=Paper=
{{Paper
|id=Vol-2215/paper13
|storemode=property
|title=An Agent Framework to Support Air Passengers in Departure Terminals
|pdfUrl=https://ceur-ws.org/Vol-2215/paper_13.pdf
|volume=Vol-2215
|authors=Maria Nadia Postorino,Luca Mantecchini
|dblpUrl=https://dblp.org/rec/conf/woa/PostorinoM18
}}
==An Agent Framework to Support Air Passengers in Departure Terminals==
https://ceur-ws.org/Vol-2215/paper_13.pdf
An Agent Framework to Support Air Passengers in
Departure Terminals
Maria Nadia Postorino ∗ Luca Mantecchini§ ,
∗ DICEAM, University of Reggio Calabria, Feo di Vito, 89122 Reggio Calabria, Italy, e-mail:npostorino@unirc.it
§ DICAM, University of Bologna, Viale Risorgimento 2, 40136, Bologna (BO), Italy, e-mail:luca.mantecchini@unibo.it
Abstract—Airports are complex nodes performing several roles of this study, a possible approach is that of providing air
such as interchange terminal, shopping and relaxing center, travelers with personalized recommendations [13], [14] about
meeting area for short-time business activities. Airport operators the available commercial facilities that suit as more as possible
pay great attention to financial profits from their managed
assets, while passengers desire spending their slack time inside their preferences and interests. This task is rather complex to
the terminal in a pleasant way after wasting time in queues realize because it requires several steps, namely:
and controls to access the gate areas. In such a context, an • acquiring preliminary information about passengers and
agent framework is proposed to support travelers’ slack time their interests by complying with privacy rules at the same
by providing purchase suggestions potentially interesting for
them. Recommendations are computed by taking into account time [15];
passengers’ interests, their current position inside the departure • tracking passengers’ movements around the terminal [16];
terminal and the commercial opportunities available therein. • taking into account passengers’ slack time [17].
Index Terms—Airport terminal; Arrivals distribution; Multi- Providing passengers with a personalized support requires
agent system; Recommender system
the knowledge of some information about them (i.e., age,
I. I NTRODUCTION sex, job) and their main interests (i.e., preferences for product
categories). As for personal information, some of them could
Airports play a fundamental role in the mobility of people be obtained/deduced when a passenger is processed at security
and goods for middle-long range trips and a significant number checkpoints before entering the departure area1 . Unfortunately,
of studies exist on the different aspects involved in their data gathered at airport security checkpoints are subjected to
management (e.g., transport, financial and security issues) [1]– manifold restrictions mainly due to privacy rules [20], which
[6], included strict regulatory constraints to cope with air can also differ among countries, and are not available for the
travelers inside the departure terminals [7]. above aims. Therefore, the main way to acquire the desired
Three main facilities can be identified for a given airport [8]: information is that of requiring it directly to passengers [21],
i) service areas, ii) waiting areas and iii) commercial activities. for instance in return for the access to some reserved terminal
More in detail, in the service areas passengers access all the services offered for free (e.g., Wi-Fi connection, discounts).
services addressed to process the flow of travelers departing The second listed step is addressed to identify the current
from the airport (e.g., check-in, passport and security controls, traveler’s position inside the terminal for providing him/her
baggage drop) [9]. Waiting areas, where passengers may wait with suitable personalized suggestions, offers and so on, by
before boarding their flight, are equipped with seats (e.g., following a “now, here, only-for-me” approach. As the precise
lounges and open seating areas, usually close to the departure tracking of passengers’ movements by using GPS2 is practi-
gates) and free services for travelers (e.g., information desks, cally impossible, the remaining opportunities are i) the analysis
Wi-Fi, toilets). Finally, commercial areas consists of shops, of the images by dedicate cameras (e.g., different from those
food courts, currency exchange and so on, available to air belonging to the security system) and/or ii) the explotation of
travelers waiting their flights. Wi-Fi [22]/Bluetooth [23] connections used by smarthphones
In this scenario, air passengers desire to both avoid wasting and tablets. In particular, the analysis of camera images also
time in queues and controls and spend pleasantly their slack provides people density information [24], while each Wi-Fi
time inside the terminal [10], [11]. On the other hand, airport and Bluetooth fingerprints can return the number of connected
operators have to optimize all the terminal activities and, devices for each of their hot-spot (note that these two wireless
at the same time, give profitability to the assets directly or technologies should be set with different and suitable operating
indirectly managed by them [12] (e.g., commercial areas inside ranges and that, in the proposed framework, very low power
the terminal, car parks outside the terminal and so on) by Bluetooth connections will work only as detection points).
complying with national and international rules [7].
1 In large airport, security checkpoint operators realize the first identification
The challenge of satisfying both the passengers’ desire of
enjoying their slack time and the airport operator’s needs of by using ticket and document [18], while a further identification may be
based on video analysis processes realized by specialized softwares, which
increasing the revenues from the airport commercial areas is also allows the traveler’ movements around the terminal to be tracked [19].
of great interest. As for the first goal, which is the focus 2 The use of the GPS technology is very difficult inside the terminal.
75
The final required step is addressed to know in advance the II. T HE PROPOSED AGENT- BASED F RAMEWORK
number of passengers that could be present in the departure
The structure of the proposed agent-based framework (from
terminal at a given time and their estimated slack time.
hereafter only AF) is quite simple (see Figure 1) and consists
Generally, the whole amount of demand on a yearly basis
of three mutually interacting components which are:
is estimated to understand the development opportunities for
• the Personal Agent (PA);
both airlines and airports [25], [26]. However, in this case
• the Commercial Agent (CA);
it is more relevant to identify not simply the number of
• the Terminal Agent (TA);
departing passengers but mainly the air terminal processing
procedures [27]. Tasks, profiles and behaviors of all the AF components will
In particular, processing procedures are modeled by queuing be briefly described in the following sections.
theory approaches, which include the estimate of the time
required to carry out a terminal procedure, roughly made by
the time necessary to provide the service to the passenger and
the time the passenger has to spent in queue. The service time
depends by both the nature of the service and the specific TA
adopted procedure, which often follows service requirements CA
and/or security regulations. The time spent in queue depends PA
on the length of the queue that, in turn, depends on i) the
number of passenger requiring the service at the same time
interval and ii) the efficiency of the service process [16]. The
set of terminal procedures can be summarized as a chain of
processes, where two consecutive processes are separated by
time varying intervals. In large airports also acting as hubs,
the number of passengers increases quickly during peak hours
and generates congestion and delays, which generally increase
the time required to perform terminal procedures [27]. Mod-
eling the probability density function describing passengers
arrivals at airport facilities allows better managing airport
resources [28].
In the context above described, this paper intends to con- Fig. 1. The Agent-based Framework
tribute by designing an agent-based framework where each
traveler inside the airport departure area is associated with a
A. The Personal Agent (PA)
personal agent whose goal is to provide personalized sugges-
tions for enjoying slack times. More in detail, the personal A PA is an agent pre-activated and identifiable (i.e.,
agent runs on a traveler’s mobile device, by exploiting a equipped a unique identifier) associated with a free app
free app required for accessing some reserved free terminal running on a passenger’s device (i.e., a smartphone, a notebook
services, and by interacting with both its user and other com- or a tablet). This app is provided by the airport company and it
ponents of the agent-based framework, will build a light user is necessary to access to some services provided for free to the
profile as closed as possible to his/her interests [29]. In such a passengers. The PA does not require expensive computational
scenario, the user will be supported with some suggestions tasks and/or a great amount of storage resources to its host
potentially interesting for him/her generated by using both device.
content-based (CB) [30] and collaborative filtering (CF) [31] The main tasks carried out by a PA include:
techniques. Such recommendations will take into account i) the • interacting with its owner both for:
user profile, ii) the current position of the traveler (determined – acquiring some personal information;
by exploiting his/her device and the terminal hot-spots), iii) – acquiring the product categories meeting the owner’s
his/her slack time also depending on the crowd degree of interest with respect to the terminal resources;
the terminal areas and the amount of time remaining before
• interacting with CAs and TAs agents.
the scheduled departing time and iv) the terminal resources
available for the traveler. Each PA builds, manages and updates a light XML agent
The remaining of the paper is organized as follows. The profile [32], represented in Fig. 2, consisting of:
proposed agent-based framework is described in detail in • the PA Identifier (PAId), which is unique into the AF;
3
Section II, while Section III introduces some information about • some basic personal owner data ;
the computation of the slack time, and Section IV deals with
3 Note that the system recognizes a traveler by means of the identifier of
the proposed recommender algorithm. Section V presents the
its associated PA and, therefore, the required personal data only consists of
relevant literature related to the matter presented here and, age, sex, job, trip reason and similar information voluntarily provided by the
finally, in Section VI some conclusions are drawn. traveler in the respect of his/her privacy.
76
PA Profile the nearest TA6 ); ii) sends its product category list to
all the PA connected to its associated hot-spot; iii) sends
PA Identifier (PAId) periodically the number of devices and the identifiers of
Personal Data the PAs connected to its associated Bluetooth hot-spot
Product Categories and the identifiers of the PAs that interacted with the
of Interest CA, to assist a traveler in a purchase, to the nearest TA.
Fig. 2. The PA Profile C. The Terminal Agent (TA)
A TA is an agent associated with a Wi-Fi hot-spot. A TA
• the product categories of interest selected by the owner 4 also generates CF recommendations (for all the PAs connected
from a list taking into account the commercial resources with its associated Wi-Fi hot-spot) by taking into account both
available inside the terminal (see below). the travelers’ interests and the commercial resources available
The PA behavior consists of two main activities, namely: inside its operating range.
More TAs can be active into the AF and, from a functional
• setup: The first time it is active, it receives the list of
point of view, they are interchangeable.
the AF product categories of interest (from a TA) and Each TA performs manifold activities, more precisely it:
interacts with the PA owner to acquire both personal and
• takes into account the number of wireless connections
interests information.
• operative: To support its owner, the PA:
and agents (e.g., PAs and CAs) active in the operating
range of its associated Wi-Fi hot-spot;
– interacts with CAs and TAs, associated with • stores the resources available in its operating range.
Bluetooth/Wi-Fi hot-spots, each time its device en-
• stores the profiles of each connected PA agent;
ters into their operating ranges5 ; • collects the information received by CAs (see above) to
– updates its profile and sends a copy to the nearest roughly monitor the travelers’ position inside the terminal
TA (see below); and their potential interests;
– computes (CB) and presents (CB and CF) recom- • generates CF recommendations (see Section IV);
mendations generated for the PA owner (see Sec- • maintains an updated list of the AF product categories;
tion IV) and other opportunities (like advertising,
discount codes and so on) proposed, also based To realize its goals a TA builds, manages and updates a
on his/her current position inside the terminal (see XML profile [34], represented in Fig. 3, and uses the infor-
below). mation stored in its profile to realize its goals. In particular,
a TA profile is formed by the three sections i) Working Data,
B. The Commercial Agent (CA) ii) PA Data and iii) CA Data. More in detail:
• the Working Data section stores:
This agent is associated with both a commercial facility
placed inside the departure terminal and a Bluetooth hot-spot. 6 When a new product category needs, its insertion into the AF list of
The main activities of a CA consist of: product categories can be required to a TA by the associated CA.
• storing and updating the list of the product categories
made available by the associated commercial facility;
• considering the number of devices and PAs active in the TA Profile
operating range of its associated Bluetooth hot-spot; TA Identifier
• taking into account the detected purchases performed
...
Agent List Agent Identifier
with the PA assistance (e.g., by using a TA discount Working Data ...
code). Product Category
• interacting with PAs and TAs agents; List
The CA behavior consists of the following activities: ...
Agent Identifier
• setup: The first time the CA is active, it registers its
Personal Data
presence with the closer TA from which receives its PA Data
...
identifier and the AF list of all the product categories Product Categories Category
of Interest
available into the terminal. ... ...Time
• operative: The CA : i) selects from the AF list of product
categories those present into its associated commercial ...
facility (when this CA list changes, it sends a copy to Agent Identifier
SA Data Product Category
List
4 Note that for privacy reasons the PA will not monitor the traveler’s activity
on his/her device for automatically extracting his/her interests [33].
...
5 Note that the PA exploits Bluetooth connections only to communicate its
presence in a narrower range with respect to that of a Wi-Fi hot-spot. Fig. 3. The TA Profile
77
– the TA Identifier; • If BP T is obtained when the passenger accesses to the
– the Agent List containing the identifiers of all the security area, before queuing for the security service, the
contacted PAs and CAs; related slack time computed by Eq.1 is overestimated. To
– the Product Category List storing all the product obtain a more reliable estimate, the time spent in queue
categories managed by the TA. by each passenger should be used. It is worthwhile to note
• The PA Data stores the PA profiles and, for each product that only the average waiting time can be estimated under
category of interest, takes into account the PA owner’s the hypothesis that there is no other automatic detection
interest (see Section IV); system before passing through the metal detector. To this
• The CA Data stores the identifier of each CA and aim, the approach proposed in [28] is briefly summarized.
the product categories made available by its associated Each 15 minutes data are aggregated and the discrete
commercial facility. arrival distribution is obtained based on how many pas-
The behavior of a TA consists of the following activities: sengers arrive in each time interval. Once obtained this
discrete arrival distribution, the underlying probability
• assigning an identifier (unique into the AF) to all the
density describing the arrival process is identified by
CAs requiring to be registered and providing PAs and
means of a Chi-Square test. The estimated passenger
CAs with the updated list of the Categories of Interest
probability density function f (x), where x is the Early
managed by the AF;
Scheduled Delay ESD, is then used to forecast the
• interacting with PAs in both receiving their profiles and
number of Expected Passengers EP of flight ∫ j in the
transmitting the CF recommendation suitably generated j
given interval ∆t , computed as EP∆t = Nj · ∆t f (x)dx,
for their associated travelers (see Section IV) for sup-
where Nj is the expected number of passengers on flight
porting their slack time;
j. The total number of Expected
∑ Passengers in interval ∆t
• interacting with CAs to receive the product categories j
is then given by EP∆t = j EP∆t . Finally, the Slack
available in their associated commercial facility and
Time ST of passenger i during ∆t can be estimated as:
(periodically) the number of devices connected to the
associated hot-spot, the identifiers of all the connected STi = ESDi − LTi (2)
PAs at a given time and those of the PAs assisting a
passenger in a purchase; where LTi is the time spent by passenger i at the check-
• maintaining updated the list of the AF categories of point, while queuing to be processed, which depends on
interest. the expected number of passengers in ∆t, EP∆t and the
expected∑number of passengers for all flights j in ∆t,
III. E STIMATION OF THE T RAVELER ’ S S LACK T IME
N∆t = j Nj .
The recommender algorithm works by considering the slack
time of each traveler, which depends on the arrival time at IV. T HE R ECOMMENDER S YSTEM
the terminal checkpoints with regard to the expected take-
off time of his/her flight. The slack time can be estimated Travelers are supported by personalized suggestions gen-
based on data collected at the security control desks. In fact, erated by an hybrid approach adopting both CB and CF
collecting data coming from the automatic detection of the bar- techniques. In particular, the CB recommendation system are
coded Boarding Pass (BP) (e.g., the barcode reading of paper computed by PAs, while the CF component is generated by
and/or mobile boarding cards) of each passenger identifies TAs. Recommendations take into account information coming
both his/her arrival time at security checkpoints and the time from: i) PA profiles; ii) positions and time spent inside the
he/she enters inside the departure area to take his/her flight, hot-spots ranges; iii) detected purchases; iv) travelers’ slack
other than information on departure and boarding times. time. The recommender process consists of three main steps,
By following [28], an estimate of the slack time can be namely: i) selecting the categories potentially interesting for a
obtained by the Early Scheduled Delay (ESD) measuring the traveler; ii) locating the resources (i.e., commercial facilities)
earliness arrival of passengers at checkpoints, which is here also based on the traveler’s position; iii) generating some
defined as the difference between the BP scan Time (BP T ) personalized suggestions by considering the traveler’s slack
and the Scheduled Boarding Time (SBT ) for the detected time ST .
flight. Therefore, for passenger i his/her ESD is obtained as To realize the CB stage, a PA selects for its owner the first
ESDi = SBTi − BP Ti and may represent an estimate of the m (a system parameter) categories on the basis of a measure
slack time STi for each passenger i. However, the reliability of his/her interest. In particular, the measure of interest for the
of this estimate depends on the nature of BP T : k-th category (i.e., Ik ) is computed as:
• If BP T refers to the time the passenger scans the board-
Ik = (w1 · lk + w2 · T B k + w3 · T W k ) · pk (3)
ing pass just before passing through the metal detector,
the corresponding slack time is very close to ESD and where:
can be estimated as:
• l is set to 1 / 0 if the associated category was selected or
STi ∼
= ESDi = SBTi − BP Ti (1) not by the traveler as a category of his/her interest.
78
• T B (i.e., T W ) is a parameter, belonging to [0.1] ∈ R, accessibility, transparency and scalability [47]; although these
which considers the time T B (i.e., T W ) spent in Blue- RSs are very attractive, only a few systems are really operative
tooth (i.e., Wi-Fi) hot-spot ranges, where there are items given their intrinsic complexity.
belonging to the k-th category, as a rough measure of the The proposed RS is characterized by locating travelers in
interest for thosecategories. T B is computed as: order to identify both their interest and the better facilities
0 if τ1 > T B for them. Different localization schemes (based on wireless
TB = T B/τ2 if τ1 < T B ≤ τ2 connections and a wide range of different sensors) have
1 if τ2 < T B been investigated [48] for understanding shoppers behavior
where τ1 and τ2 are system time thresholds (note that within retail spaces. In [49] a framework that should identify
after a time greater than τ2 the value of T B is set to 1). customers malling behaviors by using smartphones, named
T W is computed in a similar way. MallingSense, is presented. It consists of three steps; customer
• p is set to 0.5 / 1 if an item of that category has been data collection, customer trace extraction, and behavior model
purchased or not. It decreases the value of Ik if an item analysis. MallingSense was positively tested on real data. A
belonging to k-th category has been already purchased in store-type RS for physical stores considering the learned cus-
order to give priority to other categories. tomers’ preference’ and temporal influence is proposed in [50].
• w1 , w2 and w3 are system weights ranging in [0.1] ∈ R, It finds customers’ preferences in physical stores from their
∑3
with w1 ≪ w2 ≪ w3 and i=1 wi = 1. interaction behaviors, non-intrusively generated from WiFi
Similarly, a TA will select, for each PA in its operating logs, confirming that customers preferences are influenced by
range, the first m categories popular among similar travelers intrinsic interests and temporal data. In the same context, [51]
(based on their interest degree measures). The similarity describes a location-aware RS matching customers shopping
between two travelers u and q (i.e., σu,q ) is calculated by needs with location-dependent vendor offers and promotions.
using the Jaccard similarity measure [35] on the basis of The other main considered question is how identifying
|P ∩P | specific travelers’ interests on the basis of their location. In
their associated PA profiles (P ) as σuq = |Puu ∪Pqq | , where the
number of categories shared by u and q is divided by the total this paper, it is proposed to assign the same value of interest
number of unique categories in u and q. to all the categories present in the range of a hot-spot on the
Finally, let X be the set of the m CB and the m CF basis of his/her stop time therein. This solution is due to the
categories selected for each traveler. Then the commercial impossibility of identifying a specific topic of interest. The
facilities where there are items belonging to the selected problem is similar to that of measuring the interest in the
categories will be identified by the PA (by using the data stored topics contained into a visited Web page. In this case, the
in its profile). Based on the current traveler’s position, his/her same measure of interest is assigned to all the topics present
data (i.e., age, sex, job and so on) and his/her slack time ST , in a visited page, for instance, based on the time spent by a
the most relevant personalized suggestions will be presented user on the Web page, its length or a score assigned by the
to the traveler. visitor. A RS using a similar approach is described in [52]
where the visiting time of a Web page is the main parameter
V. R ELATED W ORK to estimate the user’s interest in the instances present therein,
Recommender systems (RSs) have been widely investigated while in [53], [54] the typology of the device exploited in the
in the literature and their contextualization is beyond our aims. page access is also considered.
However, interested readers can refer to [36]–[38]
RSs are generally classified in [39]: (i) Content-based (CB), VI. C ONCLUSIONS AND F UTURE W ORK
based on past users’ interests [40]; (ii) Collaborative Filtering
This paper proposed the design of an agent framework to
(CF), searching people having similar interests [41], [42]; (iii)
support air travelers slack time inside departure terminal. To
Demographic, identifying people belonging to the same demo-
this aim, for each passenger some suggestions about the com-
graphic niche [43]; (iv) Knowledge-based, inferring people’s
mercial opportunities available inside the terminal, potentially
needs and references [44]. However, the most performing RSs
interesting for him/her, are generated. Such recommendations
are usually the hybrid systems [45], which combine several
suitably take into account traveler’s interest, current position
approaches to promote mutual synergies, as in [46].
inside the terminal and slack time.
Another common way to classify RSs is based on the
Currently, an app for the free access to some terminal
adoption of a centralized or distributed architecture. The first
services is in the designing phase. This app also should
one is adopted by many RSs because it is easy to implement
contribute to collect data coming from both travelers and some
given that it exploits a unique server and a unique database to
hot-spots to perform a preliminary check about the potential
perform all the tasks. Many e-Commerce sites like Amazon
feasibility of the proposed framework.
(www.amazon.com) and eBay (www.ebay.com) implement
this type of RS, mainly by combining CB and CF techniques.
ACKNOWLEDGMENT
However, these RSs are affected by efficiency, fault tolerance,
scalability and privacy problems. Differently, distributed RSs This study has been supported by NeCS Laboratory
exploit more computational resources but guarantee openness, (DICEAM, University Mediterranea of Reggio Calabria).
79
R EFERENCES [29] D. Rosaci, G. M. L. Sarnè, and D. Ursino, “A multi-agent model for
handling e-commerce activities,” vol. 2002-Jan. IDEAS, 2002, pp.
[1] G. Francis, I. Humphreys, and J. Fry, “The benchmarking of airport 202–211.
performance,” J. air Transport Manag., vol. 8, no. 4, pp. 239–247, 2002. [30] P. Lops, M. De Gemmis, and G. Semeraro, “Content-based recommender
[2] R. Guimera and L. A. N. Amaral, “Modeling the world-wide airport systems: State of the art and trends,” in Recommender systems handbook.
network,” The European Physical J., vol. 38, no. 2, pp. 381–385, 2004. Springer, 2011, pp. 73–105.
[3] A. Kazda and R. Caves, Airport design and operation. Emerald, 2010. [31] X. Su and T. M. Khoshgoftaar, “A survey of collaborative filtering
[4] M. N. Postorino and F. G. Praticò, “An application of the multi-criteria techniques,” Advances in artificial intelligence, vol. 2009, p. 4, 2009.
decision-making analysis to a regional multi-airport system,” Research [32] P. De Meo, D. Rosaci, G. M. L. Sarnè, D. Ursino, and G. Terracina,
in Transportation Business & Management, vol. 4, pp. 44–52, 2012. “Ec-xamas: supporting e-commerce activities by an xml-based adaptive
[5] M. N. Postorino, L. Mantecchini, and F. Paganelli, “Green airport invest- multi-agent system,” Applied Artificial Intelligence, vol. 21, no. 6, pp.
ments to mitigate externalities: Procedural and technological strategies,” 529–562, 2007.
in Sustainable Entrepreneurship and Investments in the Green Economy. [33] D. Rosaci and G. M. L. Sarnè, “Masha: A multi-agent system handling
IGI Global, 2017, pp. 231–256. user and device adaptivity of web sites,” User modeling and user-
[6] C. Malandri, L. Mantecchini, and M. N. Postorino, “Airport ground adapted interaction, vol. 16, no. 5, pp. 435–462, 2006.
access reliability and resilience of transit networks: a case study,” [34] P. De Meo, D. Rosaci, G. M. L. Sarnè, G. Terracina, and D. Ursino,
Transportation Research Procedia, vol. 27, pp. 1129–1136, 2017. “An xml-based adaptive multi-agent system for handling e-commerce
[7] M. Milde, Int. air law and ICAO. Eleven Int. Pub., 2008, vol. 4. activities,” in Web Services-ICWS-Europe ’03. Springer, 2003, pp. 152–
[8] B. Edwards, The modern airport terminal: New approaches to airport 166.
architecture. Taylor & Francis, 2004. [35] Grefenstette G., Explorations in Authomatic Thesaurus Construction.
[9] N. Ashford, P. Coutu, and J. Beasley, Airport operations, 2013. Hingham, MA, USA: Kluwer Academic Pub., 1994.
[10] D. Fodness and B. Murray, “Passengers’ expectations of airport service [36] G. Adomavicius and A. Tuzhilin, “Toward the next generation of
quality,” J. of Services Marketing, vol. 21, no. 7, pp. 492–506, 2007. recommender systems: A survey of the state-of-the-art and possible
[11] Y. Lin and C. Chen, “Passengers’ shopping motivations and commercial extensions,” IEEE trans. on knowledge and data engineering, vol. 17,
activities at airports–the moderating effects of time pressure and impulse no. 6, pp. 734–749, 2005.
buying tendency,” Tourism Management, vol. 36, pp. 426–434, 2013. [37] F. Ricci, L. Rokach, and B. Shapira, “Introduction to recommender
[12] K. A. Mew, P. C. Alexander, I. Bergera, M. Chanes, G. Crysler, K. Dana, systems handbook,” in Recommender systems handbook. Springer,
and M. Desai, Airport financial management. Ashgate, 2012. 2011, pp. 1–35.
[13] L. Palopoli, D. Rosaci, and G. M. L. Sarnè, “A multi-tiered recom- [38] X. Yang, Y. Guo, Y. Liu, and H. Steck, “A survey of collaborative
mender system architecture for supporting e-commerce,” in Intelligent filtering based social recommender systems,” Computer Comm., vol. 41,
Distributed Computing VI. Springer, 2013, pp. 71–81. pp. 1–10, 2014.
[14] M. N. Postorino and G. M. L. Sarnè, “A neural network hybrid [39] R. Burke, “Hybrid Web Recommender Systems,” in The Adaptive Web,
recommender system,” in Proceedings of the 2011 conference on neural ser. LNCS, vol. 4321. Springer, 2007, pp. 377–408.
Nets WIRN10, 2011, pp. 180–187. [40] P. Lops, M. Gemmis, and G. Semeraro, “Content-based Recommender
[15] L. Palopoli, D. Rosaci, and G. M. L. Sarnè, “Introducing specialization Systems: State of the Art and Trends,” in Recommender Systems
in e-commerce recommender systems,” Concurrent Engineering, vol. 21, Handbook. Springer, 2011, pp. 73–105.
no. 3, pp. 187–196, 2013. [41] J. Breese, D. Heckerman, and C. Kadie, “Empirical Analysis of Predic-
[16] S. Takakuwa and T. Oyama, “Modeling people flow: simulation analysis tive Algorithms for Collaborative Filtering,” in Proc. 14th Int. Conf. on
of international-departure passenger flows in an airport terminal,” in Uncertainty in Artificial Intell. Morgan Kaufmann, 1998, pp. 43–52.
Proc. 35th conf. on Winter simulation: driving innovation. Winter [42] X. Su and T. Khoshgoftaar, “A Survey of Collaborative Filtering
Simulation Conf., 2003, pp. 1627–1634. Techniques,” Adv. in Artif. Intell., vol. 2009, pp. 4:2–4:2, 2009.
[17] M. L. Tam, W. H. Lam, and H. P. Lo, “Modeling air passenger travel [43] C. Stiller, F. Ross, and C. Ament, “Demographic Recommendations
behavior on airport ground access mode choices,” Transportmetrica, for WEITBLICK, an Assistance System for Elderly,” in Comm. and
vol. 4, no. 2, pp. 135–153, 2008. Information Technologies, 2010 Int. Symp. . IEEE, 2010, pp. 406–411.
[18] K. Artur and K. Tomasz, “A model of check-in system management [44] S. Trewin, “Knowledge-based recommender systems,” Encyclopedia of
to reduce the security checkpoint variability,” Simulation Modelling Library and Information Science, vol. 69, no. Suppl. 32, p. 69, 2000.
Practice and Theory, vol. 74, pp. 80–98, 2017. [45] R. Burke, “Hybrid Recommender Systems: Survey and Experiments,”
[19] K. Yoo and Y. Choi, “Analytic hierarchy process approach for identifying UMUAI, vol. 12, no. 4, pp. 331–370, 2002.
relative importance of factors to improve passenger security checks at [46] D. Rosaci and G. M. L. Sarnè, “Multi-agent technology and ontologies
airports,” J. of Air Transport Manag., vol. 12, no. 3, pp. 135–142, 2006. to support personalization in b2c e-commerce,” Electronic Commerce
[20] K. W. Bowyer, “Face recognition technology: security versus privacy,” Research and Applications, vol. 13, no. 1, pp. 13–23, 2014.
IEEE Technology and society magazine, vol. 23, no. 1, pp. 9–19, 2004. [47] A. S. Tanenbaum and M. Van Steen, Distributed systems: principles and
[21] G. M. L. Sarnè, “A collaborative filtering recommender exploiting a paradigms. Prentice-Hall, 2007.
som network,” in Recent Advances of Neural Network Models and [48] S. Rallapalli, “Mobile localization: approach and applications,” Ph.D.
Applications. Springer, 2014, pp. 215–222. dissertation, 2014.
[22] F. Ohrtman and K. Roeder, Wi-Fi handbook: Building 802.11 b wireless [49] S. Lee, C. Min, C. Yoo, and J. Song, “Understanding customer malling
networks. McGraw-Hill New York, NY, 2003, vol. 67. behavior in an urban shopping mall using smartphones,” in Proc of ACM
[23] B. A. Miller and C. Bisdikian, Bluetooth revealed: the insider’s guide conf. on Pervasive and ubiquitous comp. ACM, 2013, pp. 901–910.
to an open specification for global wireless communication. Prentice [50] Y. Chen, Z. Zheng, S. Chen, L. Sun, and D. Chen, “Mining customer
Hall PTR, 2001. preference in physical stores from interaction behavior,” IEEE Access,
[24] S. A. M. Saleh, S. A. Suandi, and H. Ibrahim, “Recent survey on crowd vol. 5, pp. 17 436–17 449, 2017.
density estimation and counting for visual surveillance,” Engineering [51] J. Bao, Y. Zheng, and M. Mokbel, “Location-based and preference-aware
Applications of Artificial Intelligence, vol. 41, pp. 103–114, 2015. recommendation using sparse geo-social networking data,” in Proc. of
[25] A. Andreoni and M. N. Postorino, “Time series models to forecast air conf. on advances in geographic inform. sys. ACM, 2012, pp. 199–208.
transport demand: a study about a regional airport,” IFAC Proceedings [52] J. Parsons, P. Ralph, and K. Gallagher, “Using Viewing Time to Infer
Volumes, vol. 39, no. 12, pp. 101–106, 2006. User Preference in Recommender Systems,” in AAAI Work. on Semantic
[26] M. Postorino, “A comparison among different approaches for the evalu- Web Personalization. Menlo Park, CA, USA: AAAI., 2004, pp. 52–64.
ation of the air traffic demand elasticity,” WIT Transactions on Ecology [53] D. Rosaci, G. M. L. Sarnè, and S. Garruzzo, “Muaddib: A distributed
and the Environment, vol. 67, 2003. recommender system supporting device adaptivity,” ACM Transactions
[27] R. De Neufville, “Airport systems planning and design,” Air Transport on Information Systems (TOIS), vol. 27, no. 4, p. 24, 2009.
Management: An International Perspective, p. 61, 2016. [54] D. Rosaci and G. M. L. Sarnè, “A multi-agent recommender system
[28] M. N. Postorino, L. Mantecchini, C. Malandri, and F. Paganelli, “Airport for supporting device adaptivity in e-commerce,” Journal of Intelligent
passenger arrival process: Estimation of earliness arrival functions,” Information Systems, vol. 38, no. 2, pp. 393–418, 2012.
Transportation Research Procedia, In Press.
80