Search

The introduction of trust-based approaches in social scenarios modeled as multi-agent systems (MAS) has been recognized as a valid solution to improve the effectiveness of these communities. In fact, they make interactions taking place in social scenarios much fruitful as possible, limiting or even avoiding malicious or fraudulent behaviors, including collusion. This is also the case of multi-layered neural networks (NN), which can face limited, incomplete, misleading, controversial or noisy datasets, produced by untrustworthy agents. Many strategies to deal with malicious agents in social networks have been proposed in the literature. One of the most effective is represented by Eigentrust, often adopted as a benchmark. It can be seen as a variation of PageRank, an algorithm for determining result rankings used by search engines like Google. Moreover, Eigentrust can also be viewed as a linear neural network whose architecture is represented by the graph of Web pages. A major drawback of Eigentrust is that it uses some additional information about agents that can be a priori considered particularly trustworthy, rewarding them in terms of reputation, while the non pre-trusted agents are penalized. In this paper, we propose a different strategy to detect malicious agents which does not modify the real reputation values of the honest ones. We introduce a measure of effectiveness when computing reputation in presence of malicious agents. Moreover, we define a metric of error useful to quantitatively determine how much an algorithm for the identification of malicious agents modifies the reputation scores of the honest ones. We have performed an experimental campaign of mathematical simulations on a dynamic multi-agent environment. The obtained results show that our method is more effective than Eigentrust in determining reputation values, presenting an error which is about a thousand times lower than the error produced by Eigentrust on medium-sized social networks.

Artificial Neural Networks (NNs) have been recognized as a powerful tool for automatically learning complex relationships in data. In this paper, we propose to apply such a tool for modeling forest regeneration, a possibility not yet investigated in the literature. In order to evaluate the capability of NNs to simulate initial recruitment of pine species in Mediterranean forests, a feed-forward multi-layer neural network has been applied to seed germination and seedling survival of four pine species under three soil conditions, with or without seed protection, in Castilla La Mancha (Central-Eastern Spain). The experimental campaign has shown good performance in predicting the two pine initial recruitment stages. The proposed approach may help to predict the success of natural regeneration in Mediterranean pine forests under different basal areas and management strategies.

The Internet of Things (IoT) promises to change many aspects of our daily lives thanks to the opportunity of interconnecting a massive numbers of smart objects with increasing computational, storage, communication, and power capabilities, in such a way making “smart” and “interactive” most of the world around us. In other words, smart objects and humans will be involved together in pervasive, proactive, and collaborative activities to orchestrate and execute increasingly complex and sophisticated tasks. In such a scenario, thanks also to a greater ecological awareness of people, a promising field of application for IoT technology is the monitoring of natural habitats in automatic or semiautomatic mode. A potentially effective and efficient solution is to form composite teams joining human operators and IoT devices. In such teams, not only the kind of team members will be different, i.e., humans and IoT devices, but also the IoT devices will be heterogeneous among them in terms of characteristics and performance. However, a basic requirement for a good team is the existence of high levels of mutual trust among its members. In other words, for the formation of a good team, it becomes of primary relevance to know and adequately represent the trustworthiness of the individual team members. To this end, our contribution can be summarized as (i) introducing a trust measure that takes into account both the reputation of devices and the accuracy of their measures; (ii) designing a framework that based on the proposed trust measure forms temporary teams for environmental ecosystems made of humans and IoT devices; and (iii) testing the proposed framework by simulating a collaborative environmental monitoring activity. The simulation results confirmed the advantages of the proposed approach in terms of performance and appreciation of the composite temporary teams that have been formed in this way.

The blockchain paradigm is currently modifying all the major Internet of Things (IoT) application domains, giving the opportunity of constructing decentralized environments in which trustful and anonymous activities can be efficiently performed. Blockchain proposes an approach which assures data saved on a distributed ledger would be continuously synchronized and in such a way the ledger will remain consistent. The distributed ledger has the responsibility to maintain a higher level of consistency. Given a connected network, the Optimum Neighbor Selection (ONS) of paths can be obtained by finding the Minimum Spanning Tree of the network. However, none of past approaches proposed in the literature to construct ONS considered the problem of having nodes with low levels of reliability in the network and even malicious or fraudulent nodes, that is a situation very common in an IoT environment. In this paper, we propose an optimized blockchain ONS algorithm, called Trust-based ONS (TONS), which allows the miners to communicate with a globally-optimized selection of neighbors. The algorithm can also guarantee that these nodes are the most reliable miners. We also describe an experimental simulation campaign we have performed to evaluate the effectiveness and the efficiency of our approach.

The introduction of the IoT technology and its pervasive penetration in our daily life implies that IoT smart objects can often participate in social events, and in this case their paradigm of interaction is commonly denoted as social IoT (SIoT). To make reliable transactions into a SIoT scenario, in this article, we introduce a multi-agent SIoT architecture, which integrates a reputation system based on a clustering of the SOs. In our framework, when a SO looks for a resource and detects a reliable partner having that resource, then the two SOs can interact to make a transaction, and at the end of the transaction each of them provides a feedback about the partner to the local reputation system of the edge server, to update the SOs' reputation scores. Moreover, each edge server sends these reputation scores to the cloud, which updates the associated reputation values of the SOs, deriving from experiences coming from all the edge domains. This architecture provides a given object, moving from an edge domain to another one, with the possibility to have an updated value of its reputation, represented by the value stored in the cloud. We have validated our approach by a campaign of simulations, whose results seem particularly promising.

In the near future, the growth of personal mobility is expected to continue all over the world and to cause impacts such as increased levels of traffic congestion and worsened quality of life, mainly in highly populated urban areas. Alternative solutions for individual mobility have been promoted to dissuade people from using private cars. Particularly, renting private vehicles scarcely used by their owners would result in several benefits, including environmental ones, and traffic reduction. This type of solution, based on the willingness of individuals to rent their own vehicles, is called Peer-to-Peer Car-Sharing (P2P-CS). This study proposes a methodological approach focused on the adoption of both a reputation system and blockchain technology to support P2P-CS activities. Reputation scores are introduced to meet CS actors’ expectations of dealing with trusted partners, by allowing both access to service and the opportunity to customize CS fares, while the blockchain makes reputation scores publicly accessible and unchangeable and allows the P2P-CS transport solution to be managed without third parties. The effectiveness of the proposed approach has been verified by several tests carried out on real and simulated data. The obtained results are satisfactory and encourage the adoption of these further sustainable travel mode opportunities.

In real and virtual communities, complex and sophisticated forms of social interactions and cooperation are increasingly taking place between heterogeneous actors such as people, intelligent objects and virtual entities. At the same time, in such communities the risks of interacting with unreliable partners engaged in malicious activities increase. Therefore, it is important to provide all players with adequate information in order to allow them to choose the most reliable partner to interact with. In this context, we have focused our attention on colluding activities. In particular we propose a reputation method that preliminarily identifies those actors who most likely can be considered colluding players. This method does not introduce side effects to trust scores of honest actors while detecting colluding ones with high accuracy. A simple example supports our results.

In this paper, we deal with the possibility of simulating dynamic polarization and proximity depolarization processes in a software multi-agent community, modeling homophily and trust relationships usually present in human processes. Group polarization involves various disciplines, such as economics, social psychology, political science, sociology and many others, and it can be considered a critical process underlying relevant behaviors as, for instance, voting and conflictual intergroup relations in the society. Moreover, being a human social phenomenon, the polarization processes are subject to change over time or also effect overturning (i.e., depolarization). Our contribution consists of proposing a compactness-based model for equipping agents in order to simulate the complexity of such processes. We have simulated two case studies where polarization is ruled by compactness measures, combining similarity and trust with different percentages. We evaluated the results provided by compactness measures in order to verify the role of similarity and trust in the agent polarization processes.

A key issue to realize community of agents involving social aspects is that of modeling trustworthiness between the actors of the society and, to this purpose, many trust and reputation models have been proposed in the past. Most of these proposals focused on representing trust dimensions using apposite scalar measures, and integrating these measures in synthetic indicators of trustworthiness, possibly collected into a trust vector. We highlight as this choice is an evident limitation and to overcome it, we propose a new model of trust and reputation for a community of social agents, where the trust perceived by an agent about another agent is modeled by a directed, weighted graph whose nodes and edges represent trust dimensions and their relationships, respectively. This way, we can represent also those situations in which an agent does not knows a given trust dimension, e.g. the honesty, but it is capable to derive it from other correlated dimensions, e.g. the reliability. Our model, called T-pattern, has been specifically designed to represent any situation in which many different trust dimensions are mutually dependent. We also introduce the notion of T-Pattern Network (TPN) as an integrated framework to represent both the trust and reputation values as well as the dependencies between trust dimensions for all the pairs of agents.

In this work we introduce a clustering reputation-based approach for edge-based IoT (Internet of Things) environments where it’s possible to process and save data on edge servers. In particular, once in an edge domain two Smart IoT Objects (SOs) exchanged a service, supported by their associated agents, they will release a feedback about their respective counterparts to the edge server managing their current domain to update the counterparts’ reputation score. However, even though in the proposed systems the reputation is managed and hosted in the Edge, the SOs’ reputation scores are also periodically saved on the Cloud. In this way, when a SO moves itself from an edge domain to another edge domain its reputation score will be available on the Cloud. Moreover, each edge server provides to cluster its managed SOs on the basis of their reputation scores to strengthen the competitiveness of SOs into the domain. This paper describes the proposed distributed clustering reputation-based framework for heterogeneous SOs and its performance in a simulated scenario.

Few experiences of Artificial Neural Networks (ANNs) for hydro-logical predictions in forest soils after wildfire and post-fire treatments are available in literature. To fill this gap, an ANN model has been adapted to predict surface runoff and soil erosion in Mediterranean burned pine forests (Central Spain), and tested against hydro-logical observations at plot scale throughout 2 years. The model gave very accurate runoff and erosion predictions in burned and non-burned soils as well as for all soil treatments (mulching and/or logging or not). Although further experimental tests are needed to validate the ANN applicability to soils in burned and treated forests in other ecosystems, the use of ANN may be useful for landscape planners as decision support system for the integrated assessment and management of forests.

Increasing road traffic levels in urban areas require actions and policies to manage and control the number of road users. Travelers' choices of transport modes, particularly private cars, that generate the main share of road traffic levels, depend on many factors, which include both personal preferences and level-of-service variables. Understanding how travelers choose transport modes according to the above factors is an important challenge in order to adopt the most suitable policies and facilitate a sustainable mobility. In the literature, behavioral models have been mainly proposed in order to both estimate mode choice percentages and capture travel behaviors by suitable estimation of some parameters associated to the above factors. However, behavior is complex in itself and the mechanisms underlying user behavior might be difficult to be captured by traditional models. In this paper, a neuro-fuzzy approach is proposed to extract mode choice decision rules by evaluating different sets of rules and different membership functions of the neuro-fuzzy model. Particularly, to determine which inputs are the most relevant in such decision process, fuzzy curves and surfaces have been considered in order to take into account nonlinear effects. The neuro-fuzzy model proposed in this paper has been thought to be embedded in an agent-based methodological framework where user agents - representing travelers - make travel choices based on the rules learnt by means of the neuro-fuzzy system.

The “Internet of Things” (IoT) provide humans and smart objects with attractive services, based on the advanced features of the IoT devices, like high sensing, real-time acting and reasoning. In our previous research we have highlighted that these features can be improved by promoting cooperation between smart objects, and we introduced the association between Multi-Agent Systems and IoT devices. In that context, we focused on the issue of accurately choosing the best partners for cooperation, in a scenario composed by several federations. We proposed a reputation model and we have shown that the model leads to detect agents having unreliable or misleading behaviors and that the model itself can be profitably used to form groups of agents that mutually cooperate for improving the effectiveness of their tasks. In this further contribution, we focus on the important issue of the group formation, by arguing that in practical IoT situations it is necessary to improve the group formation strategy to provide it with greater adaptability. To this end we introduce – in a particular IoT context described in this work – a two-phase group formation algorithm to support the reputation model. Experimental results prove that the adoption of the group formation algorithm, along with the proposed reputation model provides a few benefits to the whole IoT ecosystem.

In competitive Multi-Agent Systems (MASs) for IoT, an important task is to form friendships and groups for mutual cooperation. The existing proposals try to improve the profit of individual agent or group but, on the other hand, the most aggressive agents may take advantages from these approaches by means of malicious behaviors. In this work we model a MAS framework of non cooperative games by taking into account i) the reliability of agents relationships through a trust model and ii) the community ability to furnish a good environment to its members through its social capital. Thus, we designed an algorithm that maximizes the social capital. We emphasize two main findings: i) the result is a Nash equilibrium and ii) the agents are awarded only if they behave correctly.

Alzheimer’s Disease (AD) is an incurable neurodegenerative disorder which mainly affects older adults. An early diagnosis is essential because medical treatments can slow down the progression of the disease only if provided during the first stage, called Mild Cognitive Impairment (MCI). Starting from the study of electroencephalografic signals, brain functional connectivity analyses can be performed with the support of the graph theory. In particular, the purpose of this work is to verify the performances of three indexes, typically adopted to evaluate the graph robustness, in order to estimate the functional connectivity for three groups of subjects: healthy controls and people affected by dementia at two different stages (MCI and AD). The results obtained by the Connection Density Index, the Randić Index, and a normalized version of the Kirchhoff Index revealed a higher robustness in the brain networks of healthy people, followed by MCI and, finally, by AD patients, consistent with the hallmarks of Alzheimer’s disease. The statistical analysis showed that there is a significant difference between controls and AD for all three indexes. Finally, all three indexes were compared, revealing that the the Randić Index outperformed the other two indexes. These preliminary outcomes will be exploited to address further in-depth and time-expensive analyses for improving the diagnosis of Alzheimer’s disease.

In a massive IoT systems, large amount of data are collected and stored in clouds, edge devices, and terminals, but the data are mostly isolated. For many new demands of various intelligent applications, self-organized collaborated learning on those data to achieve group decisions has been a new trend. However, in order to reach the goal of group decisions, trust problems on data fusion and model fusion should be solved since the participants may not be trusted. We propose a consistent and trust fusion method with the consortium chain to reach a consensus, and complete the self-organized trusted decentralized collaborated learning. In each consensus process, consensus candidates check others’ trust levels to ensure that they tends to fuse consensus with users with high trust, where the trust levels are evaluated by scores according to their historical behaviors in the past consensus process and stored in the public ledger of blockchain. A trust rewards and punishments method is designed to realize trust incentive consensus, the candidates with higher trust levels have more rights and reputation in the consensus. Simulation results and security analysis show that the method can effectively defend malicious users and data, improve the trust perception performance of the whole federated learning network, and make the federated learning more trusted and stable.

Smart Workshops are experiencing the need of a mobile intelligence for mining both learning patterns and knowledge from the wide sea of data generated by both mobile users and mobile technologies. Indeed, mobile intelligence would represent the ideal substratum for providing "agentified" robots with a plethora of advanced capabilities (e.g., visual recognition, fault detection, self-recovery) and, hence, with high-level functionalities, like production line control, asset movement, connectivity restore. Besides the operational plane, however, mobile intelligence can be successfully exploited also in organizational tasks, like the formation of temporary, ad-hoc teams for accomplishing a given target. The complexity of some industrial operations, indeed, often demands the involvement of several, heterogeneous group of robots and the adequate representation of the reciprocal trustworthiness represents a key pre-requisite. It holds particularly for the Automated Guided Vehicles (AGVs) which are increasingly involved in collaborative activities aimed to optimise storage, picking, and transport functions in a wide variety of workshop areas. Therefore, in this paper we define a trustworthiness model for agentified AGVs based on the mix of their reputation and reliability and we present an agent-based framework implementing the related team formation strategy. The improvements obtained in terms of effectiveness and efficiency from the AGV team are observed and measured through a simulation activity, in which realistic settings for an industrial applications have been considered.

Flying vehicles are receiving more and more attention and are becoming an opportunity to start a new urban mobility paradigm. The most interesting feature of flying cars is the expected opportunity they could offer to reduce congestion, traffic jams and the loss of time to move between origin/destination pairs in urban contexts. In this perspective, urban air mobility might meet the concept of "sustainable mobility", intended as the ideal model of a transport system that minimizes the environmental impacts by maximizing efficiency and travel speed. For transport engineering planning issues, further knowledge is required in this field to understand the effects that a possible urban air mobility system, including the ground traffic component, could have in terms of sustainable mobility in the above meaning. This paper contributes to this topic by providing an analysis of different urban flying car scenarios by using an agent-based approach with different traffic conditions. The preliminary results obtained on some test networks and focusing on travel cost effects suggest that the expected advantages the flying car will depend on trip origin/destination points, average distances travelled in the urban contexts and the location of transition nodes, which are introduced as interchange nodes between aerial and ground mode.

In the last decades, individual urban traffic flows have increased all over the world with a consequent growth of road congestion and environmental pollution. In this context, car-pooling is an interesting car-based alternative to satisfy the individual mobility demand by optimizing the car loading factor with respect to the number of passengers, provided that all the participants share trip origin and destination in the same time slot. To make the system more appealing, this paper proposes an on-demand car-pooling service adopting variable fares, on the basis of trip length and number of participants. Multi-agent, reputation and recommender system technologies in synergy with a routing algorithm have been used to this aim. Experiments on simulated data proved the potentiality of the proposed approach.

In this paper, we are interested in answering the following research question: Is it possible to form effective groups in virtual communities by exploiting trust information without significant overhead, similarly to real user communities? In order to answer this question, instead of adopting the largely used approach of exploiting the opinions provided by all the users of the community (called global reputation), we propose to use a particular form of reputation, called local reputation. We also propose an algorithm for group formation able to implement the proposed procedure to form effective groups in virtual communities. Another interesting question is how to measure the effectiveness of groups in virtual communities. To this aim we introduce the Gk index in a measure of the effectiveness of the group formation. We tested our algorithm by realizing some experimental trials on real data from the real world EPINIONS and CIAO communities, showing the significant advantages of our procedure w.r.t. another prominent approach based on traditional global reputation.

In Smart Factories, automated guided vehicles (AGVs) accomplish heterogeneous tasks as moving objects, restoring connectivity, or performing different manufacturing activities into production lines. These kinds of devices combine several capabilities, as artificial intelligence (visual and speech recognition, automatic fault detecting, proactive behavior) and mobility, into the so-called 'mobile intelligence.' A typical scenario is represented by a workshop with a large number of mobile intelligent devices with associated agents, mutually interacting on their behalf. Here, to reach a given target by contemporary satisfying some basic requirements like effectiveness and efficiency, it is often necessary to organize ad hoc teams of free-moving vehicles, sensors, and smart devices. Therefore, a specific issue is the adequate representation of the reciprocal agent/device trustworthiness for advantaging such team formation processes within a smart factory environment. To this end, in this article, first, we define a trust measure based on reliability and reputation of AGVs, which are computed based on the feedbacks released for the AGVs activities in the factory; second, we design a trust framework exploiting the defined measures to support the formation of virtual, temporary, and trust-based teams of mobile intelligent devices; and third, we present a set of experimental results highlighting that the proposed trust framework can improve the workshop performance in terms of effectiveness and efficiency.

The purpose of the next internet of things IoT is that of making available myriad of services to people by high sensing intelligent devices capable of reasoning and real time acting. The convergence of IoT and multi-agent systems MAS provides the opportunity to benefit from the social attitude of agents in order to perform machine-to-machine M2M cooperation among smart entities. However, the selection of reliable partners for cooperation represents a hard task in a mobile and federated context, especially because the trustworthiness of devices is largely unreferenced. The issues discussed above can be synthesized by recalling the well known concept of social resilience in IoT systems, i.e., the capability of an IoT network to resist to possible attacks by malicious agent that potentially could infect large areas of the network, spamming unreliable information and or assuming unfair behaviors. In this sense, social resilience is devoted to face malicious activities of software agents in their social interactions, and do not deal with the correct working of the sensors and other information devices. In this setting, the use of a reputation model can be a practicable and effective solution to form local communities of agents on the basis of their social capabilities. In this paper, we propose a framework for agents operating in an IoT environment, called ResIoT, where the formation of communities for collaborative purposes is performed on the basis of agent reputation. In order to validate our approach, we performed an experimental campaign by means of a simulated framework, which allowed us to verify that, by our approach, devices have not any economic convenience to performs misleading behaviors. Moreover, further experimental results have shown that our approach is able to detect the nature of the active agents in the systems i.e., honest and malicious , with an accuracy of not less than 11 compared to the best competitor tested and highlighting a high resilience wi

IoT devices dealing with complex tasks usually require powerful hardware capabilities or, as a possible alternative, to get on the Cloud those resources they need. When an IoT device is 'virtualized' on the Cloud, it can take benefit from relying on one or more software agents and their social skills to mutually interact and cooperate. In particular, in a Cloud of Things scenario, where agents cooperate to perform complex tasks, the choice of a partner is a sensitive question. In such a context, when an agent is not capable to perform a reliable choice then, like real social communities, it can ask information to other agents it considers as trustworthy. In order to support agents in their partner choices, we conceived a local trust model, based on reliability and reputation measures coming from its ego-network, adopted to partition the agents in groups by exploiting trust relationships to allow agents to be associated with the most reliable partners. To this aim, we designed an algorithm to form agent groups by exploiting available local trust measures and the results obtained in a simulated scenario confirmed the potential advantages of this approach.

In the next years, flying cars are expected to become a real opportunity to realize Urban Air Mobility (UAM) systems. Most of the appeal is given by the opportunity of avoiding congestion, gaining time and reducing environmental impacts with respect to conventional mobility. However, UAM implementation is not trivial as it has several implications in manifold areas like safety, security, traffic control, legal issues and urban design among the others. To investigate on the impacts of UAM, a dedicated agent-based framework has been designed. The results of some preliminary tests carried out to verify the capabilities of this simulator are presented.

In the Internet of Things (IoT) age, the majority of real environments will become smart through the massive spread of novel devices empowered with cyber-physical abilities. Provided with increasing computation capabilities and pervasively deployed, smart IoT devices (SDs) will exhibit pro-Active behaviors and will perform more and more complex tasks, thus enabling the provision of advanced cyber-physical services which make the environment smarter and smarter. Such technology advancements together with an increased environmental awareness suggested the exploitation of these SDs for natural environments monitoring purposes. However, in presence of many and heterogeneous SDs, the formation of good teams requires high levels of trustworthiness among the members and, therefore, it is necessary to adequately represent their mutual trustworthiness. To this aim, the contribution provided by this paper consists in (i) defining a trust measure combining the reputation of SDs and the precision of their sensory data; (ii) designing a framework which adopts such measures as main criteria for the formation of temporary teams of humans and SDs; (iii) testing the proposed trust-based framework on a case study simulating the collaborative monitoring of a natural environment. The obtained results confirmed the potential improvements in the teams composition in terms of both performance and appreciation.

In the Cloud of Things, data are stored and processed in the cloud and the results are sent to IoT Smart Objects (SOs) with, generally, a network latency overhead due to the distances between the cloud and the local IoT networks. Moreover, trusting inappropriate counterparts can expose SOs to threats and to mitigate them we propose a clustering reputation-based approach for IoT Edge-based platform processing and storing data on the “edge”, nearby the SOs. Whenever SOs interact for services, a feedback is sent to an Edge server to calculate their reputation scores. In this way, the reputation systems is moved from the cloud to the edge servers, but available on the cloud if a SO will change edge domain. To this aim, we designed a distributed Trusted Object Framework (TOF) where heterogeneous OSs host and exploit the assistance of associated software agents and verified its performance in a simulated scenario which confirmed the potential advantages of TOF.

In the next future, an incredible amount of objectswill be mutually interconnected to exchange information and services to realize more and more complex tasks. Unfortunately, this process places significant risks on the interconnected objects and their users, due to the adoption of unreliable information sources. Consequently, an emerging issue is of improving thisaspect. In open and dynamic software agent contexts, this is a relevant question given the agents capability to quickly evaluate the potential effects of a greatest number of malicious behaviors in automated way. The adoption of authentication techniques is unable to assure the reliability of the information sources, differently from trust and reputation systems that fit well with such problems. The reliability of the information sources isbecoming relevant also in the context of urban mobility where vehicles could acquire information coming not only from the infrastructures but also from other vehicles. Reliable information sources are fundamental to contribute in reducing the negativetraffic effects in urban centers, which is among the most important factors affecting citizens’ life quality. In the above scenario, we designed a distributed reputation model, working within an agent framework, to assure the reliability of the informationsources in vehicle-to-vehicle interactions. In particular, the designed reputation model is able to detect malicious and cheating information sources as shown by some preliminary simulationscarried out on a simple urban test network.

The need of managing complex and interactive activities is becoming a key challenge in the “Internet of Things” (IoT) and leads to request large hardware and power resources. A possibility of facing such a problem is represented by the possibility of virtualizing physical IoT environments over the so called Cloud-of-Things (CoT), where each device is associated with one or more software agents working in the Cloud on its behalf. In this open and heterogeneous context, IoT devices obtain significant advantages by the social cooperation of software agents, and the selection of the most trustworthy partners for cooperating becomes a crucial issue, making necessary to use a suitable trust model. The cooperation activity can be further improved by clustering agents in different groups on the basis of trust measures, allowing each agent will to interact with the agents belonging to its own group. To this purpose, we designed an algorithm to form agent groups on the basis of information about reliability and reputation collected by the agents. In order to validate both the efficiency and effectiveness of our approach, we performed some experiments in a simulated scenario, which showed significant advantages introduces by the use of the trust measures.

Nowadays, a challenge for the ‘‘Internet of Things’’ (IoT) world is represented by the necessity of facing very complex and interactive tasks, such that IoT devices have to be equipped with hardware having very powerful capabilities. All this becomes particularly critical in presence of small and low-cost IoT devices. A way to deal with such problems is represented by the possible virtualization of the IoT environments over the cloud, the so called Cloud-of-Things (CoT), and then to associate each device with one or more software agents working in the Cloud environment. Moreover, the convergence of these technologies allows IoT devices to take significant benefits also by the social attitude of software agents to interact and cooperate. In this context, based on Machine-to-Machine (M2M) interactions, the choice of the partner for cooperating is a sensitive question, particularly in open and heterogeneous environments. If an agent does not hold suitable information to carry out a reliable choice then, similarly to real communities, it can ask information to other agents it considers as trustworthy. In this context, agents cooperation must be supported by a proper trust model which helps to select potential partners. This process can be further improved by partitioning the agents in different groups based on trust relationships. This way, each agent has the possibility to prefer the interactions with the agents belonging to its group that are, from its viewpoint, the most reliable for avoiding malicious behaviors and threats of different nature. To this purpose, we designed an algorithm, named CoTAG (CoT Agent Grouping algorithm), to form agent groups on the basis of information about reliability and reputation collected by the agents. To verify the efficiency and effectiveness of this algorithm, we carried out some experimentations in a simulated scenario. The results confirm the potential advantages deriving by the adoption of our proposal.

An important issue in Online Social Networks consists of the capability to generate useful recommen- dations for users, as peers to contact in order to establish friendships and collaborations, interesting resources to use and so on. This implies the necessity of evaluating the trustworthiness a user should assign to other members of his/her online community. In the past literature, a common approach for predicting trust is represented by a number of models that rely on “global” reputation: they are based on the evaluation of the behaviors of the users, that is shared across the entire community. These models, however, show an evident limitation due to the difficulty of taking the effects of malicious or fraudulent behaviors into account, thus making the feedback themselves. Other approaches, that consider also a local perspective of the trust, are limited by the fact they are supervised, i.e. they need a training phase in generating recommendations. In this paper, we propose a novel approach to extend global reputation models with a local reputation, computed on the ego-network of the user, by means of an unsupervised approach. We characterize our model by considering (i) the different relevance given to local and global repu- tation, (ii) the threshold that is used to consider a user unreliable and (iii) the dimension of the user’s ego-network. Experiments performed on a real data set show that global reputation is useful only if the size of the user ego-network is small, as in the case of a newcomer. Moreover, the combined usage of global and local reputation leads to predict the expected trust with a high level of precision.

Graph robustness-the ability of a graph to preserve its connectivity after the loss of nodes and edges-has been extensively studied to quantify how social, biological, physical, and technical systems withstand to external damages. In this paper, we prove that graph robustness can be quickly estimated through the Randic index, a parameter introduced in chemistry to study organic compounds. We prove that Erdos-Renyj (ER) graphs are a good specimen of robust graphs because they lack of a clear modular structure; we derive an analytical expression for the Randic index of ER graphs and use ER graphs as an effective term of comparison to decide about graph robustness. Experiments on real datasets from different domains (scientific collaboration networks, content-sharing systems, co-purchase networks from an e-commerce platform, and a road network) show that real-life large graphs are more robust than ER ones with the same number of nodes and edges. We also observe that if node degree distribution closely follows a power law, then few edges contribute for more than half of the Randic index, thus indicating that the selective removal of those edges has devastating impact on graph robustness. Finally, we describe sampling-based algorithms to efficiently but accurately approximate the Randic index.

The use of Recommender Systems (RSs) to support customers and sellers in Business-to-Consumer activities is emerged in the last years and several RSs have been proposed on different e-Commerce platforms to provide customers with automatic and personalized suggestions. However, the information such tools catch in supporting B2C customers in their Web activities then are unused to support them on the traditional commerce. In other words, these two environments operate separately without implementing synergistic actions to share knowledge and experiences between these two modality of commerce. In this paper, we propose a distributed RS, called ICR-IoT, based on a multi-tiered agent architecture, conceived to realize such a synergy. The key of our idea is that of using a tier, based on the Internet-of-Things technology, designed to catch information about customers of traditional markets in order to generate very effective suggestions to support commercial activities both on a traditional store as well as on an e-Commerce site.

In this paper we address the problem of measuring the overall effectiveness of group formation in virtual communities. Group formation is often driven by the combination of similarity and trust measures which are usually exploited with the recommendations provided by all the community members (global reputation). In this work propose a specific index to measure the effectiveness of group formation, and to exploit the local reputation in place of the global one. The use of local reputation will allow group administrators to save a significant amount of computationally and/or communicational tasks. We designed an algorithm to form effective groups in virtual communities and tested it on real data.

In social communities the agent groups composition could vary over time due to changes occurring in agents’ behaviors. To study the time evolution of such processes, we propose a conceptual framework exploiting a distributed algorithm driving the group formation. The results of tests carried out on real data, extracted by the social network CIAO, show as groups formed by combining similarity and trust measures are i) more timestable, independently by the weight of the trust component, and ii) more time-homogeneous, independently by the presence of uncorrelated random agents’ behaviors affecting the similarity component.

In recent years the increasing rate of vehicular traffic due to private mobility caused congestion and environmentalimpacts in urban contexts all over the world. To face such problems an important contribution might be given bytransit systems. However, transit systems are characterized by a discontinuous spatial and time coverage so that other formsof mobility, like car-sharing, can be an effective complementto it by providing the same flexibility and comfort of private cars. Several studies confirmed that car-sharing is almost as highly appreciated as private cars but having the advantage of lower costs. For such reasons, in recent years the car-sharing market increased continuously as it has been resulting more and more attractive for investors, although its market shareremains limited. To encourage the car-sharing philosophy, from one hand traditional car-sharing companies are trying to reduce operational costs to offer lower fares and, from the other hand, several individual owners are starting to share their cars suitablysupported by technology. To promote car-sharing activities, in this paper a multi-agent system able to monitor car-sharing users’driving habits is proposed. In particular, agents assist users in improving their driving, as well as in building their individualreputation measures over time. Such reputation scores can beused to allow the access to car-sharing services and personalized fares. Experiments on real and simulated data are encouragingand show the potentiality of this proposal.

E-Learning class formation will take benefit if common learners’ needs are taken into account. For instance, the availability of trust relationships among users can represent an additional motivation for classmates to engage activities. Common experience also suggests that there are many similarities within dynamics of formation for thematic social network groups and e-Learning classrooms. Since Online Social Networks provide data concerning users interactions (e.g. trust relationships), we propose a model aimed at managing the formation and the evolution of e-Learning classes based on information available on Online Social Networks — skills, interactions, and trust relationships — which are properly combined in a unique measure. The aim is to suggest both to a user the best classes to join with and to the classes themselves the best students to accept. The proposed approach has been tested by simulating an e-Learning scenario within a large social network. Experiments show that this proposal is able to support students and class managers in order to satisfy their expectations in a scalable manner.

Matching users profile in virtual communities can represent the most natural way of representing group homogeneity, i.e. how much the group members are mutually linked. However, optimizing profile matching does not guarantee the group cohesion, i.e. that the group will continue to be homogeneous in time. Moreover, computing profile matching in large virtual communities can be very expensive, and cannot be integrated in a fully distributed system. In the past, we have demonstrated that using users mutual trust, along with profile matching, can help to improve groups homogeneity. In this work we demonstrate, by an extended set of experiments on datasets extracted from real communities, that trust measures can effectively replace profile matching in order to optimize group's cohesion. A further interesting result is represented by the fact that it is also possible to replace the global trust measure with a local measure of trust, called local reputation, which is not highly sensitive to the size of the network, thus allowing to perform computations which are limited on the size of the ego-network of the single node.

Detecting communities in graphs is a fundamental data analysis tool to understand the structure of Web-based systems and predict their evolution. Many community detection algorithms handle undirected graphs (i.e., graphs with bidirectional edges) but many graphs on the Web -- e.g. microblogging Web sites, trust networks or the Web graph itself -- are often directed. Few community detection algorithms deal with directed graphs but we lack their experimental comparison. In this paper we evaluated some community detection algorithms across accuracy and scalability. A first group of algorithms (Label Propagation and Infomap) are explicitly designed to manage directed graphs while a second group (e.g., WalkTrap) simply ignores edge directionality; finally, a third group of algorithms (e.g., Eigenvector) maps input graphs onto undirected ones and extracts communities from the symmetrized version of the input graph. We ran our tests on both artificial and real graphs and, on artificial graphs, WalkTrap achieved the highest accuracy, closely followed by other algorithms; Label Propagation has outstanding performance in scalability on both artificial and real graphs. The Infomap algorithm showcased the best trade-off between accuracy and computational performance and, therefore, it has to be considered as a promising tool for Web Data Analytics purposes.