Distributed Computing and Artificial Intelligence

In this paper, we focus on the learning aspect of NEAT and its variants in an attempt to solve benchmark problems through fewer generations. In NEAT, genetic algorithm is the key technique that is used to complexify artificial neural network. Crossover value, being the parameter that dictates the evolution of NEAT is reduced. Reducing crossover rate aids in allowing the algorithm to learn. This is because lesser interchange among genes ensures that patterns of genes carrying valuable information is not split or strayed during mating of two chromosomes. By tweaking the crossover parameter and with some minor modification, it is shown that the performance of NEAT can be improved. This enables NEAT algorithm to evolve slowly and retain information even while undergoing complexification. Thus, the learning process in NEAT is greatly enhanced as compared to evolution.

Ciudad Juárez, a large city located along the Mexico-United States border with a population over a million people in 87 km

2

has recently experienced a history of violence and insecurity related directly to organized crime: assaults, kidnappings, multi-homicides, burglary, between others. However the second leading cause of death in the city is associated with traffic accidents: 1,377 deaths in 2011 alone. For this reason, citizens have actively pursued specific programs that would decrease the overwhelming statistics: 3,897 deaths from 2008 to 2012. The reason of the following project is to provide drivers with a technological tool with indicators and sufficient information based off statistics compiled by the

Centro de Investigaciones Sociales

(Social Research Centre) at Autonomous University of Ciudad Juárez and other public sources. Then drivers would have more information on possible traffic accidents before they happened. This research tries to combine a Mobile Device based on Cultural Algorithms and Data Mining to determine the danger of suffering a traffic accident in a given part of the city during a specific time.

When agents and interactions grow in a situated agent-based simulations, requirements in memory or computation power increase also. To be able to tackle simulation with millions of agents, distributing the simulator on a computer network is promising but raises issues related to time consistency and synchronization between machines. This paper study the cost in performances of several synchronization policies and their impact on macroscopic properties of simulations. To that aims, we study three different time management mechanisms and evaluate them on two multi-agent applications.

Decision trees are well known, widely used algorithm for building efficient classifiers.We propose the modification of the Parallel Hoeffding Tree algorithm that could deal with large streaming data. The proposed method were evaluated on the basis of computer experiment which were carried on few real datasets. The algorithm uses parallel approach and the Hoeffding inequality for better performance with large streaming data. The paper present the analysis of Hoeffding tree and its issues.

A decentralized solution to the unsplittable flow problem (UFP) in a transport network is considered, where each flow uses only one route from source to sink and the flows cannot be separated into parts in intermediate nodes. The flow costs in each edge depend on the combination of the assigned flows as well as on external random variables. The distributions of the random variables are unknown, only samples are available. In order to use the information available in the samples more effectively, several resamples are constructed from the original samples. The nodes agree on the resamples in a decentralized way using a cooperative resampling scheme. A decentralized asynchronous solution algorithm for the flow routing problem in these conditions is proposed, which is based on the ADOPT algorithm for asynchronous distributed constraint optimization (DCOP). An example illustrating the proposed approach is presented.

This paper explores the main challenges to be tackled for more accessible, and easy to use, Smart Spaces. We propose to use Semantic Web principles of interoperability and flexibility to build an end-user graphical model for rapid prototyping of Smart Spaces applications. This approach is implemented as a visual rule-based system that can be mapped into SPARQL queries. In addition, we add support to represent imprecise and fuzzy knowledge. Our approach is exemplified in the experimental section using a context-aware test-bed scenario.

OPTICS (Ordering Points To Identify the Clustering Structure) is an algorithm for finding density-based clusters in data.We introduce an adaptive dynamical clustering algorithm based on OPTICS. The algorithm is applied to clustering emails which are represented by quantitative profiles. Performance of the algorithm is assessed on public email corpuses TREC and CEAS.

In this paper we study the optimal saving problem in the framework of possibility theory. The notion of possibilistic precautionary saving is introduced as a measure of the way the presence of risk (represented by a fuzzy number) influences a consumer in establishing the level of optimal saving. The equivalence between the prudence condition (in the sense of Kimball) and a positive possibilistic precautionary saving is proved. Some relations between possibilistic risk aversion, prudence and possibilistic precautionary saving are established.

Recently, the economy has taken a downturn, which has forced many companies to reduce their costs in IT. This fact has, conversely, benefited the adoption of innovative computing models such as cloud computing, which allow businesses to reduce their fixed IT costs through outsourcing. As the number of cloud services available on the Internet grows, it is more and more difficult for companies to find those that can meet their needs. Under these circumstances, enabling a semantically-enriched search engine for cloud solutions can be a major breakthrough. In this paper, we present a fully-fledged platform based on semantics that (1) assist in generating a semantic description of cloud services, and (2) provide a cloud-focused search tool that makes use of such semantic descriptions to get accurate results from keyword-based searches. The proposed platform has been tested in the ICT domain with promising results.

The study of the impact of emotion and affect in decision making processes involved in a working team stands for a multi-disciplinary issue (e.g. with insights from disciplines such as Psychology, Neuroscience, Philosophy and Computer Science). On the one hand, and in order to create such an environment we look at a team of affective agents to play into a battlefield, which present different emotional profiles (e.g. personality and mood).On the other hand, to attain cooperation, a voting mechanism and a decision-making process was implemented, being Robocode used as the simulation environment. Indeed, the results so far obtained are quite satisfying; the agent team performs quite well in the battlefield and undertakes different behaviours depending on the skirmish conditions.

On the way to autonomous service robots, spatial reasoning plays a main role since it properly deals with problems involving uncertainty. In particular, we are interested in knowing people’s pose to avoid collisions. With that aim, in this paper, we present a qualitative acceleration model for robotic applications including representation, reasoning and a practical application.

A medium size (and bigger) company has to have a good (i.e. adjusted to the companies profile and its environment for example economic situation or seasonality of the demand) informatics system. The reason for it is that large amount of date has to be collected on time, transformed and transmitted to allow managers to make adequate and made on time decision. Time factor in this situation is very important and without a good logistic system it is difficult or even impossible to deal with contemporary market. In this paper a framework of a hybrid intelligent system is presented, which will optimize the lead time needed to the transport goods from the company to the receiver. Consequently the time, when the final product will appear on the market will be optimized and thus the product can be both due the client. Transport is often associated with the Travelling Salesman Problem and the method of its approximated solution will be considered.

New Internet technologies can considerably enhance contemporary traffic control and management systems (TCMS). Such systems need to process increasing volumes of data available in clouds, and so new algorithms and techniques for statistical data analysis are required. A very important problem for cloud-based TCMS is the selection of the shortest itinerary, which requires route comparison on the basis of historical data and dynamic observations. In the paper we compare two non-overlapping routes in a stochastic graph. The weights of the edges are considered to be independent random variables with unknown distributions. Only historical samples of the weights are available, and some edges may have common samples. Our purpose is to estimate the probability that the weight of the first route is greater than that of the second one. We consider the resampling estimator of the probability in the case of small samples and compare it with the parametric plugin estimator. The analytical expressions for the expectations and variances of the proposed estimators are derived, which allow theoretical evaluation of the estimators’ quality. The experimental results demonstrate that the resampling estimator is a suitable alternative to the parametric plug-in estimator. This problem is very important for a vehicle decision-making procedure to choose route from the available alternatives.

This paper studies the application of a genetic algorithm (GA) for determining closest efficient targets in Data Envelopment Analysis. Traditionally, this problem has been solved in the literature through unsatisfactory methods since all of them are related in some sense to a combinatorial NP-hard problem. This paper presents and studies some algorithms to be used in the creation, crossover and mutation of chromosomes in a GA, in order to obtain an efficient metaheuristic which obtains better solutions.

We consider an array sensing system of odors and adopt a layered neural network for classification. Measurement data obtained from fourteen metal oxide semiconductor gas (MOG) sensors are used, where some sensors exhibit relatively weak responses.We propose two methods for enhancing such weak signals to obtain better classification results. One method is to apply scaling to magnify the weak signals as to increase their significance in the classification criteria. The other method also involves magnifying the weak signals. However, predetermined values are assigned in the order of the magnitude of the actual signals. In both methods the group of weak signals is first determined. Then their values are negated prior to scaling, in order to be distinguished from stronger signals. An experiment shows that the accuracy of classifying five kinds of odors is improved from 74% to 85%.

This paper discusses a process to graphically view and analyze information obtained from a network of urban streets, using an algorithm that establishes a ranking of importance of the nodes of the network itself. The basis of this process is to quantify the network information obtained by assigning numerical values to each node, representing numerically the information. These values are used to construct a data matrix that allows us to apply a classification algorithm of nodes in a network in order of importance. From this numerical ranking of the nodes, the process finish with the graphical visualization of the network. An example is shown to illustrate the whole process.

The Case-Based Reasoning (CBR) is an appropriate methodology to apply in diagnosis and treatment. Research in CBR is growing and there are shortcomings, especially in the adaptation mechanism. In this paper, besides presenting a methodological review of the technology applied to the diagnostics and health sector published in recent years, a new proposal is presented to improve the adaptation stage. This proposal is focused on preparing the data to create association rules that help to reduce the number of cases and facilitate learning adaptation rules.

This paper proposes a distributed processing framework inspired from data processing. It unique among other data processing for large-scale data, socalled bigdata, because it can locally process data maintained in distributed nodes, including sensor or database nodes with non-powerful computing capabilities connected through low-bandwidth networks. It usesmobile agent technology as amechanism to distribute and execute data processing tasks to distributed nodes and aggregate their results. The paper outlines the architecture of the framework and evaluates its basic performance.

Many distributed systems (task scheduling,moving priorities,mobile environments, ...) can be linked as Dynamic Optimization Problems (DOPs), since they require to pursue an optimal value that changes over time. We have focused on the utilization of Distributed Genetic Algorithms (dGAs), one of the domains still to be investigated for DOPs. A dGA essentially decentralizes the population in islands which cooperate through

migrations

of individuals. In this article, we analyze the effect of the migrants selection and replacement on the performance of dGAs for DOPs. Quality and distance based criteria are tested using a comprehensive set of benchmarks. Results show the benefits and drawbacks of each setting for DOPs.

The Project being presented is related to the use of ubiquitous computing to reduce the domestic violence rate. The general framework presented is for the design and development of an emotional support social network. The technologies used are a combination of artificial intelligence, data mining, speech processing and Android based services. In this same way a comparative is going to be obtaining from the Basque Country in Spain and Mexico so an Iberoamerican emotional support social network is constructed. The assessment is based on satisfaction surveys and longitudinal analysis of stored data.

Polygonal approximation is based on the division of a closed curve into a set of segments. This problem has been traditionally approached as a single-objective optimization issue where the representation error was minimized according to a set of restrictions and parameters. When these approaches try to be subsumed into more recent multi-objective ones, a number of issues arise. Current work successfully adapts two of these traditional approaches and introduces them as initialization procedures for a MOEA approach to polygonal approximation, being the results, both for initial and final fronts, analyzed according to their statistical significance over a set of traditional curves from the domain.

Modern and efficient GPUs evolve towards a new integration paradigm for parallel processing systems, where Message-Passing Interfaces (MPI), Open MP and GPU architectures (CUDA) may be joined to perform a powerful high performance computation system (HPC). Nevertheless, this challenge requires much effort to properly integrate both technology and software programming. This paper describes GSched (Grid Scheduler), an optimized scheduler that allows distributing both CPU and GPU processor execution using a previous calculated optimum pattern, obtaining the best elapsed execution time for overall execution. Furthermore, high-level algorithm description is introduced to efficiently distribute processing and network resources.

This research presents the knowledge discovery using Data Mining from the organization and with a KPI management point of view. The stages presented here are based on techniques and Data Mining models, with emphasis on clustering techniques, such as the C-MEANS algorithm. We both consider the classic and fuzzy perspectives, namely Fuzzy C-MEANS and K-MEANS, and then compare the results based on the level of support which each algorithm provides to information management. The CRISP-DM methodology is used in our implementation, which is then applied to three case studies.

This paper presents the work in progress of a mobile-based distributed system which aims to minimize the social impact of abandoned or lost animals. System is based on the use of smart mobile devices to provide message warnings of animals localized. Messages are stored in a database to be processed. In order to enter data such as photography, audio and artificial images, system uses different mobile device interfaces. Data processing consists mainly in matching localized animals with lost animals, assigning abandoned animals at shelters and generating notifications for animal shelters or authorities. Currently, the system is in the development phase. The technical challenges in which we are working are to optimize data and metadata matching, and the management of message warning.

Open Shortest Path First (OSPF) is a widely used routing protocol that depends on weights assigned to each link to make routing decisions. If traffic demands are known, the OSPF weight setting (OSPFWS) problem can be defined to seek a set of weights that optimize network performance, typically by minimizing a congestion measure. The OSPFWS problem is NP-hard and, thus, meta-heuristics such as Evolutionary Algorithms (EAs) have been used in previous work to obtain near optimal solutions. However, the dynamic nature of this problem leads to the necessity of addressing these problems in a more robust manner that can deal with changes in the conditions of the network. Here, we present EAs for two of those tasks, defining objective functions that take into account, on the one hand, changes in the traffic demand matrices and, on the other, single link failures. Those functions use weighting schemes to provide trade-offs between the behaviour of the network in distinct conditions, thus providing robust sets of OSPF weights.The algorithms are implemented in the open-source software NetOpt framework.

In this paper, an Intelligent Tutoring System (ITS) for teaching computational logic called SIAL is described. Several basic topics in computational logic are covered. The more complex part in SIAL is the module in charge of the diagnosis, which performs model-based diagnosis although sometimes, a knowledge-based (expertise) model is necessary in order to yield a more accurate diagnosis. The inherent complexity of the ITS is approached using a Multi-Agent System (MAS). The classical approach in ITS, which divides them into four independent modules, is adapted to a MAS creating an agent for each module and other agent for any other subsystem needed. The results obtained from an experiment of usage of SIAL are presented.

Disasters like terrorist attacks, earthquakes, hurricanes, and volcano eruptions are usually unpredictable events that affect a high number of people. We propose an approach that can be used as a decision support tool for a post-disaster response that allows the assignment of victims to hospitals and organizes their transportation via emergency vehicles. Exploiting Operational Research and Constraint Programming techniques we are able to compute assignments and schedules of vehicles that save more victims than heuristic based approaches.

By offloading some computation to graphical processing units (GPUs), agent-based simulation (ABS) can be accelerated up to thousands of times faster. To exploit the power of GPUs, modellers can use available simulation frameworks to auto-generated GPU codes without requiring any knowledge of GPU programming languages. However, such frameworks only support computation on the GPUs of a particular vendor. This paper proposes techniques, implemented in a synchronous parallel discrete event simulation (PDES) tool, to allow modellers to create ABS models, and to specify computation regions in the models for multiple vendor’s GPUs or CPUs. The technique comprises a set of meta-language tags and a compilation framework to convert user-defined GPU execution regions to OpenCL. A well-known cellular ABS models, the Conway’s Game of Life, have been implemented and evaluated on two platforms

i.e.

, the NVIDIA GeForce 240M LE and AMD Radeon HD6650M. The preliminary results demonstrate two findings: (a) the proposed technique allows the example ABS model to be executed on a PDES engine successfully; (b) the generated GPU-enabled ABS model can achieve fourteen times faster than its multicore version.

Program Supervision aims at automating the use of complex programs, independently of any particular application domain. Program Supervision Systems offer original approaches to plan and control program processing activities. Since real applications imply more and more participants on various sites, we worked on the distribution of such systems. Therefore, given distributed data, programs and knowledge, our aim is to propose convenient and efficient models for distributing program supervision systems dealing with image processing and image analysis. In this paper, we present the Agent and Knowledge Models we used for the distributed, collaborative and intelligent assistant for image processing and its analysis. It is collaborative because the participants (knowledge-bases designers, program developers and other users) can work collaboratively to enhance the quality of programs and then the quality of the results. It is intelligent since it is a knowledge-based system including, but not only, a knowledge base, an inference engine said “supervision engine” and ontologies.

The design patterns have attracted increasing attention in the field of software engineering, since effectively selecting the fits pattern for a given problem can seriously improve the quality of the software, on the contrary of the expert developers selecting the suitable pattern process consider to be critical phase especially for novice developers which have to be provided with mechanism to help them find a suitable pattern to a particular solution. This paper introduces a design pattern selection architecture (DPS) based on a Multi-Agent System (MAS) that aim to obtain the appropriate recommendation to reduce development efforts, facilitate and assist the developers in selecting the suitable patterns for their problems.

The Qualitative Rectilinear Projection Calculus (QRPC) is a novel representation model for describing qualitatively motion patterns of two objects through the possible relationships among the rectilinear projection of their trajectories. The paper introduces the key issues of the model (i) the set of geometric relations defined in terms of the front-back and left-right dichotomies, (ii) how it can be possible enumerate an exhaustive set of qualitative states by the composition of these relations and (iii) the possible transitions among states based on the notion of conceptual neighborhood. The representational ability of the model is illustrated by an example extracted from the traffic engineering field where the relative motion of two objects is analyzed and described in terms of the QRPC-states.

This paper presents a proposal to extend the event management sub-system of the Distribution Data Service standard (DDS). The proposal allows user to optimize the use of DDS in networked control systems (NCS). DDS offers a simple event management system based on message filtering. The aim of the proposal is to improve the event management with three main elements: Events, Conditions and Actions. Actions are the new element proposed. Actions perform basic operations in the middleware, discharging the process load of control elements. The proposal is fully compatible with the standard and can be easily added to an existing system. Proposal has been tested in a distributed mobile robot navigation system with interesting results.

This article provides an algorithm that is dedicated to repeated trajectory optimization with a fixed horizon and addresses processes that are difficult to describe by the established laws of physics. Typically, soft-computing methods are used in such cases, i.e. black-box modeling and evolutionary optimization. Both suffer from high dimensions that make the problems complex or even computationally infeasible. We propose a way how to start from very simple problems and - after the simple problems are covered sufficiently - proceed to more complex ones. We provide also a case study related to the dynamic optimization of the HVAC (heating, ventilation, and air conditioning) systems.

One of the open issues in risk literature is the difference between risk perception and effective risk, especially when the risk is clearly defined and measured. Until now, the main focus has been given on the behaviour of individuals and the evidences of their biases according to some stimulus. Consequently, it is important to analyse what are the main reasons for those biases and identify the dimensions and mechanisms involved. To that purpose, we tackle the classic problem of tax fraud as a case study. In this paper, we will look into how agent based modelling methodology can help unfold the reasons why individuals commit errors of judgment when risk is involved.

In this paper, we deal with automated multi-issue concurrent negotiations. A buyer utilizes a number of threads for negotiating with a number of sellers. We propose a method based on the known PSO algorithm for threads coordination. The PSO algorithm is used to lead the buyer to the optimal solution (best deal) through threads team work. Moreover, we propose a weights adaptation scheme for optimizing buyer behavior and promoting efficiency. This way, we are able to provide an efficient mechanism for decision making in the buyer’s side. This is proved by our results through a wide range of experiments.

We have an increasingly sedentary population without the care to make a healthy diet. Therefore, it becomes necessary to give the population the opportunity, despite living a very busy and tiring life, to have control over important aspects to their health. This work aims to present a model of an ambient intelligence system for monitoring the weight and physical activity in active individuals. To accomplish this objective we have developed a mobile application that allows users to monitor their weight over a period of time, identify the amount of food they consume and the amount of exercise they practice. This mobile application will give information to users about dietary and physical activity guidelines in order to improve their lifestyles. It is expected that students improve their lifestyles.

In an idea generation meeting, the facilitator role is essential to obtain good results. The emotional context of the meeting partially determines the (un)success of the meeting, so the facilitator needs to obtain and process this information. Thus, the facilitator role is to assist the participants to reach their goals, i.e., to generate ideas with quality. In this paper is proposed an emotional aware architecture whose aim is to assist the facilitator in the process of maximizing the results of the meeting.

Intelligent environments and ambient intelligence provide means to monitor physical environments and to learn from users, generating data that can be used to promote sustainability. With communities of intelligent environments, it is possible to obtain information about environment and user behaviors which can be computed and ranked. Such rankings are bound to be dynamic as users and environments exchange interactions on a daily basis. This work aims to use knowledge from communities of intelligent environments to their own benefit. The approach presented in this work uses information from each environment, ranking them according to their sustainability assessment. Recommendations are then computed using similarity and clustering functions ranking users and environments, updating their previous records and launching new recommendations in the process.

In this paper we present a distributed solution for the acquisition, processing, archiving and diffusion of endoscopic procedures. The goal is to provide a system capable of managing all administrative and clinical information (including audiovisual content) since the acquisition process to the searching process of previous exams, for comparison with new cases. In this context, a device for the acquisition of the endoscopic video was designed (

MIVbox

), regardless of the endoscopic camera that is used. All the information is stored in a structured and standardized way, allowing its reuse and sharing. To facilitate this sharing process, the video undergoes several processing steps in order to obtain a summarized video and the respective content characteristics. The proposed solution uses an annotation system that enables content querying, thus becoming a versatile tool for research in this area. A streaming module in which the endoscopic video is transmitted in real time is also provided.

Texture classification poses a well known difficulty within computer vision systems. This paper reviews a method for image segmentation based on the classification of textures using artificial neural networks. The supervised machine learning system developed here is able to recognize and distinguish among multiple feature regions within one or more photographs, where areas of interest are characterized by the various patterns of color and shape they exhibit. The use of an enhancement filter to reduce sensitivity to illumination and orientation changes in images is explored, as well as various post-processing techniques to improve the classification results based on context grouping. Various applications of the system are examined, including the geographical segmentation of satellite images and a brief overview of the model’s performance when employed on a real time video stream.

Nowadays there are several systems that help people with disabilities on their quotidian tasks. The visual impairment is a problem that affects several people in their tasks and movements. In this work we propose an architecture capable of processing information from the environment and suggesting actions to the user with visual impairments, to avoid a possible obstacle. This architecture intends to improve the support given to the user in their daily movements. The idea is to use speculative computation to predict the users’ intentions and even to justify the reactive or proactive users’ behaviors.

The research offers a quite simple view of methods to classify edible and poisonous mushrooms. In fact, we are looking for not only classification methods but also for an application which supports experts’ decisions. To achieve our aim, we will study different structures of neural nets and learning algorithms, and select the best one, according to the test results.

Artificial Intelligence has been applied to dynamic games for many years. The ultimate goal is creating responses in virtual entities that display human-like reasoning in the definition of their behaviors. However, virtual entities that can be mistaken for real persons are yet very far from being fully achieved. This paper presents an adaptive learning based methodology for the definition of players’ profiles, with the purpose of supporting decisions of virtual entities. The proposed methodology is based on reinforcement learning algorithms, which are responsible for choosing, along the time, with the gathering of experience, the most appropriate from a set of different learning approaches. These learning approaches have very distinct natures, from mathematical to artificial intelligence and data analysis methodologies, so that the methodology is prepared for very distinct situations. This way it is equipped with a variety of tools that individually can be useful for each encountered situation. The proposed methodology is tested firstly on two simpler computer versus human player games: the rock-paper-scissors game, and a penalty-shootout simulation. Finally, the methodology is applied to the definition of action profiles of electricity market players; players that compete in a dynamic game-wise environment, in which the main goal is the achievement of the highest possible profits in the market.

The Today’s video games are highly technologically advanced, giving users the ability to step into virtual realities and play games from the viewpoint of highly complex characters. Most of the current efforts in the development of believable bots in videogames — bots that behave like human players — are based on classical AI techniques. Specifically, we design virtual bots using Continuous-Time Recurrent Neural Network (CTRNNs) as the controllers of the non-player characters, and we add a learning module to make an agent be capable of re-learning during its lifetime. Agents controlled by CTRNNs are evolved to search for the base camp and the enemy’s camp and associate them with one of two different altitudes depending on experience.We analyze the best-evolved agent’s behavior and explain how it arises from the dynamics of the coupled agent-environment system. The ultimate goal of the contest would be to develop a computer game bot able to behave the same way humans do.

This paper presents a system composed by a set of tools that facilitate the work of disabled people in their work environment. The PANGEA platform was used to build the base architecture of the system, where each tool is designed as a collection of intelligent agents that offer the services as Web-services. Moreover, all the system is implemented as an Open MAS. In this paper two tools are presented in detail, the proximity detection tool and the translator tool for people with hearing impairments.

Cloud computing is revolutionizing the services provided through the Internet, and is continually adapting itself in order to maintain the quality of its services. This study presents the platform +Cloud, which proposes a cloud environment for storing information and files by following the cloud paradigm. This study also presents Warehouse 3.0, a cloud-based application that has been developed to validate the services provided by +Cloud.

State-of-the-art approaches to collaborative filtering are based on the use of an input matrix that represents each user profile as a vector in a space of items and, analogically, each item as a vector in a space of users. When the behavioral input data have the form of (

userX, likes, itemY

) and (

userX, dislikes, itemY

) triples, one has to propose a bi-relational data representation that is more flexible than the ordinary user-item ratings matrix. We propose to use a matrix, in which columns represent RDF-like triples and rows represent users, items, and relations. We show that the proposed behavioral data representation based on the use of an element-fact matrix, combined with reflective matrix processing, enables outperforming state-of-the- art collaborative filtering methods based on the use of a ’standard’ user-item matrix.

This document provides an approach to biometrics analysis which consists in the location and identification of faces in real time, making the concept a safe alternative to Web sites based on the paradigm of user and password. Numerous techniques are available to implement face recognition including the principal component analysis (PCA), neural networks, and geometric approach to the problem considering the shapes of the face representing a collection of values. The study and application of these processes originated the development of a security architecture supported by the comparison of images captured from a webcam using methodology of PCA, and the Hausdorff algorithm of distance as similarity measures between a general model of the registered user and the objects (faces) stored in the database, the result is a web authentication system with main emphasis on efficiency and application of neural networks.

From the last decade, Web services technology has witnessed a great adoption rate as a new paradigm of communication and interoperability between different software systems. This fact, has led to the emergence of Web services and to their proliferation from outside the boundary of the UDDI business registry to other potential service resources such as public and private service registries, service portals, and so on. The main challenge that arises from this situation is the fact that for the same service implementation, several service descriptions are published in different service registries. Accordingly, if the service implementation is updated all of its descriptions have to be updated too over all of these registries. Otherwise, the service user may not bind to the suitable Web service if its descriptions are inaccurate or outdated. To address the above challenge, we propose in this paper a multi agent-based model that focuses on synchronizing the description of Web services, especially their quality of service, to maintain their consistency and sustainable use.

The paper applies an agent-based simulation to investigate the effectiveness of basic design options for management accounting systems. In particular, different settings of how to improve the information base by measurement of actual values in the course of adaptive walks are analyzed in the context of different levels of complexity and coordination modes. The agent-based simulation is based on the idea of NK fitness landscapes. Results provide broad, but no universal support for conventional wisdom that lower inaccuracies of information lead to more effective adaptation processes. Furthermore, results indicate that the effectiveness of improving judgmental information by actual values subtly depends on the complexity of the decisions and the coordination mode applied.

In this paper, we propose a norms mining algorithm that detects a domain’s potential norms, which we called the Potential Norms Mining Algorithm (PNMA). According to the literature, an agent changes or revises its norms based on variables of local environment and amount of thinking about its behaviour. Based on these variables, the PNMA is used to revise the norms and identify the new normative protocol to comply with the domain’s norms. The objective of this research is to enable an agent to revise its norms without a third party enforcement unlike most of the work on norms detection and identification, which entail sanctions by an authority. We demonstrate the execution of the algorithm by testing it on a typical scenario and analyse the results on several issues.

This paper presents a new methodical approach to the problem of collision avoidance of mobile robots taking advantages of multi-agents systems to deliver solutions that benefit the whole system. The proposed method has the next phases: collision detection, obstacle identification, negotiation and collision avoidance. In addition of simulations with virtual robots, in order to validate the proposed algorithm, an implementation with real mobile robots has been developed. The robots are based on Lego NXT, and they are equipped with a ring of proximity sensors for the collisions detections. The platform for the implementation and management of the multi-agent system is JADE.

Modern insurance information systems need intelligence to provide new functions that till now as a rule have been carried out by humans. Introduction of intelligent mechanisms into information systems allows the insurance companies to automate processes in the insurance business and achieve two benefits. Firstly, the amount of work done by humans is reduced and secondly more services can be provided to customers electronically, which increases the level of customer service. Additionally, insurance information systems need to communicate with many other systems to get the needed data. These demands fit the characteristics of intelligent agents. Thus the paper proposes to implement an insurance information system as a multi-agent system using intelligent agents to realize the modules of insurance information systems. A novel multi-agent architecture for insurance information system development is proposed.

This paper analyzes the main barriers that user model developers have to face when evaluating a statistical user model. Main techniques used to evaluate statistical user models, mostly borrowed from the areas of Machine Learning and Information Retrieval, are examined. Then an evaluation methodology for statistical user models is proposed together with a set of metrics to specifically evaluate statistical user models. Finally, a benchmark for statistical user models is proposed, thus making possible to compare and replicate the evaluations. Thus, main contribution of this paper is to enable that several user model evaluations were comparable.

Agent technology has been applied in e-commerce to help coping with problems that arise due to its rapid growth. However, despite the amount of research in this area, the level of automation achieved is still limited. This is mainly due to the natural diversity existent in e-commerce environments, where agents may possess different conceptualizations about their needs and capabilities, giving rise to interoperability issues. In this paper we approach this problem and present the AEMOS system as a possible solution. AEMOS is an agent-based e-commerce platform that includes ontology matching services facilitating the interoperability between agents that have different conceptualizations about the same domain of knowledge. The system also explores emergent social networks in order to improve its efficiency by enhancing its ontology matching services and supporting agents in their decisions.

In this paper, we describe a system based on FIPA standards to help the process of advertisement, discovery, invocation and reuse of traffic information on the web of data. The use of semantic web services (SWS) can be exploited to improve the outcomes in the discovery process, allowing end users to specify their need using concepts not keywords. Most of the traffic information is generally recovered by end users through web forms that specify their requirements, and must refill each time the same parameters to obtain the updated value from the web sites. Using agents besides Service Oriented Architecture (SOA), we will achieve interoperability between systems and also automatize the process to obtain the data updated periodically. The amount of SWSs and the development of domain ontologies is increasing. However, in our domain (Intelligent Transport Systems) they are not very extended, so we had the need to develop our own ontologies. Our ontologies will allow to annotate semantically web services and also to translate the data provided by web forms to web services. To provide semantic search and retrieval information in real time we store the semantic WS profiles and our domain ontologies in a knowledge repository(KB) .This framework should improve the Human-Machine and the Machine-to-Machine interaction with the web of data, thanksto agents and the use of semantically annotated web service profiles.

This paper reports the development of a B2B platform for the personalization of the publicity transmitted during the program intervals. The platform as a whole must ensure that the intervals are filled with ads compatible with the profile, context and expressed interests of the viewers. The platform acts as an electronic marketplace for advertising agencies (content producer companies) and multimedia content providers (content distribution companies). The companies, once registered at the platform, are represented by agents who negotiate automatically the price of the interval timeslots according to the specified price range and adaptation behaviour. The candidate ads for a given viewer interval are selected through a matching mechanism between ad, viewer and the current context (program being watched) profiles. The overall architecture of the platform consists of a multiagent system organized into three layers consisting of: (

i

) interface agents that interact with companies; (

ii

) enterprise agents that model the companies, and (

iii

) delegate agents that negotiate a specific ad or interval. The negotiation follows a variant of the Iterated Contract Net Interaction Protocol (ICNIP) and is based on the price/s offered by the advertising agencies to occupy the viewer’s interval.

We tackle two symbolic approaches to bounded model checking (BMC) for an existential fragment of the soft real-time epistemic computation tree logic (RTECTLK) interpreted over interleaved interpreted systems. We describe a BDD-based BMC method for RTECTLK, and provide its experimental evaluation and comparison with a SAT-based BMC method. Moreover, we have attempted a comparison with MCMAS on several benchmarks.

Over the last decade, many methodologies for developing agent based systems have been developed, however no complete evaluation frameworks have been provided. Agent Oriented Software Engineering (AOSE) methodologies enhance the ability of software engineering to develop complex applications such as games; whilst it can be difficult for researchers to select an AOSE methodology suitable for a specific application. In this paper a new framework for evaluating different types of AOSE, such as qualitative and quantitative evaluations will be introduced. The framework assists researchers to select a preferable AOSE which could be used in a game development methodology. Furthermore the results from this evaluation framework can be used to determine the existing gaps in each methodology.

This paper presents a design method for optimal cancer chemotherapy schedules using genetic algorithm (GA). The main objective of chemotherapy is to reduce the number of cancer cells or eradicate completely, if possible, after a predefined time with minimum toxic side effects which is difficult to achieve using conventional clinical methods due to narrow therapeutic indices of chemotherapy drugs. Three drug scheduling schemes are proposed where GA is used to optimize the doses and schedules by satisfying several treatment constraints. Finally, a clinically relevant dose scheme with periodic nature is proposed. Here Martin’s model is used to test the designed treatment schedules and observe cell population, drug concentration and toxicity during the treatment. The number of cancer cells is found zero at the end of the treatment for all three cases with acceptable toxicity. So the proposed design method clearly shows effectiveness in planning chemotherapy schedules.

The future air transportation systems being developed under the NextGen (USA) and SESAR (European Commission) research initiatives will imply new levels of connectivity requirements between the concerned parties. Within this new aeronautical connectivity scenario, this paper proposes a new communication architecture for non-critical delay-tolerant communication (e.g. passenger data communications such as email and news services and non-critical telemetry data) based on mobile agents.Mobile agents carry both the user data and the routing algorithm used to decide the next hop in the path to the final destination. Therefore, mobile agents allow to increase the dynamism of the routing process. The proposed architecture constitutes an evolution of DTN (Delay Tolerant Networks), more flexible than the traditional layer-based approaches such as the Bundle Protocol and Licklider Transmission Protocol (LTP). This paper also presents the results obtained after network emulation and field experimentation of our proposed architecture.

Recently, the human stroke is gathering the focus as one of the diseases with higher mortality and social impact. In addition, it has a long-term treatment and high rehabilitation costs. Therefore, the early diagnosis of stroke can take advantage in avoiding the stroke itself or highly reducing its effects. Up to our knowledge, no previous study on stroke early diagnosis has been published in the literature. This study deals with the early detection of the stroke based on accelerometers and mobile devices. First, a discussion on the problem is presented and the design of the approach is outlined. In a first stage, it is necessary to determine what is the subject doing at any moment; thus, human activity recognition is performed. Afterwards, once the current activity is estimated, the detection of anomalous movements is proposed. Nevertheless, as there is no data available to learn the problem, a realistic proposal for simulating stroke episodes is presented, which lead us to draw the conclusions.

Mobile agent is the software paradigm that has been widely employed from comparatively small system to complex industrial system to realize their activities. The key features of a mobile agents’ management system are the agent management, communication management, agent mobility and monitoring management. The real challenge arises when developing such a system from scratch for each application domain. Such a process is effort and time consuming. System developers have to spend more time and effort to realize these features for each application domain. If system developers were to be provided by a software system approach that can adapt the system key features for different application domains, the development time and effort will be reduced considerably. In this paper, a mobile agents’ management system approach is developed that can be adapted by different application domains and is proved by implementing a case study traffic management system.

Gestural interfaces can be naturally adopted for ambient intelligence applications given the fact that most interactions we have with our environment are performed through our hands. Human hands can be extremely expressive and can perform very precise actions which in turn could benefit ambient interactions and enhance user experience. However, today’s common gesture-sensing technologies make heavily use of motion in the detriment of fine hand posture and finger movements. We propose in this work a glove-based input technique for inferring properties about grasped objects by using measurements of finger flexure alone to explore new interaction opportunities in intelligent environments.

Suboptimal restraint use is a prevalent problem worldwide. In developed countries injuries and deaths related to vehicle accidents persist despite increases in restraint use. In this study we investigate the emergence of patterns of restraint use in groups of agents and the population at large. Using age as an influential factor we simulate random encounters between group members where dominant individuals repeatedly alter the knowledge of less influential individuals. Belief spaces implemented as part of a cultural algorithm are used to preserve prevalent patterns of restraint use both at the group and population levels. The objective is to demonstrate restraint selection and use patterns emerging within a population and to determine whether a focus on influential members might have a positive effect towards optimal restraint use. We demonstrate that prominent patterns of behavior similar to the influential members of the groups do emerge both in the presence of social and cultural influence.

Consensus is a negotiation process where agents need to agree upon certain quantities of interest. The theoretical framework for solving consensus problems in dynamic networks of agents was formally introduced by Olfati-Saber and Murray, and is based on algebraic graph theory, matrix theory and control theory. Consensus problems are usually simulated using mathematical frameworks. However, implementation using multi-agent system platforms is a very difficult task due to problems such as synchronization, distributed finalization, and monitorization among others. The aim of this paper is to propose a protocol for the consensus agreement process in MAS in order to check the correctness of the algorithm and validate the protocol.

The provision of valuable e-government services depends upon the capacity to integrate the disperse provision of services by the public administration and thus upon the availability of interoperability platforms. These platforms are commonly built according to the principles of service oriented architectures, which raise the question of how to dynamically orchestrate services while preserving information security. Recently, it was presented an e-government interoperability model that preserves privacy during the dynamic orchestration of services. In this paper we present a prototype that implements that model using software agents. The model and the prototype are briefly described; an illustrative use case is presented; and the advantages of using software agents to implement the model are discussed.

User profiles are the base to obtain knowledge about users of recommender systems. We propose a context- and social-aware user profiling for audiovisual recommender systems that combines explicit preferences, implicit preferences and stereotypes modeling, taking advantage of information available in social networks and the current user context. We examine how the user profile is represented, acquired, built and updated; and how the profile information is exploited by an audiovisual recommender system that uses both collaborative filtering and the content-based method.

Biological data integration is currently one of the major challenges in the field of Bioinformatics. Several studies show that biological knowledge is growing at a continuous rate and is usually distributed among many databases. This paper presents an overview of data integration and the existing approaches that aim to solve this problem. The paper will also review the different ontology approaches that were introduced by researchers for representing biological data.

This paper presents a case-study regarding development challenges of multi-agent system for multi-robot system management based on our previous research of the given topic. During the development and implementation of multi-agent system prototype using JADE platform, several implementation challenges regarding messaging system were faced. These challenges may negatively impact system maintenance, burden system evolution and also cause performance issues. The latter is of special importance in the context of multi-robot systems that operate under real-time constraints. In this paper we adopt our previous research as a case study and share challenges faced during prototype multi-robot system development. We believe that potential drawbacks and pitfalls of multi-agent system development such as challenges identified in this paper should be considered with great care especially when applying multi-agent systems to real-time constrained applications such as multi-robot systems.

In this paper we propose an intelligent control scheme based on a multiagent system composed of two main components: a logical planner and forecaster obtained by machine learning techniques. The chosen benchmark application lays in the field of smart energy management, were we consider the concept of a prosumer node. We discuss a case-study where a simplified multi-agent system is applied to the supervision of an air conditioner, leading to an energy saving of about 17%.

The fast increasing amount of articles published in the biomedical field is creating difficulties in the way this wealth of information can be efficiently exploited by researchers. As a way of overcoming these limitations and potentiating a more efficient use of the literature, we propose an approach for structuring the results of a literature search based on the latent semantic information extracted from a corpus. Moreover, we show how the results of the Latent Semantic Analysis method can be adapted so as to evidence differences between results of different searches. We also propose different visualization techniques that can be applied to explore these results. Used in combination, these techniques could empower users with tools for literature guided knowledge exploration and discovery.

Energy systems worldwide are complex and challenging environments. Multi-agent based simulation platforms are increasing at a high rate, as they show to be a good option to study many issues related to these systems, as well as the involved players at act in this domain. In this scope the authors research group has developed three multi-agent systems: MASCEM, which simulates the electricity markets; ALBidS that works as a decision support system for market players; and MASGriP, which simulates the internal operations of smart grids. To take better advantage of these systems, their integration is mandatory. For this reason, is proposed the development of an upper-ontology which allows an easier cooperation and adequate communication between them. Additionally, the concepts and rules defined by this ontology can be expanded and complemented by the needs of other simulation and real systems in the same areas as the mentioned systems. Each system’s particular ontology must be extended from this top-level ontology.

This paper proposes a practical methodology to implement a mobile robot agent based on a Google Android Smartphone. The main computational unit of the robot agent is a Smartphone connected through USB to a control motor board that drives two motors and one stick. The agent program structure is implemented using multi-threading methods with shared memory instances. The agent uses the Smartphone camera to obtain images and to apply image processing algorithms in order to obtain profitable information of its environment. Moreover, the robot can use the sensors embedded in the Smartphone to gather more information of the environment. This paper describes the methodology used and the advantages of developing a robot agent based on a Smartphone.

The new model of labor relations established by the Spanish Royal Decree-Law (3/2012) on urgent measures for labor reform has among its objectives the promotion of inclusion in the labor market of more advantaged groups, including the people with disabilities. This paper presents a cloud-based platform aimed at obtaining an on-line workspace to provide facilities to inform, train and evaluate the competencies of disabled people, and more specifically those skills required to facilitate the labor integration of individuals with auditory disabilities. This platform presented in this paper has been tested in a real environment and the results obtained are promising.

In the original version, the fourth and fifth author names were missed in this chapter. The names are given below:

Adrián Sánchez-Carmona

1

and Joan Borrell

1

1

Department of Information and Communication Engineering, Universitat Autònoma de Barcelona, Edifici Q. Bellaterra, Barcelona, Spain