A multi-layer model for diffusion of urgent information in mobile networks

doi:10.1016/j.jocs.2017.02.002

Journal of Computational Science

Volume 20, May 2017, Pages 129-142

https://doi.org/10.1016/j.jocs.2017.02.002 Get rights and content

Highlights

•
The multi-layer model of information spreading containing three layers is proposed.
•
Three types of agents, identified using two real datasets, represent regular people, active people (organizers) and busy people.
•
Information spreading uses susceptible-infected model for the information transfer between agents.
•
Model has two scenarios of execution − urgent and normal.
•
Experiments show reasonable results close to the one of other researches with a number of ways for improvement.

Abstract

Information spreading analysis is an important problem in the scientific community and is widely studied today using different techniques, from the data analysis to the agent-based modelling. For some extreme situations, like fire or flood, there is little or no reliable information about users’ activity available. That is why an efficient simulation of the urgent scenarios is very important, because analysis of the simulated data can help to provide fast and accurate reaction and save human lives. In this paper, we present a multi-layer agent-based network model for the information diffusion simulation in the urgent scenarios, which allows to investigate agents’ behavior in a variety of situations in the absence of the real data. This model can be used for the urban scenarios simulation in the integration with other agent-based human interaction models.

Experimental: results demonstrate good results in comparison with existing works in this area and give a number of insights regarding the further model development.

Introduction

Today it is hard to imagine our life without cell phones. People make calls every day all around the world. Despite the fact that every person can call anyone, whose phone number he or she knows or has in the contact list, people often have a certain social circle − relatively small group of people, whom they call regularly [1]. Members of this group may differ by their roles and number from one person type to another: family members and some friends for the ordinary people; a lot of friends and colleagues for very communicative people; business partners and clients for the businessmen, etc. If we know social circle of a specific person, we might assume that we can quite accurately predict, whom this person will call in the particular situation. However, there are several factors that should be taken into consideration: cellular network and mobile device may have their own impact from the point of accessibility; social circle of the person is also dynamic, reflecting changes in the life state of the person (family, work, education, etc.). Taking all this together, we can say that the high-quality simulation of the calling activity is a very hard problem. But, using the call details records (CDR) provided by telecom companies, researchers are able to analyze the behavior of users in the mobile networks [2], [3] and provide viable information for building mobile networks models. Even more complicated and interesting problem is analyzing and simulation of the information diffusion in the cellular networks. This property cannot be directly obtained from the large mobile networks, because we cannot ask all participants whether or not did they say some information to the other user. That is why the only option for investigation of the information spreading on the large real networks is building the network of user-to-user interactions and usage of various spreading models on them [4], [5]. But when it comes to the analysis of the people behavior in terms of the calling activity in extreme situations, like flood, fire or terrorist attacks, we have even less options, because it is very hard to gather CDRs for the period of some extreme event [6]. All the above along with the need to accurately predict the consequences of the extreme events leads to the idea of the development of a flexible and precise model for the information diffusion simulation to analyze various urgent scenarios in the absence of the real data. In our previous work [7] we presented the agent-based multi-layer model for the information spreading in the different scenarios. The main contributions of this paper are: more thorough analysis of the available datasets for agents types identification; simplification of the model to eliminate numerous assumptions and arbitrary parameters; and comparison of the modelling results with the existing researches in the area of information spreading in the mobile networks.

Section snippets

Problem statement

Information diffusion in the mobile networks is a very complex process, which involves a great number of various aspects, from the hardware communications between mobile devices through the cell towers to interpersonal relations between people, who make calls and disseminate the information. All these aspects are connected one with another and create a sophisticated structure, which can be represented using layers (Fig. 1), where each layer stands for a specific aspect with its specific

Social networks analysis

Analysis of the information about the characteristics and evolution of social networks has got a lot of attention throughout years, as it can provide insights about the human behavioral patterns. The analysis of several datasets representing different types of the social networks, like the movie actors collaboration or citation network, was discussed in [3]. A network of phone calls made during August 10, 1998 was examined in this work. Dataset contains over 53 million nodes. Values of the

Model description

In our work for the processes of making calls and information transfer we used a multi-agent simulation due to its flexibility and the ability to accurately represent complex systems [26], [40], [41]. The core of this approach is a set of agent types, their characteristics and rules of agents’ interaction. Thus to create a proper multi-agent model, we need to describe agents’ parameters, identify several types of agents, define rules of their behavior and specify a number of global model

Urgent situation scenario

To investigate the information spreading process in the developed model for the urgent situations, like information about floods or terrorist attacks, we made significant changes into the calling process of the agents. On the step of selection the person from the contacts list to call, agents now select contacts by their call probability, which means that agents now call people, who are important for them, in the first place. Agents also remember the contacts they already called to and do not

Experiments

To compare the results produced by our model with the results obtained by other researchers in the area of information diffusion in the cellular networks, we conducted a set of experiments, where we varied the network size and measured the speed of the information spreading in the network. We refer to the paper by Karsai et al. [60], which is dedicated to the investigation of the SI spreading dynamics with simulations using the event sequences. The paper uses the same class of spreading models

Conclusion and future works

In this work we presented a multi-layer agent-based model for the information spreading in the mobile network. The main idea behind the concept of the model is to try to consider all levels involved into dissemination of the information − hardware levels (cell towers and mobile devices) and social interaction levels (contact network, calls network and information spreading model). In this work we took into account only social interactions levels, development and investigation of hardware levels

Acknowledgement

This paper is financially supported by The Russian Scientific Foundation, Agreement #14-11-00823 (15.07.2014)

Alexander Visheratin is a Ph.D. student at ITMO University (Saint Petersburg) since 2014, where he participates in the designing of workflow scheduling algorithms as well as in the development of big data processing systems. He got his specialist's Degree in informational systems in the ITMO University in 2014. Alexander's research interests are distributed computing, data mining and data analysis, and agent simulation of social processes.

References (61)

A.A. Visheratin et al.
Urgent information spreading multi-layer model for simulation in mobile networks
Procedia − Procedia Comput. Sci.
(2016)
L.M. Bettencourt et al.
The power of a good idea: quantitative modeling of the spread of ideas from epidemiological models
Phys. A Stat. Mech. Appl.
(2006)
M. Safar et al.
Network robustness and irreversibility of information diffusion in Complex networks
J. Comput. Sci.
(2011)
G. Wang et al.
A computational model for multi-agent E-commerce negotiations with adaptive negotiation behaviors
J. Comput. Sci.
(2013)
A.K. Jain
Data clustering: 50 years beyond K-means
Pattern Recognit. Lett.
(2010)
P.J. Rousseeuw
Silhouettes: a graphical aid to the interpretation and validation of cluster analysis
J. Comput. Appl. Math.
(1987)
C. Liu et al.
Information spreading on dynamic social networks
Commun. Nonlinear Sci. Numer. Simul.
(2014)
A. Karsakov et al.
Toolbox for visual explorative analysis of complex temporal multiscale contact networks dynamics in healthcare
Procedia Comput. Sci.
(2016)
R. Kraut et al.
Internet paradox: a social technology that reduces social involvement and psychological well-being?
Am. Psychol.
(1998)
C. Herrera-Yagüe et al.
The anatomy of urban social networks and its implications in the searchability problem
Sci. Rep.
(2015)

R. Albert et al.

Statistical mechanics of complex networks

Rev. Mod. Phys.

(2002)

J.-P. Onnela et al.

Structure and tie strengths in mobile communication networks

Proc. Natl. Acad. Sci. U. S. A.

(2007)

M. Kivelä et al.

Multiscale analysis of spreading in a large communication network

J. Stat. Mech. Theory Exp.

(2012)

F. Aleissa et al.

Wired to connect: analyzing human communication and information sharing behavior during extreme events

Kdd-Lesi

(2014)

A. McDiarmid S. Bell J. Irvine J. Banford Nodobo: Detailed Mobile Phone Usage Dataset...

V.D. Blondel et al.

A survey of results on mobile phone datasets analysis

EPJ Data Sci.

(2015)

J. Onnela et al.

Analysis of a large-scale weighted network of one- to-one human communication

New J. Phys.

(2007)

A. a Nanavati et al.

On the structural properties of massive telecom call graphs: findings and implications

Proc. 15th ACM Int. Conf. Inf. Knowl. Manag.

(2006)

P. Erdös et al.

On random graphs

Publ. Math.

(1959)

S.H. Strogatz

Exploring complex networks

Nature

(2001)

D.J. Watts et al.

Collective dynamics of ‘small-world’ networks

Nature

(1998)

A.-L. Barabási et al.

Emergence of scaling in random networks

Science (80-)

(1999)

W. Aiello et al.

A random graph model for massive graphs

Proc. Thirty-second Annu. ACM Symp. Theory Comput. − STOC ’

(2000)

E.M. Royer

An analysis of the optimum node density for ad hoc mobile networks

IEEE Int. Conf. Commun.

(2001)

S. Zeadally et al.

Vehicular ad hoc networks (VANETS): status, results, and challenges

Telecommun. Syst.

(2010)

A. Khelil et al.

An epidemic model for information diffusion in MANETs

Proc. 5th ACM Int. Work. Model. Anal. Simul. Wirel. Mob. Syst. − MSWiM ’02

(2002)

K. Lerman et al.

Information contagion: an empirical study of the spread of news on digg and twitter social networks

Proc. Fourth Int. AAAI Conf. Weblogs Soc. Media

(2010)

W. Galuba et al.

Outtweeting the twitterers − predicting information cascades in microblogs

Proc. 3rd Conf. Online Soc. Networks

(2010)

W.M. Tam et al.

Complex-network modeling of a call network

IEEE Trans. Circuits Syst. I Regul. Pap.

(2009)

L. Lü et al.

The small world yields the most effective information spreading

New J. Phys.

(2011)

Cited by (12)

An agent-based model to simulate inhabitants’ behavior during a flood event
2021, International Journal of Disaster Risk Reduction
Citation Excerpt :
The combination of flood and inhabitant behavior, is not the key element of these models. For instance Ref. [37], provided an ABM to simulate the information spreading in the mobile network. The model is not specifically dedicated to flood, but can be used facing this risk [20].
Flood management is particularly important for many urban territories. In order to assess the interest of different risk mitigation strategies, it is necessary to consider the human behavior. Different models and approaches are available to model and simulate people behaviors during a flood event. Despite the considerable recent progress of these models, none of them succeeds in answering all the required challenges: (1) simulate the flood event, (2) integrate geographical information, (3) take into account complex behaviors for inhabitants considering emotion and partial knowledge, (4) enable different strategies for inhabitants, (5) take into account the degradation of infrastructures and its impact on their functioning, and (6) ensure the genericity and flexibility of the model in order to be able to apply it to any territory. In order to meet these challenges, we provide a new Agent based Model, called SiFlo. In this paper, we present this model and an application to La Ciotat (France).
Social awareness of crisis events: A new perspective from social-physical network
2020, Cities
Citation Excerpt :
In terms of social network topology, Huang, Shen, and Cheng (2016) proposed a super network model to describe the topology of public opinion network based on the theory of complex networks. Visheratin, Trofimenko, Mukhina, Nasonov, and Boukhanovsky (2017) proposed a time-delay evolution model in social network to analyze the interaction mechanism between two social events. As for community detection, which is used to identify meaningful subgroups within a complex network, also plays an important role in extracting the evolution information of crisis events from the social network.
The emergence of social media provides a rich source of posts to facilitate situational awareness and management during crisis events. Since interactions within social media users are frequently used to establish social networks, previous studies have explored how social media contributes to the information diffusion about crisis events. However, most studies have concentrated on the dynamics of the online social network, whereas the information diffused within social space and physical world during different event evolution stages tends to be overlooked. To address this problem, this study introduces a methodology for modeling and analyzing the information diffusion of crisis events based on the retweet patterns of social media users at city level. The proposed methodology consists of two parts: 1) Fisher ordinal clustering method is applied to divide the evolution stages of crisis events based on social media user activities. 2) To explore the information diffusion of crisis events in the physical world, the social-physical network (SPnet) is constructed by extending the retweet network to the geographical space for each evolution stage. The Weibo data related to the Kunming terrorist attack in 2014 are analyzed to demonstrate the methodology. By identifying the geo-location of users (90.8% of the total active users), two types of SPnet are constructed based on the retweet patterns of ordinary users and media users, respectively. The results show that the information diffusion in the SPnet of ordinary users are mainly from Kunming, while in the SPnet of media users, major developed cities are the key nodes for information diffusion. This study provides useful insights for understanding how human activities react to crisis events through the social-physical network at city level.
A novel negative feedback information dissemination model based on online social network
2019, Physica A: Statistical Mechanics and its Applications
Citation Excerpt :
The traditional network evolution analysis method has failed to meet the requirements. There is an urgent need for new models and analytics for large-scale social networks [11–15]. An important research direction of the social network evolution calculation method is to microscopically model the users of social networks.
Due to the existing social network information dissemination model does not consider the influencing factors of node feedback information. The characteristics of social networks and the social attributes of the disseminators are analyzed; we propose a new social network information dissemination with negative feedback NFSIR (negative feedback susceptible infected removed) model combined with traditional epidemiological models The model first introduces attenuation coefficients and noise coefficients to describe the different effects of information transmission among different users. The feedback function is defined in order to describe the information feedback mechanism. Then, according to the mechanism of information interaction evolution, an information dissemination tree is constructed, and a differential equation of propagation dynamics is established. It reveals the complex interactions and interactions between user relationships, social communities, and cyberspace information in social networks. Finally, the model is applied to two typical real social networks to carry out simulation experiments and compared with the traditional model. Experimental results show that, the intensity of information feedback has a significant impact on the process of information dissemination. It can be seen that the Twitter and Sina microblog networks are significantly different through the analysis of the transmission time and propagation life cycle indicators. The comparison of experimental results shows that the proposed NFSIR model can better reflect the characteristics of real social networks. It proves that the proposed mathematical propagation model is objective, reasonable, and effective. The proposed model not only has strong scalability, but also has application value. It also provides theoretical support for research in related fields.
Modeling information dissemination and evolution in time-varying online social network based on thermal diffusion motion
2018, Physica A: Statistical Mechanics and its Applications
Due to the traditional information dissemination model cannot accurately simulate information dissemination process in the real world, this paper proposes an information diffusion mathematical model and information state node evolution mechanism by using thermodynamic molecular thermal diffusion motion theory, combined with epidemic infection model. Four different network topologies (regular network, small-world network, random network and scale-free network) are applied in the time-varying online social network (OSN) of information dissemination process. Information entropy is also introduced in the information dissemination of OSN. Information is essentially a special form of matter; the propagation process is a process in which the system transitions from one stable state to another. A transfer function is built by some information parameters such as information energy, information temperature, and energy entropy. It reveals the relationship between the state of microscopic network nodes and the macro iterative evolution rules, and carries out simulation experiments and empirical comparative experiments in a variety of networks with different topological structures. The proposed model is trained and evaluating using truth experimental data collected in Baidu network. The experimental results show that the similarity between the simulation results and the real data is greater than 0.96, the correlation is greater than 0.95, and the peak value of the local error is less than 0.2. The mathematical model accurately describes the internal laws and mechanisms of the information dissemination behavior and proves the proposed mathematical propagation model. The transfer function and evolution mechanism are reasonable and effective; the proposed information propagation model not only has strong extensibility, but also provides theoretical support for research in related fields.
Data through the Computational Lens
2017, Journal of Computational Science
Today, many advanced studies in computational science are enforced by data collected and processed in distributed systems, obtained and assimilated in real (or near-real) time, used to identify and build models, or even used for advanced visualization to have insights through visual images. Taking such an important role of data into account, 16th International Conference on Computational Science (ICCS 2016), an event that promotes leading-edge research in the area, proposed the special theme “Data through the Computational Lens” to bring diverse ideas and recent developments together from computational science society aimed towards this vector. This special issue contains extended papers selected from the conference proceedings that promotes leading edge research.
Combining the Information Diffusion Model for the Propagation of Zhouyi Ideology in the New Era
2024, Applied Mathematics and Nonlinear Sciences

View all citing articles on Scopus

Tamara B. Trofimenko is a graduate of ITMO University. She got a double master’s degree at University of Amsterdam and ITMO University with a thesis dedicated to the development of agent-based information spreading model. Her major research interests focus on machine learning, data analysis and agent-based simulation.

Ksenia Mukhina is a Ph.D. student at ITMO University (Saint Petersburg). She received a M.S. in 2016 from ITMO University at Saint Petersburg with a thesis dedicated to visual analysis of dynamic processes in temporal networks. Her research interests are in the areas of complex network visualization, map visualization, and geographical information systems.

Denis Nasonov is a researcher at eScience Research Institute at ITMO University (Saint-Petersburg). He is leading infrastructure research group, which is focused on several fields, including distributed computation systems, cloud computing, optimization scheduling algorithms, data placement algorithms, meteorological simulation, complex network processes.

Alexander V. Boukhanovsky is the Chair of High Performance Computing (HPC) Department in ITMO University. In 2005, he defended his dissertation on Concurrent Software Statistical Measurements of Spatial-Temporal Fields. Since 2006 he has been working as a Professor of Information Systems and Head of the Parallel Software lab in ITMO University. In 2007 he created the eScience Research Institute. His research interests are high-performance computing, computer modelling of complex systems, intelligent computational technologies, distributed environments for multidisciplinary researches, statistical analysis and simulation in marine sciences.

View full text

A multi-layer model for diffusion of urgent information in mobile networks

Highlights

Abstract

Introduction

Section snippets

Problem statement

Social networks analysis

Model description

Urgent situation scenario

Experiments

Conclusion and future works

Acknowledgement

Procedia − Procedia Comput. Sci.

Phys. A Stat. Mech. Appl.

J. Comput. Sci.

J. Comput. Sci.

Pattern Recognit. Lett.

J. Comput. Appl. Math.

Commun. Nonlinear Sci. Numer. Simul.

Procedia Comput. Sci.

Internet paradox: a social technology that reduces social involvement and psychological well-being?

Am. Psychol.

The anatomy of urban social networks and its implications in the searchability problem

Sci. Rep.

Statistical mechanics of complex networks

Rev. Mod. Phys.

Structure and tie strengths in mobile communication networks

Proc. Natl. Acad. Sci. U. S. A.

Multiscale analysis of spreading in a large communication network

J. Stat. Mech. Theory Exp.

Wired to connect: analyzing human communication and information sharing behavior during extreme events

Kdd-Lesi

A survey of results on mobile phone datasets analysis

EPJ Data Sci.

Analysis of a large-scale weighted network of one- to-one human communication

New J. Phys.

On the structural properties of massive telecom call graphs: findings and implications

Proc. 15th ACM Int. Conf. Inf. Knowl. Manag.

On random graphs

Publ. Math.

Exploring complex networks

Nature

Collective dynamics of ‘small-world’ networks

Nature

Emergence of scaling in random networks

Science (80-)

A random graph model for massive graphs

Proc. Thirty-second Annu. ACM Symp. Theory Comput. − STOC ’

An analysis of the optimum node density for ad hoc mobile networks

IEEE Int. Conf. Commun.

Vehicular ad hoc networks (VANETS): status, results, and challenges

Telecommun. Syst.

An epidemic model for information diffusion in MANETs

Proc. 5th ACM Int. Work. Model. Anal. Simul. Wirel. Mob. Syst. − MSWiM ’02

Information contagion: an empirical study of the spread of news on digg and twitter social networks

Proc. Fourth Int. AAAI Conf. Weblogs Soc. Media

Outtweeting the twitterers − predicting information cascades in microblogs

Proc. 3rd Conf. Online Soc. Networks

Complex-network modeling of a call network

IEEE Trans. Circuits Syst. I Regul. Pap.

The small world yields the most effective information spreading

New J. Phys.