defense, security, and

sensing. International Society for Optics and Photonics,

pp 83,880M–83,880M

37. Khalifa IH, El Kamel A, Yim P (2011) Transportation process of

containers BPMN-modelling and transformation into ACTIF

model. ROMJIST 14(1):67–80

38. Kim KH (2005) Models and methods for operations in port

container terminals. In: Logistics systems: design and optimiza-

tion. Springer, pp 213–243

39. Kim KH, Hong BH (2010) Maritime logistics and applications of

information technologies. In: Proceedings of 2010 40th interna-

tional conference on computers and industrial engineering (CIE).

IEEE, pp 1–6

40. Koniewski R, Dzielinski A, Amborski K (2006) Use of petri nets

and business processes management notation in modelling and

simulation of multimodal logistics chains. In: Proceedings 20th

European conference on modeling and simulation, Institute of

Control and Industrial Electronics, Warsaw University of

Technology

41. Lim CH, Wan Y, Ng BP, See C (2007) A real-time indoor WiFi

localization system utilizing smart antennas. IEEE Trans Consum

Electron 53(2):618–622

42. Object Management Group (OMG): business process model and

notation (BPMN), version 2.0. http://www.omg.org/spec/BPMN/

2.0 (2011). Accessed 26 May 2015

43. Obogne MH, LIDASAN HS (2005) A study on the impact of

information and communication technology on urban logistics

system: a case in Metro Manila. J East Asia Soc Transp Stud

6:3005–3021

44. Organization for the Advancement of Structured Information

Standards (OASIS): Web services business process execution

language (WS-BPEL), version 2.0. http://docs.oasisopen.org/

wsbpel/2.0/OS/wsbpel-v2.0-OS.html (2007). Accessed 26 May

2015

45. Palumbo F, Barsocchi P, Chessa S, Augusto JC (2015) A stig-

mergic approach to indoor localization using Bluetooth Low

Energy beacons. In: Proceedings of 2015 12th IEEE International

conference on advanced video and signal based surveillance

(AVSS). IEEE, pp 1–6

46. Palumbo F, Barsocchi P, Furfari F, Ferro E (2013) AAL middle-

ware infrastructure for green bed activity monitoring. J Sens 2013

47. Palumbo F, Barsocchi P, Gallicchio C, Chessa S, Micheli A

(2013) Multisensor data fusion for activity recognition based on

238 Inf Technol Manag (2017) 18:223–239

123

reservoir computing. In: Evaluating AAL systems through com-

petitive benchmarking. Springer, pp 24–35

48. Palumbo F, Gallicchio C, Pucci R, Micheli A (2016) Human

activity recognition using multisensor data fusion based on

reservoir computing. J Ambient Intell Smart Environ 8(2):87–107

49. Patwari N, Hero AO, Perkins M, Correal NS, O’dea RJ (2003)

Relative location estimation in wireless sensor networks. IEEE

Trans Signal Process 51(8):2137–2148

50. Priyantha NB, Chakraborty A, Balakrishnan H (2000) The cricket

location-support system. In: Proceedings of the 6th annual

international conference on Mobile computing and networking.

ACM, pp 32–43

51. Psaraftis HN, Panagakos G (2012) Green corridors in European

surface freight logistics and the SuperGreen project. Proc Soc

Behav Sci 48:1723–1732

52. Rezapour TY, Atani RE, Abolghasemi MS (2014) Secure posi-

tioning for shipping containers in ports and terminals using WSN.

In: Proceedings of 2014 11th international ISC conference on

information security and cryptology (ISCISC). IEEE, pp 10–14

53. Saanen YA (2004) An approach for designing robotized marine

container terminals. Delft University of Technology, Delft

54. Sikora A, Groza VF (2007) Fields tests for ranging and local-

ization with time-of-flight-measurements using chirp spread

spectrum rf-devices. In: Instrumentation and measurement tech-

nology conference proceedings, IMTC 2007. IEEE, pp 1–6

55. Stahlbock R, Voß S (2008) Operations research at container

terminals: a literature update. Or Spectr 30(1):1–52

56. Stajniak M, Guszczak B (2011) Analysis of logistics processes

according to BPMN methodology. In: Information technologies

in environmental engineering, pp 537–549

57. Steenken D, Voß S, Stahlbock R (2004) Container terminal

operation and operations research-a classification and literature

review. OR spect 26(1):3–49

58. Taylor N (1992) Dover’s smart bridge. Proc Inst Mech Eng Part I:

J Syst Control Eng 206(1):9–18

59. Vis IF, De Koster R (2003) Transshipment of containers at a

container terminal: an overview. Eur J Oper Res 147(1):1–16

60. Webb M et al (2008) Smart 2020: enabling the low carbon

economy in the information age. Clim Group Lond 1(1):1–1

61. Weske M (2012) Business process management: concepts, lan-

guages, architectures. Springer Science & Business Media, Berlin

62. Yang GH, Xu K, Li VO (2010) Hybrid cargo-level tracking

system for logistics. In: Vehicular Technology conference (VTC

2010-Spring), 2010 IEEE 71st. IEEE, pp 1–5

63. Yang SH (2014) Hybrid RFID/WSNs for logistics management.

In: Wireless Sensor networks. Springer, pp 235–246

Inf Technol Manag (2017) 18:223–239 239

123