Big Data for Operational Efficiency of Transport and Logistics: A Review

Abstract

In the new information and communication era, digital transformation and adoption of recent technological advances have become a must for all transport and logistics providers who aim to significantly improve their activities. Consequently, this digitalization is inevitably giving birth to voluminous and rapidly growing sets of large-scale data generated from heterogeneous data sources, also known as Big Data. With particular management infrastructures and advanced data analysis methodologies, these huge amounts of data can be efficiently harvested to optimize the logistics andtransport operations and provide a higher quality of service. This paper provides a review of the application of Big Data technologies in improving the operational efficiency of transport and logistics, exposes the main use cases and identifies some future research challenges.

Full text

Big Data for Operational Efficiency of Transport
and Logistics : A Review
Tawfik Borgi∗, Nesrine Zoghlami∗, Mourad Abed†Mohamed Saber Naceur∗
∗Universit´
e de Tunis El Manar
Ecole Nationale d’Ing´
enieurs de Tunis, LTSIRS Laboratory,1002,Tunis,Tunisia
Email: tawfik.borgi@gmail.com
†LAMIH, University of Valenciennes and Hainaut-Cambr´
esis
Le Mont Houy, 59313 Valenciennes, Cedex 9, France
Abstract—In the new information and communication era,
digital transformation and adoption of recent technological
advances have become a must for all transport and logistics
providers who aim to significantly improve their activities.
Consequently, this digitalization is inevitably giving birth
to voluminous and rapidly growing sets of large-scale data
generated from heterogeneous data sources, also known as
Big Data. With particular management infrastructures and
advanced data analysis methodologies, these huge amounts of
data can be efficiently harvested to optimize the logistics and
transport operations and provide a higher quality of service.
This paper provides a review of the application of Big Data
technologies in improving the operational efficiency of transport
and logistics, exposes the main use cases and identifies some
future research challenges.
Keywords- Big Data, Transport, Logistics
I. INT ROD UC TI ON
With the recent technical and scientific advancements in
information and communication technologies, and the prolif-
eration of use of sensors and connected devices, global data
sources have rapidly increased, leading to exponentially grow-
ing sets of data (see figure 1) [1]. This phenomenal generation
of massive data along with the opportunities it provides for
discovering new values and deriving new insights, and the
various challenges it attempts to raise in terms of analysis and
management, have created a new concept commonly referred
to as Big Data.
Logistics and Transportation sectors are among the most
ideally placed to take advantage of the analytical capabilities
and the methodological advancements of Big Data technolo-
gies. While managing the massive flows of individuals, freight
and goods, logistics and transportation services providers are
permanently creating vast and enormous datasets, especially
with the recent and continuous digitization of these sectors. For
millions of deliveries and shipments around the world every
day, content, location, weight, size, source and destination,
and many other information are getting tracked and collected
across the global transport and delivery network, presenting
thus multiple perspectives for applying Big Data analytics to
harvest all these large-scale data and provide new opportunities
in terms of service quality and operational efficiency.
The current paper aims to present a survey on Big Data op-
portunities in improving the operational efficiency of transport
and logistics. It is organized as follows. In section II, a general
overview of Big Data concept and technologies is provided,
defining terms and introducing some of the main aspects of Big
Data management and analytics. In section III, the different
types of data and the sensing modalities are exposed. In section
IV, the Big Data opportunities in improving the operational
efficiency of logistics and transport are analysed, with a review
of the main applications and use cases. A brief review of the
main perspectives and challenges Big Data technologies have
to raise in transport and logistics sectors is presented in section
V, before we conclude in section VI.
Fig. 1. The exponential data growth estimated between 2010 and 2020 [1]
II. BI G DATA DEFINITIONS AND TECHNIQ UE S
Big Data generally refers to voluminous data which cannot
be managed with classical methods and techniques within
a tolerable time. Laney [2] presented the 3Vs Model to
define this concept of Big Data : Volume (a great volume
of data), Velocity (data rapidly generated) and Variety (data
with various natures and modalities). Later definitions added
other characteristics such as the Veracity (to emphasize the
uncertainty of data) and Value (a great value with a low
density) to introduce the 4Vs and 5Vs models [3].
This expanding and enormous quantity of data is generated
by a huge and increasing number of sources. The complexity
of modern life and human activities as well as the rapid ad-
vances in some scientific and technological fields such as elec-
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
113
978-1-5386-1623-9/17/$31.00 ©2017 IEEE

tronic and communications, have leaded to the proliferation of
use of connected devices and ubiquitous networking, which
represent one of the major sources of data. Social Media,
enterprise data, healthcare systems, transport and logistics are
also sources of various types of information, generated at an
exponentially increasing rate and leading to huge amounts of
data sets.
Traditional Business Intelligence techniques and classical
processing and storage approaches were generally dedicated
to the management and analysis of structured Data (with
predefined formats). Yet, the sourced data in the Big Data
context are mostly unstructured data (without any defined
format) or a combination of both unstructured and structured
data (semi-structured), gathered from heterogeneous sources
including image, text, video and many types of data that
need novel methodologies in terms of processing, storage and
analysis.
A. Big Data Management Technologies
Due to the huge volume and the heterogeneity of Big Data,
new infrastructures must be used and a whole data manage-
ment process including more operations and pre-processing
steps must be carried out to derive the required business
insights and allow any data-driven decision making. Therefore
many new management platforms, tools and techniques, as
well as various new practices and user skills are being applied
and further investigated under the new Big Data Management
discipline [4]. A. Siddiqa et al.[5] proposed a general Big Data
management process flow consisting of four main elements as
illustrated in figure 2 :
-Data Storage : one of the fundamental concerns of Big
Data management is how to build systems with capacities
to store data on a petabyte scale. Such storage man-
agement systems are required to be capable of keeping
enough and sufficient data, improving its retrieval and
optimizing its availability to processing[6]. The storage
process is based on parallel activities to optimize the
storage operations.
-Data Pre-processing : in order to perform an efficient
data analysis, the data must first meet some quality
requirements. The pre-processing operations have the
main purpose of transforming inconsistent or incomplete
raw data into an understandable format ready for more
analysis process. Thus, they prepare data for further
processing by structuring them in a standard format to
obtain integrated data [7].
-Data Processing : The key step in Big Data management
process is the Data Processing. During this stage, insights
regarding the relationships between features are gained
which allows the development of effective analytics. It
includes various analysis methods, such as statistical
analysis, data mining and machine learning [8].
-Security : Due to the heterogeneity and multitude of data
sources in the big data environment, security is becoming
a more and more serious concern. Security involves
many issues such as privacy, integrity, confidentiality and
availability [5].
As for the infrastructures used in the Big Data management,
Cloud computing, which is an internet-based computing that
provides shared processing resources (application and services,
storage, servers, etc) [9], allows an important support to Big
Data management. The storage and computing resources can
also be expanded by replicating the current infrastructure and
clustering multiple additional nodes to create a distributed
environment and take advantages of parallelism (Scale Out),
instead of replacing the currently used technology with another
more powerful one (Scale Up) [10].
Big Data technologies are rapidly evolving thanks to the
increasing innovation activities in many information and com-
munication areas such as the Cloud Computing. A large
number of platforms and tools are being available for Big
Data analytics providing new distributed architectures and high
levels of memory and processing power. Big Data technologies
involve services and platforms for storage, processing and
security of data.
Fig. 2. Big Data Management Process
The Hadoop open-source framework, is a widely used
framework designed to manage large-scale data using a com-
modity hardware. It consists of a processing component :
MapReduce programming model, and a distributed storage
component : Hadoop Distributed File System (HDFS). The
HDFS storage approach consists in replicating the data into
multiple machines and storing them in blocks by a main server
(master node). On the other hand, the MapReduce model
divides a task into smaller ones (Map) and then combines
the obtained results (Reduce), to allow a powerful computing
(using parallelism) with a relatively low cost. The figure 6
illustrates the HDFS and MapReduce architecture.
Apache Spark [11] is another alternative to Hadoop which
uses in-memory primitives to perform faster distributed com-
puting. Spark loads data into memory and reuse it repeatedly to
overcome the problems presented by MapReduce in terms of
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
114

iterative and online processing [12]. Many other open-source
platforms have lastly emerged as competitors to Apache Spark,
such as Apache Storm [13] and Apache Flink [14]. These real-
time processing platforms use pure streaming processing and
represent an alternative to the limitations of Spark from online
and streaming side [12].
Fig. 3. HDFS and MapReduce architecture [15]
B. Big Data Analytics
The data analytics concept combines different disciplines
and presents multiple analytical methods applied to explore
and investigate the data and extract the meaningful information
for new business values and decision support [16]. These
analytics are generally about extracting useful information
and new insights from the large collections of data, to give
decision makers effective operational reporting and dashboards
to monitor the performance, KPI, thresholds, etc. They also
use statistical and mathematical methods to describe and
understand data, validate hypotheses, simulate scenarios and
make predictive forecasts.
Fig. 4. Data Mining tasks
As a key concept in Big Data Analytics, the Data Mining
consists in applying statistical methods and data science tech-
niques to explore and analyse a large collection of data with
the purpose of finding useful and meaningful patterns. It uses
sophisticated algorithms and statistical models (e.g. Machine
Learning algorithms) to perform three kinds of analytics:
Predictive analytics,Descriptive analytics and Prescriptive
analytics.
The descriptive analytics transforms the data into mean-
ingful information for monitoring, visualization and reporting
purposes (e.g. clustering, association rules, etc), while the
predictive analytics makes prediction about the future for
decision support (e.g. regression, time-series analysis, etc).
As for the prescriptive analytics, they include predictive and
descriptive analytics to suggest decision options for real-
time process optimization [17]. Examples of predictive and
descriptive analytics models and algorithms are given in the
figure 4.
III. DATA SOU RC ES A ND SE NS IN G MODA LI TI ES I N
TRA NS PO RT AN D LOG IS TI CS
Supply chain management, reverse logistics, express de-
livery, freight logistics, e-commerce, air, maritime and land
logistics, are all types of logistics operations that were totally
transformed and significantly improved by the adoption of
new information and communication technologies in transport
chains. The digital transformation triggered by these tech-
nologies is creating growing sets of large-scale data. The
Internet of Things (IoT), as a key trend in logistics and
transport industry, deploys recent communication technologies
such as the Machine to Machine Comunication (M2M) to
virtually connect objects to the internet [18]. In fact, IoT
can be deployed for car-to-car communication and various
Intelligent Transportation Systems (ITS), especially with the
proliferation of use of sensors, Radio-Frequency Identification
(RFID), Global Positioning Systems (GPS) and WIFI, which
gives objects and vehicles more and more connectivity, and
hence transform the logistics and transport activities into an
important source of Big Data.
Fig. 5. Using Tri-axial accelerometer data to determine travel modes [19]
The collected data may generally be either born analogue
data or born digital data [20]. Digital data are those created
by a computing device to be specifically used in a machine
processing environment. In the transport and logistics context,
they may include data produced by vehicles, devices and
networked objects, GPS or other spatial data stamps, public
transport cards, portal access or RFID tags data, mobile
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
115

devices connected to various kinds of networks (e.g. Wi-
fi,GSM), process logs and time stamps, SMS, Emails, etc.
For digital data sources, only specific data are produced and
retained to address some specific needs, in order to avoid
inflating storage and processing costs and capacity [19].
On the other hand, born analogue data come originally
from imprints of some physical phenomenon such as sound,
light, or presence of particular chemical compound, acquired
by a specific sensor and converted into a digital signal.
Sensing devices may include microphones, cameras, thermal
sensors, accelerometers, etc. These born analogue data can
be video streams from in-vehicle, surveillance or roadside
cameras (e.g. CCTV cameras for road traffic monitoring),
motion (accelerometers data), audio records of voice phone
calls, infrared radiation, light or electromagnetic reflectance
of objects, air pressure, temperature, etc [19].
Fig. 6. Interface of OpenStreetMap [21]
The gathered data in transport and logistics context are
generally sourced based on one of the three main sensing
modalities : purposely sensing, opportunistic sensing and
crowd sensing [19]. As for the opportunistic sensing, data
collected for one specific purpose and can be used for another.
Based on this approach, mobility can for example be better
understood by simply leveraging data arising from mobile
phones or credit card companies records [22]. The recent
transition in processing power and reporting frequency of
existing networks, as well as the increasing use of smart
phones with GPS capability has furthered the potential of
opportunistic sensing in gathering mobility data. The example
in figure 5 shows the possibility of determining the transport
mode by using data from an accelerometer-equiped phone de-
vice. In contrast of this modality, purposely-sensing modality
acquire data sets from ad-hoc sensor networks pre-configured
to study and analyse a specific object or phenomenon. With
the advances in computation and microelectronics, the use of
such sensors and their proliferation is often referred to a new
concept termed ”Smart Dust”. In this case, the gathered data
sets are usually more uniform and better aligned to process
and analyze for information and new insights extraction.
However, collecting real-time travel conditions or traffic
information (weather, traffic incidents, etc) in the purposely-
sensing modality, by using vehicle mounted sensing devices
(Vehicle to Vehicle communication - V2V and Vehicle to
Infrastructure communication - V2I), is still impractical in
many cases, for reliability and cots issues [23]. Instead, the
wide spread of use of smart phones and communication
devices has introduced the crowdsensing as an alternative in
which large-scale people can participate in the collection and
analysis of transport data. As an example, travel time and real-
time traffic flow information are usually inferred using crowd
sensing from taxis equipped with GPS devices [24]. Another
famous example is the OpenStreetMap [21] project which aims
at collaboratively create a free and editable map of the world
(see figure 6).
From a technical perspective, all these voluminous and
heterogeneous data might be seen as real problem and a
forbidding challenge in terms of storage, management and
processing. Yet, a huge opportunity in enhancing the quality of
processes and improving the operational efficiency of logistics
and transport services can also be gained from the untapped
potential of these Big sets of data.
IV. IMP ROV IN G OPE RATI ONA L EFFIC IE NC Y OF
TRA NS PO RT AN D LOG IS TI CS
One of the most challenging purposes of adopting Big Data
in logistics and transportation is to improve the operational
efficiency, by providing better decision making in terms of
operational and strategic resource planning, and improving
performance and process quality. In fact, operational efficiency
is primarily about providing high quality services in the
most cost-effective manner. Thus, taking advantages of the
capabilities of Big Data technologies to optimize the utilization
of resources and reduce operational costs, can be an important
advantage for any logistics providers [25].
A. Last-mile delivery
A relevant use case in which Big Data can significantly
improve the operational efficiency is the last-mile optimization
[1]. Traditionally, the last-mile delivery is known to be the
most complicated and the most expensive part of the delivery
process. In fact, transit points and transport routes used to be
coordinated and planned based on inconsistent information and
resources such as historical averages or even personal expe-
rience. Yet, Big Data analytics makes it possible to optimize
the matching between demand and available resources, and
perform real-time optimization of delivery routes to transform
the delivery network into a self-organizing infrastructure.
Yamato, the largest Japanese door-to-door delivery service
used to rely on the know-how of its sales drivers and their
personal experience to plan the delivery route on a daily basis.
Although this strategy allowed the company to build over
years a high-density delivery network in Japan, which could
be compared to an infrastructure like electric power or water
supply, it unfortunately can’t enable its replication in another
country. Therefore, Yamato started to collect Big Data of its
sales drivers to extract all the knowledge they acquired over
many years (the delivery route order, the map of the city and
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
116

its fastest routes, and recipients schedules) and replicate it in
a new delivery infrastructure [26].
Fig. 7. GUI of Routific : a last-mile delivery optimization tool [27]
Another challenging issue in the last-mile delivery optimiza-
tion is the delivery route monitoring and optimization. In fact,
by tracking and monitoring the freight vehicles and shipment
items, the delivery routes can be analysed and automatically
changed according to new real-time information on traffic
conditions, weather, or any other incidents. This automated
control of large fleets of moving delivery resources can be ef-
fectively carried out using the real-time processing capabilities
of Big Data technologies [1]. The graphical user interface of
a commercial route optimization software is exposed in figure
7. This tool named Routific [27] allows the tracking of freight
vehicles fleet for a last mile delivery optimization.
B. Route Optimization
Big Data descriptive and predictive capabilities can also
provide all the needed information to perform optimization
tasks, resource planning and determine an optimal path re-
garding certain parameters and a certain cost. For instance,
by gathering and analysing large sets of traffic data, traffic
flow can be predicted and a real-time route optimization can
be performed by detecting traffic congestion on roads and
suggesting the best options for each case [28]. Lv et al. [29]
proposed a deep-learning traffic flow prediction approach by
training a deep neural network using a two months traffic flow
data collected each 30 seconds from 15000 detectors deployed
in freeway systems accross California. The traffic flow moni-
toring and congestion prediction is a key requirement in many
route optimization applications.
Besides the extraction of insights on travel, road and traf-
fic conditions, new optimization methodologies and efficient
routing approaches can be implemented via Big Data analysis
of the large-scale collected data. As an example, the huge
aviation data can be harvested to efficiently decide the routes
and manage the freights and passengers, in order to allow
effective flight plans and reduce route distance, fuel costs
and lost revenues from payload that could not be carried.
Kasturi et al. [30] used large scale aviation data to propose
an airline route profitability optimization model based on
Big Data analytics under multiple meta-heuristic methods.
Ning et al. [31] proposed a real-time dynamic medical rescue
vehicles scheduling and optimization to reduce running time
and improve punctuality, by using genetic algorithms and Big
Data technologies. Optimal route discovery and shortest path
problems can likewise be solved by analyzing Big Data (e.g.
Big trajectory data) and using data-driven approaches [32]
[33].
C. Crowdsourcing and Social Transportation
Crowdsourcing is a novel mean of solving problems through
contributions of a large online communities. In transport
and logistics applications, crowdsourcing approaches use the
collective wisdom to make easier the real-time collection and
update of data, allowing thus the design of more efficient
transportation systems. For instance, the GPS-based geograph-
ical navigation application Waze uses shared or user-generated
information to provide traffic monitoring and real-time route
guide [34] (see figure 8). Uber [35], the transportation and
delivery service is another relevant example of crowdsourcing
location-based applications, that uses shared location data and
transportation information.
Furthermore, the increasing use of smart phones and com-
munication devices, ubiquitous social networking websites and
apps such as Tweeter or Facebook, are becoming a space
in which people can provide a live description of the traffic
situation and instantly report any new incidents (traffic acci-
dent, traffic jams). By applying several analytics techniques
such as text analytics, natural language processing and pattern
recognition, traffic state information can be collected in real-
time [36][37].
Fig. 8. GUI of Waze : a real-time route guide [34]
Traffic forecasting and traffic monitoring using social sig-
nals from social media, wearable devices and mobile phones,
represents a typical example of application in the emerging
Social Transportation research field. By including dynamics
of social behaviour, cultures and organizations, Big Data and
social transportation can build a world of connected people,
vehicles, infrastructures and services to enter a new era of
Intelligent Transportation Systems [23]. By analysing the large
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
117

sets of collected social signals data such as travel behaviour
or chosen trajectories, Song et al. [38] demonstrated that
the human mobility can be highly predicted. Consequently,
many new insights can be inferred such as the estimation of
travel time, the prediction of traffic flow, and the detection
of congestions in roads, which allows a more efficient traffic-
aware travel planning and routing optimization [39].
D. Smart Logistics
Adopting Big Data technologies to improve the operational
efficiency of transport and logistics is still a new trend.
However, the new advances in Data Science, Information
Technologies, Artificial Intelligence and Robotics, are leading
to the rapid emergence of many new scientific and techno-
logical concepts. The fourth industrial revolution also termed
Industry 4.0, aims at fully integrating the Internet of Things,
Artificial Intelligence, and Information Technologies in manu-
facturing, leading thus to the emergence of many new innova-
tive concepts such as Smart Technologies (cyber-physical and
computer-assisted systems with the ability of automatically
control entire work processes) and Smart Factory (self-learning
and self-regulating processes) [40].
These new capabilities of autonomous process and opera-
tion management are leading to the implementation of fully
autonomous decision processes in the transport and logistics
area, also known as Smart Logistics. The introduction of
self-driving and intelligent vehicles in a Big Data and IoT-
based infrastructure [41] is one of the key trends in Smart
Logistics, and can enable fully automated and more flexible
logistics solutions. As an example of futurist autonomous
vehicles, the ”Future Truck 2025” [42] is a prototype of a
self-driving truck designed by Mercedes for freight transport.
Amazon is also exploring an alternative for home-delivery
service using autonomous drones, which will be able to track
customer location using his phone GPS and deliver the parcel
within 30 minutes [26]. In this context, Big Data technologies
represent a key asset in improving the operational efficiency of
transport and logistics, since the more extensively information
on traffic, weather, or vehicles sensors data are available, the
more efficiently the logistics and transport flows can be self-
managed and optimized.
E. Anticipatory Logistics
The Anticipatory Logistics concept relies on Big Data
predictive analytics to anticipate demand before requests and
orders are placed, which boosts the service quality and process
efficiency and leads to shorter delivery times. The Smart City
Planning uses anticipatory algorithms with the purpose of
matching logistics resources with demand. In this context,
Anticipatory Shipping enable a same-day or even a one-hour
delivery services, by anticipating the customer demand and
moving goods to distribution points and centers closer to zones
where potential customers were detected based on the analysis
of their purchasing behaviour.
A pertinent example of anticipatory shipping was imple-
mented by Amazon, who patented in 2013 an effective method
for a anticipatory package shipping system [43] (see figure 9).
However, there are still many challenges in integrating these
prediction-based anticipatory shipping systems with traditional
order-based delivery methods.
Fig. 9. Anticipatory Shipping method developed by Amazon [43]
Anticipatory Logistics also involves the concept of pre-
dictive maintenance, which consists in predicting the main-
tenance needs of a machine right before the failure occurs.
By analysing the data collected from real-time monitoring
of machines or vehicles, predictive maintenance systems can
detect the faults and predict the maintenance requirements
early, reducing hence the downtime costs for both logis-
tics service providers and the customers [44]. For instance,
Siemens is heading to innovative and next-generation pre-
dictive maintenance services, by introducing the concept of
Internet of Trains, which consists in enabling new data-driven
methodologies for predictive maintenance of trains fleet [45].
V. CH AL LE NG ES A ND PE RS PE CT IV ES
The large-scale multi-modal data collected in transport and
logistics applications are leading to numerous challenges to
raise in terms of processing capabilities, computer architecture,
and in data science and analytics. The incompleteness, the
inconsistency and the heterogeneity of these sensed data re-
quire a new set of methodologies and analysis algorithms to be
explored and further enhanced. The Big Data streams sourced
from sensors and devices used in logistics and transportation
systems are one of the key areas of data mining applications.
Consequently, new perspectives are opened for exploring novel
approaches of Big Data a such as the Streaming Analytics
[46][44] which deals with the analysis of data-in-motion, and
the IoT Big Data Stream Mining [47] to derive insights from
the internet of things data.
Data fusion methodologies are also widely used in trans-
port and logistics applications to improve the descriptive and
predictive analytical capabilities of the implemented systems.
However, data fusion is still in its preliminary stage of
engineering and scientific fields especially in logistics and
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
118

transportation area, and requires new systematic methodolo-
gies on data analysis and organization [23]. Furthermore, the
Big Data predictive analytics have also their own limitations.
In fact, predictive models based on data fitting are not able
to resolve all the issues and provide accurate predictions,
since history cannot always predict the future. Thus existing
predictive models are still not able to offer reliable anticipatory
or fully self-regulating or self-learning processes.
Besides, the security and privacy issues represent a serious
challenge for the transmission and storage of the large-scale
sourced transport data. The generated new knowledge from
the fusion of the opportunistically sensed and crowd-sensed
data can actually open avenues for misuse, since it involves
information and insights derived from data without the consent
of their sources (people or organizations who are the object of
these data). In addition, the increasing abilities of these multi-
platform sensing technologies to precisely track and locate
people and vehicles without their permission, is creating new
challenges in terms of data anonymization, confidentiality and
policy [48].
VI. CO NC LU SI ON
This paper explored a review of Big Data technologies
and analysed the opportunities and benefits they provide to
transport and logistics in terms of operational efficiency. The
predictive and analytical capabilities of Big Data technologies
as well as the new insights and knowledge they can derive
from the huge amount of data collected along the logistics
and transport chains, offer valuable opportunities in terms of
operational efficiency. Last-mile delivery, route optimzation,
crowdsourcing and social transportation, Smart Logistics and
Anticipatory Logistics are the most likely to revolutionize
transport and logistics in the coming years. Yet, more efforts
still need to get involved into surpassing the scientific and
technical challenges and dealing with the policy and privacy
issues that transport and logistics Big Data are raising.
For future studies, it would be interesting to investigate new
data-driven approaches that use big data analytics (descriptive,
predictive and prescriptive analytics) to deal with the optimza-
tion problems encountered in the transport and logistics fields.
More specifically, smart logistics and self-driving vehicles
which are expected to increasingly become the norm, are
essentially based on fully autonomous decision processes that
need more advanced analytics models to solve in real time the
optimization problems. In this context, prescriptive analytics,
as an emerging and promising technology, can be further
investigated in order to propose new prescriptive analytic
optimization models for the decision support in transport and
logistics applications.
REF ER EN CE S
[1] F.Wei M.Jeseke, M.Grner. Big data in logistics a dhl perspective on how
to move beyond the hype. Technical report, DHL Customer Solutions
and Innovation, December 2013.
[2] Laney. D. Meta group. application delivery strategies. Mobile Networks
and Applications, 2001.
[3] A.E. Hassanien, A.T. Azar, V. Sn´
asel, J. Kacprzyk, and J.H. Abawajy.
Big Data in Complex Systems: Challenges and Opportunities. Studies
in Big Data. Springer International Publishing, 2015.
[4] Philip Russom. Managing big data. Technical report, TDWI research,
2013.
[5] Aisha Siddiqa; Ibrahim Abaker Targio Hashem; Ibrar Yaqoob; Mohsen
Marjani; Shahabuddin Shamshirband; Abdullah Gani; Fariza Nasarud-
din. A survey of big data management : Taxonomy and state-of-the-art.
Journal of Network and Computer Applications, 2016.
[6] Buza K.;Nagy G.I.;Nanopoulos A. Storage-optimizing clustering algo-
rithms for high-dimensiona ltick data. ExpertSyst.Appl, 2014.
[7] Jerry M Mendel and Mohammad M Korjani. On establishing nonlinear
combinations of variables from small to big data for use in later
processing. Information Sciences, 280:98–110, 2014.
[8] Salvador Garca, Julin Luengo, and Francisco Herrera. Data Prepro-
cessing in Data Mining. Springer Publishing Company, Incorporated,
2014.
[9] Qusay Hassan. Demystifying cloud computing. The Journal of Defense
Software Engineering, 2014.
[10] Chun-Wei Tsai, Chin-Feng Lai, Han-Chieh Chao, and Athanasios V.
Vasilakos. Big data analytics: a survey. Journal of Big Data, 2(1):1–32,
2015.
[11] H. Karau, A. Konwinski, P. Wendell, and M. Zaharia. Learning Spark:
Lightning-Fast Big Data Analytics. O’Reilly Media, Sebastopol, 2015.
[12] Salvador Garc´
ıa, Sergio Ram´
ırez-Gallego, Juli´
an Luengo, Jos´
e Manuel
Ben´
ıtez, and Francisco Herrera. Big data preprocessing: methods and
prospects. Big Data Analytics, 1(1):9, 2016.
[13] Storm. apache storm. 2017. http://storm-project.net/. accessed march
2017.
[14] Flink. apache flink. 2017. https://flink.apache.org/. accessed march 2017.
[15] Mark Rittman. Going beyond mapreduce for hadoop etl pt.2 : Introduc-
ing apache yarn and apache tez, 2014.
[16] Rajeshwari.d. State of the art of big data analytics: A survey. Inter-
national Journal of Computer Applications, 120(22):39–46, June 2015.
Full text available.
[17] Mondher Feki, Imed Boughzala, and Samuel Fosso Wamba. Big data
analytics-enabled supply chain transformation: A literature review. In
49th Hawaii International Conference on System Sciences, HICSS 2016,
Koloa, HI, USA, January 5-8, 2016, pages 1123–1132, 2016.
[18] Markus Kckelhaus. Logistics trend radar. Technical report, DHL Trend
Research, 2016.
[19] Big data and transport : Understanding and assessing options. Technical
report, International Transport Forum, 2015.
[20] PCAST. Big data and privacy: A technological perspective. Technical
report, President’s Council of Advisors on Science and Technology:
Executive Office of the President, 2014.
[21] Openstreetmap. 2017. https://www.openstreetmap.org/. accessed march
2017.
[22] K. He, J. Wang, Lianbo Deng, and P. Wang. Congestion avoidance
routing in urban rail transit networks. In 17th International IEEE
Conference on Intelligent Transportation Systems (ITSC), pages 200–
205, Oct 2014.
[23] X. Zheng, W. Chen, P. Wang, D. Shen, S. Chen, X. Wang, Q. Zhang,
and L. Yang. Big data for social transportation. IEEE Transactions on
Intelligent Transportation Systems, 17(3):620–630, March 2016.
[24] Bei Pan, Yu Zheng, David Wilkie, and Cyrus Shahabi. Crowd sensing
of traffic anomalies based on human mobility and social media. In
Proceedings of the 21st ACM SIGSPATIAL International Conference on
Advances in Geographic Information Systems, SIGSPATIAL’13, pages
344–353, New York, NY, USA, 2013. ACM.
[25] Vesna Radonjic Dogatovic Branka Mikavicaa, Aleksandra Kostic-
Ljubisavljevica. Big data: Challenges and opportunities in logistics
systems. 2015.
[26] Carlos Cordon, Pau Garcia-Mil`
a, Teresa Ferreiro Vilarino, and Pablo
Caballero. Bikes or Drones to the Consumer: The Logistical Challenge
of the Last Mile, pages 85–98. Springer International Publishing, Cham,
2016.
[27] Routific. 2016. https://routific.com/. accessed july 2016.
[28] Anshul Mahajan and Arvinder Kaur. Predictive urban traffic flow model
using vehicular big data. Indian Journal of Science and Technology,
9(42), 2016.
[29] Y. Lv, Y. Duan, W. Kang, Z. Li, and F. Y. Wang. Traffic flow
prediction with big data: A deep learning approach. IEEE Transactions
on Intelligent Transportation Systems, 16(2):865–873, April 2015.
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
119

[30] E. Kasturi, S. Prasanna Devi, S. Vinu Kiran, and S. Manivannan. Airline
route profitability analysis and optimization using {BIG} {DATA}
analyticson aviation data sets under heuristic techniques. Procedia
Computer Science, 87:86 – 92, 2016. Fourth International Conference on
Recent Trends in Computer Science and Engineering (ICRTCSE 2016).
[31] Tao Ning, Hua Jin, Meiyi Yang, and Fang Liu. Research on the shortest
route algorithm for medical rescue vehicles based on big data. Journal
of Investigative Medicine, 64(Suppl 8):A21–A22, 2016.
[32] M. Li, X. Li, and J. Yin. Tord problem and its solution based on
big trajectories data. IEEE Transactions on Intelligent Transportation
Systems, 17(6):1666–1677, June 2016.
[33] Z. Cao, H. Guo, J. Zhang, D. Niyato, and U. Fastenrath. Finding the
shortest path in stochastic vehicle routing: A cardinality minimization
approach. IEEE Transactions on Intelligent Transportation Systems,
17(6):1688–1702, June 2016.
[34] Waze mobile. 2017. https://www.waze.com/fr/. accessed march 2017.
[35] Uber technologies inc. 2017. https://www.uber.com/. accessed march
2017.
[36] Arunkumar.A Gayathiri.R. Opinion Mining On Traffic Dataset Using
Rule Based Approach. International Journal of Computer Science and
Mobile Computing, 5, 2016.
[37] H. Mo, X. Hao, H. Zheng, Z. Liu, and D. Wen. Linguistic dynamic
analysis of traffic flow based on social media : A case study. IEEE
Transactions on Intelligent Transportation Systems, 17(9):2668–2676,
Sept 2016.
[38] Chaoming Song, Zehui Qu, Nicholas Blumm, and Albert-L ´
aszl´
o
Barab´
asi. Limits of predictability in human mobility. Science,
327(5968):1018–1021, 2010.
[39] C. Chen, D. Zhang, B. Guo, X. Ma, G. Pan, and Z. Wu. Tripplanner:
Personalized trip planning leveraging heterogeneous crowdsourced digi-
tal footprints. IEEE Transactions on Intelligent Transportation Systems,
16(3):1259–1273, June 2015.
[40] K. Zhou, Taigang Liu, and Lifeng Zhou. Industry 4.0: Towards future
industrial opportunities and challenges. In 2015 12th International
Conference on Fuzzy Systems and Knowledge Discovery (FSKD), pages
2147–2152, Aug 2015.
[41] S. Misbahuddin, J. A. Zubairi, A. Saggaf, J. Basuni, S. A-Wadany,
and A. Al-Sofi. Iot based dynamic road traffic management for smart
cities. In 2015 12th International Conference on High-capacity Optical
Networks and Enabling/Emerging Technologies (HONET), pages 1–5,
Dec 2015.
[42] The long-haul truck of the future. https://www.mercedes-benz.com/.
Accessed March 2017.
[43] J.R. Spiegel, M.T. McKenna, G.S. Lakshman, and P.G. Nordstrom.
Method and system for anticipatory package shipping, December 2013.
[44] Emanuele Fumeo, Luca Oneto, and Davide Anguita. Condition based
maintenance in railway transportation systems based on big data stream-
ing analysis. Procedia Computer Science, 53:437 – 446, 2015.
[45] The internet of trains : Analysing sensor data helps siemens keep
operators on track by reducing train failures. Technical report, Teradata,
2015.
[46] Daniel D. Gutierrez. Guide to streaming analytics. Technical report,
InsideBIGDATA, 2016.
[47] Gianmarco De Francisci Morales, Albert Bifet, Latifur Khan, Joao
Gama, and Wei Fan. Iot big data stream mining. In Proceedings of the
22Nd ACM SIGKDD International Conference on Knowledge Discovery
and Data Mining, pages 2119–2120, New York, NY, USA, 2016. ACM.
[48] Data-driven transport policy. Technical report, International Transport
Forum, 2016.
2017 6th IEEE International Conference on Advanced Logistics and Transport (ICALT)
120