Trends in Practical Applications of Scalable Multi-Agent Systems, the PAAMS Collection

We investigate by means of agent-based simulations the influence of convex incentives, e.g. option-like compensation, on financial markets. We propose an agent based model already developed in Fabretti et al (2015), where the model was build with the aim of studying convex contract effect using the results of a laboratory experiment performed by Holmen et al. (2014) as benchmark. Here we replicate their results studying prices dynamics, volatility, volumes and risk preference effect. We show that convex incentives produces higher prices, lower liquidity and higher volatility when agents are risk averse, while, differently from Fabretti et al (2015), their effect is less evident if agents are risk lovers. This appears related to the fact that prices in the long run converge more likely to the equilibrium when agents are risk averse.JEL Classification G10, D40, D53

In this paper we analyze the impact of the French cancel order tax on market quality measured by market liquidity and volatility. Additionally, this paper raises the question whether this tax leads to reduction of high-frequency trading (HFT) activities and a declining in trading volume. Moreover, we test market rules that have not been yet introduced using artificial market framework.

The electricity markets environment has changed completely with the introduction of renewable energy sources in the energy distribution systems. With such alterations, preventing the system from collapsing required the development of tools to avoid system failure. In this new market environment competitiveness increases, new and different power producers have emerged, each of them with different characteristics, although some are shared for all of them, such as the unpredictability. In order to battle the unpredictability, the power supplies of this nature are supported by techniques of artificial intelligence that enables them crucial information for participation in the energy markets. In electricity markets any player aims to get the best profit, but is necessary have knowledge of the future with a degree of confidence leading to possible build successful actions. With optimization techniques based on artificial intelligence it is possible to achieve results in considerable time so that producers are able to optimize their profits from the sale of Electricity. Nowadays, there are many optimization problems where there are no that cannot be solved with exact methods, or where deterministic methods are computationally too complex to implement. Heuristic optimization methods have, thus, become a promising solution. In this paper, a simulated annealing based approach is used to solve the portfolio optimization problem for multiple electricity markets participation. A case study based on real electricity markets data is presented, and the results using the proposed approach are compared to those achieved by a previous implementation using particle swarm optimization.

With the advent of smart grids, distribution utilities have initiated a large deployment of smart meters on the premises of the consumers. The enormous amount of data obtained from the consumers and communicated to the utility give new perspectives and possibilities for various analytics-based applications. In this paper the current smart metering-based energy-theft detection schemes are reviewed and discussed according to two main distinctive categories: A) system state-based, and B) artificial intelligence-based.

Microgrids aim at providing reliable and optimised energy for all the grid participants. This requires the integration of technology solutions, especially for the management of the grid. Existing technologies, such as multi-agent systems, demonstrated their applicability in such fields by providing distributable, reliable, secure, and flexible solutions. We propose a reference architecture for microgrid approaches based on multi-agent systems with the aim of guiding software engineers and researches in the design and implementation of such solutions. Our reference architecture was validated by means of two case studies.

The intelligent management systems of the end consumers are endowed with advanced functions being one of them the interaction with external entities through the automatic participation in demand response programs. The development of the intelligent management systems is to reduce the energy consumption based on internal information and on the interaction with an external entity. Moreover, the management approaches results in an active participation of the consumers in the operation of the smart grids and microgrids concepts. The paper developed presents the application of a dynamic priority method in SCADA Office Intelligent Context Awareness Management system to manage the energy resources installed in an office building. The intelligent management method allows the dynamic active participation of the office building in the DR events considering the real data of consumption and generation of one building in Polytechnic of Porto. The main goal of the methodology is to obtain a dynamic scheduling for all energy resources with little interference in the comfort of users. The results of dynamic management model in office building are discussed for the participation in 8 hours demand response event. The power limit of the scenario depends on the consumption and micro-generation power of an October day.

Electric Vehicles are particularly adapted for frequency regulation service regarding their batteries features and their availability when not used for mobility. Many parameters have an important impact on the revenue and the cost of the service. On the one hand, the Electric Vehicle or Fleet owner behavior with the plugin hours, the driving patterns and the investments on the Vehicle to Grid technology may highly affect the service availability and cost. On the other hand, the service remuneration based on capacity remuneration and energy payment is highly dependent on the grid location, renewables penetration and Electric Vehicles presence. The aim of this study is to highlight various parameters variation depending on the context as well as their effect on the profitability of the service. The proposed tool allows apprehending the profitability for various markets situations from the contract level to the decision for the service delivery depending on the profit.

There has been continuous research in the energy distribution sector over the last years because of its significant impact in modern societies. Nonetheless, the use of high voltage power lines transport involves some risks that may be avoided with periodic reviews. The objective of this work is to reduce the number of these periodic reviews so that the maintenance cost of power lines is also reduced. This work is focused on the periodic review of transmission towers (TT) to avoid important risks, such as step and touch potentials, for humans. To reduce the number of TT to be reviewed, an organization-based agent system is proposed in conjunction with different artificial intelligence methods and algorithms. This system is able to propose a sample of TT from a selected set to be reviewed and to ensure that the whole set will have similar values without needing to review all the TT. As a result, the system provides a web application to manage all the review processes and all the TT of Spain, allowing the review companies to use the application either when they initiate a new review process for a whole line or area of TT, or when they want to place an entirely new set of TT, in which case the system would recommend the best place and the best type of structure to use.

This paper compares Ultra-Wide Band (UWB) and WiFi technologies as sources of information for accurate localization of static targets in indoor scenarios. Such technologies offer different localization accuracy and they are also characterized by different applicability. UWB provides very accurate localization information, but it requires a dedicated infrastructure and it is not yet widely available in mobile appliances. WiFi gives less accurate localization information but it is integrated in all modern mobile appliances and it does not require a dedicated infrastructure. This paper details on accuracy versus wide applicability trade-off and it provides quantitative criteria to choose one technology or the other. The discussed results are obtained using a new add-on module for JADE which allows embedding diverse sources of ranging information and localization algorithms.

During the last years, great advances has been produced in the automotive industry, a strategic sector both nationally and internationally with a high socioeconomic impact. Many efforts have focused on providing smart environments to the final user in vehicles such as cars, capable of detecting contextual vehicle’s conditions and adapting automatically to the user needs. This paper proposes an innovative solution in the automotive field consisting of a new product family which allows the transformation of a traditional bicycle to an electric bicycle by an architecture that provides the user intelligent adaptive environments and significantly improve the driving experience design enabling value-added services.

In this paper, we present the design and implementation of a novel agent-based platform to collect waste on cities and villages. A low cost sensor prototype is developed to measure the fulfilling level of the containers, a route system is developed to optimize the routes of the trucks and a mobile application has been developed to help drivers in their work. In order to evaluate and validate the proposed platform, a practical case study in a real city environment is modeled using open data available and with the purpose of identifying limitations of the platform.

Attention is strongly connected with learning and when it comes to acquiring new knowledge, attention is one the most important mechanisms. The learner’s attention affects learning results and can define the success or failure of a student. The negative effects are especially significant when carrying out long or demanding tasks, as often happens in an assessment. This paper presents a monitoring system using computer peripheral devices. Two classes were monitored, a regular one and an assessment one. Results show that it is possible to measure attentiveness in a non-intrusive way.

During recent years, research in smart cities and internet of the things has acquired a notable relevance. Current research is mainly focused on wireless sensor networks and data analysis. However, it is still necessary to provide new solutions for social problems based on mobile intelligent devices connected to the city. Mobility is a key factor for social environments in smart cities in which humans wear intelligent devices that can also be installed in vehicles and continuously vary their positions in the city. Social computing envisions a new kind of computation where humans and machines collaborate to compute and resolve a social problem. The role of mobile intelligent actors in social computing is still a challenge and require new solutions. In this paper, we present a multi-agent architecture that incorporates a new mobile sensing agent model and virtual organizations of agents for information fusion and machine learning, as well as contextual information to enrich the social knowledge representation.

This paper proposes a control system design for a mobile robot assistant based on a multi-agent architecture. The robotic platform used in this paper is a second generation Assistant Personal Robot (APR-02) with an own design. The control implementation is distributed among different agents in which each one is designed to fulfill a specific functionality such as localization, navigation, task managing, vision, hearing, communications, and environmental supervision. In addition, a set of shared memory instances are implemented to ensure the cohesion among all the agents while working together. The proposed methodology provides robustness and effectiveness by assigning each agent on a single CPU thread.

Social animals have conquered the world thanks to their ability to team up in order to solve survival problems. From ants to human beings, animals show ability to cooperate, communicate and divide labour among individuals. Cooperation allows members of a group to solve problems that a single individual could not, or to speed up a solution by splitting a task in subparts. Biological and swarm robotics studies suggest that division of labour can be favoured by differences in local information, especially in clonal individuals. However, environmental information alone could not suffice despite a task requires a role differentiation to be solved. In order to overcome this problem, in this paper, we analyse and discuss the role of a communication system able to differentiate signals emitted among a group of homogeneous robots to foster the evolution of a successful role allocation strategy.

We investigate the possibility of observing cooperative behavior in simple, autonomous, and non-communicating robotic agents when performing a harvesting task in a multi-agent environment. We evaluate a method to enforce a reflex-like territorial behavior in order to optimize individual and collective utilities.

The problem of allocating tasks to a team of robots composing a complex activity with global performance constraints to be met, is NP-hard. Automated negotiation was proposed as a viable heuristic approach allowing for the dynamic adjustment of the performance levels provided by the single robots in the case of robots with limited resources. This approach leads to an improved exploitation of robots capabilities in terms of the number of composite activities that can be successfully allocated to the team. In the present work, the proposed approach is extended to include the possibility for the robots to negotiate for task allocation, and to execute the tasks in an interleaved way, so that the capabilities of the entire team can be better exploited, reducing the time the robots are inactive.

This paper presents the core features of JADEL, an agent-oriented domain-specific programming language for the construction of JADE agents, behaviours and ontologies. The work on JADEL originates from the need to assist programmers by means of tools that reduce the complexity and speed up the construction of a JADE agents and multi-agent systems. The features of JADEL discussed in this paper include abstractions for the main entities of JADE—agents, behaviours and ontologies—and they also encompass the features needed for the construction of domain-specific tasks, thus enriching JADE APIs with novel and simple notations.

In the Internet of Things (IoT) domain, being able to propose a contextualized and personalized user experience is a major issue. The explosion of connected objects makes it possible to gather more and more information about users and therefore create new, more innovative services that are truly adapted to users. To attain these goals, and meet the user expectations, applications must learn from user behavior and continuously adapt this learning accordingly. To achieve this, we propose a solution that provides a simple way to inject this kind of behavior into IoT applications by pairing a learning algorithm (C4.5) with Behavior Trees. In this context, this paper presents new forgetting methods for the C4.5 algorithm in order to continuously adapt the learning.

The paper introduces a technique for representing quantifier relations that can have different scope order depending on context and agents. The technique is demonstrated by classes of terms denoting relations, where each of the arguments of a relation term is bound by a different quantifier. We represent a formalization of linking quantifiers with the corresponding argument slots that they bind, across $$\lambda $$-abstractions. The purpose of the technique is to represent underspecified order of quantification, for computationally efficient and adequate representation of scope ambiguity in the absence of context and corresponding information about the order. Furthermore, the technique is used to represent subclasses of larger classes of relations depending on order of quantification or specific relations.

Networks of Polarized Evolutionary Processors is a highly parallel distributed computing model inspired and abstracted from the biological evolution. This model is computationally complete and able to efficiently solve NP complete problems. Although this model is inspired from biology, basically it has been investigated from the points of view of mathematical and computer science goals with a qualitative perspective. It is true that Networks of Polarized Evolutionary Processors incorporate a numerical evaluation over the data that it processes, but this is not used from a quantitative viewpoint. In this paper we propose to enhance Networks of Polarized Evolutionary Processors of a quantitative perspective through a novel number of formal components. In particular, these components are able to evaluate quantitative conditions inherent to biological phenomena preserving the same computational power of Networks of Polarized Evolutionary Processors. Moreover, as a proof of concept, we model and simulate a simple but expressive example: a discrete abstraction of the sodium-potassium pump that includes the components proposed. Finally, we suggest that this integration enhances Networks of Polarized Evolutionary Processors model to (a) be more expressive for the algorithm design and (b) use less resources (nodes, rules, strings and computation time). This resource reduction could become a clear advantage when we will deploy hardware/software solutions of these bio-inspired computational models on top of massively distributed computational platforms.

One of the most important challenges of this decade is the Internet of Things (IoT) that pursues the integration of real-world objects in Internet. One of the key areas of the IoT is the Ambient Assisted Living (AAL) systems, which should be able to react to variable and continuous changes while ensuring their acceptance and adoption by users. This means that AAL systems need to work as self-adaptive systems. The autonomy property inherent to software agents, makes them a suitable choice for developing self-adaptive systems. However, agents lack the mechanisms to deal with the variability present in the IoT domain with regard to devices and network technologies. To overcome this limitation we have already proposed a Software Product Line (SPL) process for the development of self-adaptive agents in the IoT. Here we analyze the challenges that poses the development of self-adaptive AAL systems based on agents. To do so, we focus on the domain and application engineering of the self-adaptation concern of our SPL process. In addition, we provide a validation of our development process for AAL systems.

This paper proposes a multi-agent-based framework for the prevention of violence against women. A general description of the proposed service is presented. The service takes advantage of the Internet of Things that are/will be available in the context of Smart Cities and mobile technologies (such as smart phones). Lastly, some typical domestic violence scenarios are simulated by using Google Maps API and the results are shown.

The gathering of crowd traffic data either from videos or from visual observation has different uses. In the social simulation context, one of them is validating crowd behavior models and match the resulting traffic in control points with the real ones. When such models have been already validated, the immediate use can be aiding managers of facilities to infer, from real time data, what crowd behavior they should expect in their facilities. However, the transformation of those measurements into actual behavior patterns has not been satisfactorily addressed in the literature. In particular, most papers take into account a single measurement point. This paper contributes with an algorithm that produces possible populations that reproduces real traffic data obtained from multiple measurement locations. The algorithm has been validated against data obtained in a real field experiment.

This paper presents a middleware based on an agent model for supporting reactive agents that can change their behavior in order to evolve with time based on the accomplish of active norm defined into a dynamic normative model. This middleware named Agent Dynamic EvoLutionary at runtimE (ADELE) was developed on a Java platform using the JADE framework. ADELE middleware allows developing dynamic multi-agent systems applicable to ubiquitous applications where environments are highly dynamic. Our approach includes metaprogramming mechanisms which enable agents to be able to evolve through a behavior injection (on the fly) at runtime, instead of killing agents which probably can be implicated in other processes and cannot be interrupted to prevent problems to the entire system.

In recent years, much research has focused its attention on Ambient Intelligence (AmI). Its potential applications to smart homes, hospitals, health monitoring or daily life assistance make this paradigm a very promising field of research that can have a great and positive impact in our lives. The combination of AmI environments and Multi-Agent Systems (MAS) has emerged as a perfect solution for the development of this kind of applications. However, there are many challenges to be addressed before such applications can be put into practice. In this paper, we propose an architecture based on MAS aimed to build rehabilitation systems for people with Acquired Brain Injury (ABI) and explain how this architecture has been applied for the development of Vi-SMARt: a system for defining and planning therapies for people with ABI, and to control and evaluate their rehabilitation process.

We developed a novel approach for including metadata generated from Linked Open Data into Recommendation Systems by proposing a probabilistic view of Collective Matrix Factorization. The Linked Open Data cloud is being conceived and published to improve the usability and performance of various applications including Recommender Systems. While most previous works focus on exploiting Linked Open Data on content based Recommendation System, we include the semantic information into the collaborative filtering recommendation approach. With an unsupervised method, we generated different metadata representations for items from Linked Open Data and incorporated them into Probabilistic Matrix Factorization to get a double matrix factorization to boost the performance. Experiments showed that our proposed approach performs comparably well and in some scenarios generate significantly better results than Probabilistic Matrix Factorization methods when there is no semantic data inclusion.

Nowadays, online social networks such as Facebook and Twitter become increasingly popular. These social media channels allow people to create, share, and comment on information about anything related to their real-life. Such information is very useful for various application domains, e.g., decision support systems or online advertising.In this paper, we propose a comprehensive framework for retrieving relevant contents from online social network data. Our approach is proposed on the basic of the Vector Space Model and Support Vector Machine to process and classify raw text data. Our experiments demonstrate the utility and accuracy of the framework in retrieving fashion related contents from Twitter and Facebook.

In this paper, a novel agent-based platform for Twitter user clustering is proposed. We describe how our system tracks the activity for a given topic in the social network and how to detect communities of users with similar political preferences by means of the Louvain Modularity. The quality of this clustering method is evaluated against a subset of human-labeled user profiles. Finally, we propose combining community detection with a force-directed graph algorithm to produce a visual representation of the political communities.

This paper proposes a methodology to combine the depth information obtained from LIDAR and RGB-D sensors in order to generate enhanced 2D navigation maps which will be used by an Assistant Personal Robot. The objective of this procedure is to locate the mobile robot and to avoid collisions while performing movements in any direction.

To ensure a proficient topology discovery service in Software-Defined Networks (SDN), we propose a simple agents-based mechanism to improve the efficiency of the topology discovery process. In this work, an algorithm for a novel Topology Discovery Protocol (SD-TDP) is designed. This protocol will be implemented in each switch through an agent. Thus, this approach will provide a distributed solution to solve the problem of network topology discovery in a more simple and efficient way.

In this paper we address the issue of designing a novel distributed routing algorithm that optimizes the power consumption in large scale software-defined networks (SDN) with multiple domains. The solution proposed, called DEAR (distributed energy-aware routing), tackles the problem of minimizing the number of links that can be used to satisfy a given traffic demand under performance constraints such as control traffic delay and link utilization.

This paper proposes the design and development of a telescope network scheduler to maximize the overall observation acceptance rate. A key module of this scheduler is the telescope decision algorithm which objective is to avoid serving an observation to a telescope that cannot execute it.

Currently, the Web provides many different information sources with valuable information that is available in human friendly formats only. This makes it difficult for software agent to sift through them to extract relevant information to feed automated business processes.

The Web is an ever growing repository of valuable information. That information lacks semantics since it is buried into web documents that are represented using HTML. Information extractors are software components that help software engineers in the task of extracting structured information from web documents.

The main objective of this research is to ease organizational design of large-scale multi-agent systems (LSMAS) development. Methods including ontology engineering, metamodeling and code generation are proposed to achieve the set goal. The resulting modeling tool is expected to aid in development of LSMAS for numerous application domains.

A link discovery task is performed to automatically link instances from different datasets that describe the same real-world concept by means of owl:sameAs. The link discovery relies on a link specification [3, 5] which is an equality criteria between instances from different datasets.

In recent years, companies are demanding finer market analysis in order to increase their revenues, improve their products or services, and carry out better marketing campaigns. Most current proposals perform polarity analysis on a message level, that is, they cannot analyse the opinion regarding the different attributes that are referenced in a message.