Search

Trabajo presentado en el International Conference on Electrical, Computer and Energy Technologies (ICECET), celebrada en Cape Town (Sudáfrica), los días 9 y 10 de diciembre de 2021, Nowadays, competitiveness is a reality in all industrial fields and the plastic injection industry is not an exception. Due to the complex intrinsic changes that the parameters undergo during the injection process, it is essential to monitor the parameters that influence the quality of the final part to guarantee a superior quality of service provided to customers. Quality requirements impose the development of intelligent systems capable to detect defects in the produced parts. This article presents a first step towards building an intelligent system for classifying the quality of produced parts. The basic approach of this work is machine learning methods (Artificial Neural Networks and Support Vector Machines) and techniques that combine the two previous approaches (ensemble method). These are trained as classifiers to detect conformity or even defect types in parts. The data analyzed were collected at a plastic injection company in Portugal. The results show that these techniques are capable of incorporating the non-linear relationships between the process variables, which allows for a good accuracy ( ¿ 99%) in the identification of defects. Although these techniques present good accuracy, we show that taking into account the history of the last cycles and the use of combined techniques improves even further the performance. The approach presented in this article has a number of potential advantages for online predicting of parts quality in injection molding processes.

Maintenance is the process of preserving the good condition of a system to ensure its reliability and availability to perform specific operations. The way maintenance is nowadays performed in industry is changing thanks to the increasing availability of data and condition assessment methods. Soft sensors have been widely used over last years to monitor industrial processes and to predict process variables that are difficult to measured. The main objective of this study is to monitor and evaluate the condition of the compressor in a particular industrial gas turbine by developing a soft sensor following an autoencoder architecture. The data used to monitor and analyze its condition were captured by several sensors located along the compressor for around five years. The condition assessment of an industrial gas turbine compressor reveals significant changes over time, as well as a drift in its performance. These results lead to a qualitative indicator of the compressor behavior in long-term performance.

There has been an increasing relevance of the cultural sector in the economic and social development of different countries. However, this sector continues without much input from multi-criteria decision-making (MDCM) techniques and sustainability analysis, which are widely used in other sectors. This paper proposes an MCDM model to assess the sustainability of a musical institution&rsquo, s program. To define the parameters of the proposed model, qualitative interviews with relevant representatives of Catalan cultural institutions and highly recognized professionals in the sector were performed. The content of the 2015&ndash, 2016 season of the &lsquo, Palau de la M&uacute, sica Catalana&rsquo, a relevant Catalan musical institution located in Barcelona, was used as a case study to empirically test the method. The method allows the calculation of a season value index (SVI), which serves to make more sustainable decisions on musical season programs according to the established criteria. The sensitivity analysis carried out for different scenarios shows the robustness of the method. The research suggests that more complex decision settings, such as MCDM methods that are widely used in other sectors, can be easily applied to the sustainable management of any type of cultural institution. To the best of the authors&rsquo, knowledge, this method was never applied to a cultural institution and with real data.

Este trabajo de fin de grado consiste en el desarrollo de un sistema en el lenguaje Python que realiza inferencia lógica, implementando una redefinición del dialecto de reglas de producción del estándar Rule Interchange Format, un lenguaje para la unificación de documentos de reglas en diferentes lenguajes. En adición, el motor de inferencia desarrollado usa ontologías escritas en un subconjunto de OWL 2, e implementa una versión del algoritmo Rete para la realización de matching en reglas. Este trabajo se engloba en las tecnologías de la web semántica, y pretende ser una alternativa cómoda y fácil de usar que no requiera de traductores entre lenguajes de reglas. También se ha realizado un estudio de rendimiento del motor, en el que se observa el peso que tiene el pobre rendimiento de Python sobre el motor y la influencia del tamaño de la red Rete sobre el tiempo de ejecución. This end-of-degree project consists in the development in the language Python of a system that performs logical inference, implementing a redefinition of the production rule dialect of the Rule Interchange Format (RIF) standard, a language to unify rule documents written in different languages. Furthermore, the developed inference engine uses ontologies written in a subset of OWL 2 and implements a version of the Rete algorithm to perform rule matching. This project is encompassed in the semantic web technologies, and pretends to be an easy and convenient alternative that does not require translators between rule languages. A performance study has also been done, in which the weight of the poor performance of Python over the engine and the influence of the size of the Rete network over the execution time can be observed.

Artificial Intelligence (AI) is among most important fields of knowledge and applications in a large variety of domains. Recently however, it has become a trending research topic propelled by Cloud computing, social networks and alike. Terms like machine learning, "Big Data" and artificial neural networks very frequently appear not only in scientific media but even in the mass media. In this project, we aim to design, implement and evaluate an AI technique, namely, deep learning, which has become very popular for face recognition. The problem is formulated from an engineering perspective: to design a small size system based on Raspberry Pi and an attached camera to it to detect and recognise human faces in real time. It should be mentioned that while for humans face recognition is a trivial task, we do it every day and with a full accuracy, for a computer, this is complex task. Recent applications from many industries show a large potential of intelligent systems that need to recognise faces with high accuracy. The thesis is essentially structured into two main parts. In the first part we formulate the problem, analyse potential solutions and propose a solution for its resolution. In the second part of the project we develop the proposed solution into an implementation of an intelligent system for a computationally limited and physical portable device (Raspberry Pi). The solution is empirically evaluated in terms of accuracy and performance using real data sets. The relevance of using such a small size intelligent system relies in the fact that this application can be installed in other devices, such as drones, easily, at low cost and without compromising the performance and speed of the said intelligent system.

This work is focused on the data based modelling and monitoring of a family of modular systems that have multiple replicated structures with the same nominal variables and show temporal behaviour with certain periodicity. These characteristics are present in many systems in numerous fields such as the construction or energy sector or in industry. The challenge for these systems is to be able to exploit the redundancy in both time and the physical structure. In this paper the authors present a method for representing such granular systems using N-dimensional data arrays which are then transformed into the suitable 2-dimensional matrices required to perform statistical processing. Here, the focus is on pre-processing data using a non-unique folding-unfolding algorithm in a way that allows for different statistical models to be built in accordance with the monitoring requirements selected. Principal Component Analysis (PCA) is assumed as the underlying principle to carry out the monitoring. Thus, the method extends the Unfold Principal Component Analysis (Unfold-PCA or Multiway PCA), applied to 3D arrays, to deal with N-dimensional matrices. However, this method is general enough to be applied in other multivariate monitoring strategies. Two of examples in the area of energy efficiency illustrate the application of the method for modelling. Both examples illustrate how when a unique data-set folded and unfolded in different ways, it offers different modelling capabilities. Moreover, one of the examples is extended to exploit real data. In this case, real data collected over a two-year period from a multi-housing social-building located in down town Barcelona (Catalonia) has been used This work has been carried out by the research group eXIT (http://exit.udg.edu), funded through the following projects: MESC project(Ref. DPI2013-47450-C21-R) and its continuation CROWDSAVING (Ref.TIN2016-79726-C2-2-R), both funded by the Spanish Ministerio de Industria y Competitividad within the Research, Development and Innovation Program oriented towards the Societal Challenges, and also the project Hit2Gap of the Horizon 2020 research and innovation program under grant agreement N680708. The author Llorenç Burgas would also like to thank Girona University for their support through the competitive grant for doctoral formation IFUdG2016

In this master thesis, the problem of identification and control of a foam process in a waste treatment plant will be address. Using real data, an experimental model relating the anti-foam product injected using a pump and the measure of foam level measured used computer vision will be developed using identification techniques. Once the model have been obtained, a control law that aims at minimizing the foam level will be developed. The control system will be tested and implemented in a real waste treatment plant.

In order to create better decisions for business analytics, organizations increasingly use external structured, semi-structured, and unstructured data in addition to the (mostly structured) internal data. Current Extract-Transform-Load (ETL) tools are not suitable for this “open world scenario” because they do not consider semantic issues in the integration processing. Current ETL tools neither support processing semantic data nor create a semantic Data Warehouse (DW), a repository of semantically integrated data. This paper describes our programmable Semantic ETL (SETL) framework. SETL builds on Semantic Web (SW) standards and tools and supports developers by offering a number of powerful modules, classes, and methods for (dimensional and semantic) DW constructs and tasks. Thus it supports semantic data sources in addition to traditional data sources, semantic integration, and creating or publishing a semantic (multidimensional) DW in terms of a knowledge base. A comprehensive experimental evaluation comparing SETL to a solution made with traditional tools (requiring much more hand-coding) on a concrete use case, shows that SETL provides better programmer productivity, knowledge base quality, and performance.

Display Omitted Risk assessment of developing an inherited illness using pedigree information.Methods, structured and statistic indices are defined for knowledge-based systems.Real information from a breast cancer database is used for experimentation.Implemented and integrated into eXiT?CBR experimentation framework as a plugin. The use of family information is a key issue to deal with inheritance illnesses. This kind of information use to come in the form of pedigree files, which contain structured information as tree or graphs, which explains the family relationships. Knowledge-based systems should incorporate the information gathered by pedigree tools to assess medical decision making. In this paper, we propose a method to achieve such a goal, which consists on the definition of new indicators, and methods and rules to compute them from family trees. The method is illustrated with several case studies. We provide information about its implementation and integration on a case-based reasoning tool. The method has been experimentally tested with breast cancer diagnosis data. The results show the feasibility of our methodology.

Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2017, Director: Eloi Puertas i Prats, Currently, thanks to the Internet anyone has access to a large amount of data and for this reason it is essential to create new systems that help to understand that information in a little time. Recommender Systems are engines which allow to filter the information depending on people’s interests. There are different kinds of Recommenders and each of them has a different purpose. In this project, a case of use of a Collaborative filtering Recommender System is introduced employing every interaction users do while surfing the Stilavia web site as data input. In order to carry out this task, some scoring functions are required to generate a model. This model will be extrapolated throughout the whole dataset space thanks to a Machine Learning algorithm called Alternating Least Squares (ALS) that is available in a library of the Apache Spark framework. Lastly, the results of each scoring function will be tested and evaluated employing a statistic estimator.

The complexity of Business Intelligence activities has driven the proposal of several approaches for the effective modeling of Extract-Transform-Load (ETL) processes, based on the conceptual abstraction of their operations. Apart from fostering automation and maintainability, such modeling also provides the building blocks to identify and represent frequently recurring patterns. Despite some existing work on classifying ETL components and functionality archetypes, the issue of systematically mining such patterns and their connection to quality attributes such as performance has not yet been addressed. In this work, we propose a methodology for the identification of ETL structural patterns. We logically model the ETL workflows using labeled graphs and employ graph algorithms to identify candidate patterns and to recognize them on different workflows. We showcase our approach through a use case that is applied on implemented ETL processes from the TPC-DI specification and we present mined ETL patterns. Decomposing ETL processes to identified patterns, our approach provides a stepping stone for the automatic translation of ETL logical models to their conceptual representation and to generate fine-grained cost models at the granularity level of patterns.

The final publication is available at IOS Press through http://dx.doi.org/10.3233/THC-161289 BACKGROUND AND OBJECTIVE: Major depressive disorder causes more human suffering than any other disease affecting humankind. It has a high prevalence and it is predicted that it will be among the three leading causes of disease burden by 2030. The prevalence of depression, all of its social and personal costs, and its recurrent characteristics, put heavy constraints on the ability of the public healthcare system to provide sufficient support for patients with depression. In this research, a model for continuous monitoring and tracking of depression in both short-term and long-term periods is presented. This model is based on a new qualitative reasoning approach. METHOD: This paper describes the patient assessment unit of a major depression monitoring system that has three modules: a patient progress module, based on a qualitative reasoning model; an analysis module, based on expert knowledge and a rules-based system; and the communication module. These modules base their reasoning mainly on data of the patient's mood and life events that are obtained from the patient's responses to specific questionnaires (PHQ-9, M.I.N.I. and Brugha). The patient assessment unit provides synthetic and useful information for both patients and physicians, keeps them informed of the progress of patients, and alerts them in the case of necessity. RESULTS: A set of hypothetical patients has been defined based on clinically possible cases in order to perform a complete scenario evaluation. The results that have been verified by psychiatrists suggest the utility of the platform. CONCLUSION: The proposed major depression monitoring system takes advantage of current technologies and facilitates more frequent follow-up of the progress of patients during their home stay after being diagnosed with depression by a psychiatrist.

Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2016, Director: Salamó Llorente, Maria, In recent years there has been an exponential growth of the new technologies. The great expansion of internet and easy access to the network has increased the number of internet data and behold the context of the problem: the difficulty of finding the right information.Nowadays million of users access for internet searching products and information, cause of this it’s necessary the use of recommenders which allow users to find out what they are looking for. At the same time, most interactive applications of recommenders are a key business factor for those who sell their products through the network, enabling them to increase their incomes if they use a recommender system suitable. This project is focused on recommender systems based on critics (critiquing), which allow the user to indicate by a feedback, the characteristics of the product you are looking for in order to provide products according to their needs and preferences. Continuing the methodology Critiquing, a study of standard algorithms (STD) and incremental (IC) will be performed, as well as those belonging to the category Experience-based Critiquing: EBC, HAC, HGR, HOR and Graph based, all based on Unit Critiquing. In addition, the algorithms previously mentioned will be implemented to be used by users through Compound Critiques. These allow the user to provide simultaneously a feedback about multiples characteristics of the product. This fact differentiates them from Unit Critiques, which only provide information on a single characteristic. In this project an analysis algorithms will be held Unit Critiques vs. Compound Critiques. The result of this analysis will allow to reflect optimization that occurs in a recommender process if Compound Critiques are used.

Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2016, Director: Maria Salamó Llorente, The maintanence of a case data base is very important since it helps to improve efficiency in the Case-Based Reasoning (CBR). There are different types of maintenance and many algorithms, and the combination of both can make the CBR a very powerful tool. This Final Project consists on the implementation of different algorithms of memory of cases maintaining those that already exist in literature and that are applied on Case Based Reasoning systems. Specifically this project adds the following algorithms to the CBR_KM library: Repeated Edited Nearest Neighbor (RENN), All k-NN (ANN), Blame-Based Noise Reduction (BBNR) and Conservative Redundary Reduction (CRR). All these are pre-process algorithms, which are applied before the use of the base case in order to reduce its size removing cases that are considered harmful or less useful. The policy of each algorithm is different and it is focused on a particular aspect. The evaluation of algorithms will be made with different bases of cases from the UCI Repository and the results will be evaluated taking into account several criteria such as: final size of the case base, accuracy, rate of cases retention and percentage of forgotten cases.

Background: Reverse engineering of gene regulatory networks (GRNs) from gene expression data is a classical challenge in systems biology. Thanks to high-throughput technologies, a massive amount of gene-expression data has been accumulated in the public repositories. Modelling GRNs from multiple experiments (also called integrative analysis) has; therefore, naturally become a standard procedure in modern computational biology. Indeed, such analysis is usually more robust than the traditional approaches, which suffer from experimental biases and the low number of samples by analysing individual datasets. To date, there are mainly two strategies for the problem of interest: the first one (“data merging”) merges all datasets together and then infers a GRN whereas the other (“networks ensemble”) infers GRNs from every dataset separately and then aggregates them using some ensemble rules (such as ranksum or weightsum). Unfortunately, a thorough comparison of these two approaches is lacking. Results: In this work, we are going to present another meta-analysis approach for inferring GRNs from multiple studies. Our proposed meta-analysis approach, adapted to methods based on pairwise measures such as correlation or mutual information, consists of two steps: aggregating matrices of the pairwise measures from every dataset followed by extracting the network from the meta-matrix. Afterwards, we evaluate the performance of the two commonly used approaches mentioned above and our presented approach with a systematic set of experiments based on in silico benchmarks. Conclusions: We proposed a first systematic evaluation of different strategies for reverse engineering GRNs from multiple datasets. Experiment results strongly suggest that assembling matrices of pairwise dependencies is a better strategy for network inference than the two commonly used ones., SCOPUS: ar.j, info:eu-repo/semantics/published

This Bachelor Thesis aims to draw a first approach for building a highly versatile tool to simulate systems with adaptive modularity, that is, systems that are able to adapt their functional capabilities according to the type of mission or scanario they are involved into. The simulation framework follows an agent-based strategy, in which several agents with different roles (launcher, scenario, active agent) manage to accomplish missions. These missions consist of performing certain tasks of target observation in one or more spectral region (e.g, microwave, infrared, visible) with a certain set of temporal requirements and a score value. The goal of the system is to perform as many high-value observations as possible. If certain tasks require more requirements than a single agent can provide, it can build a coalition with other agents, which do carry enough resources for performing the task. These coalitions are temporal associations of agents that can be split or merged as considered by them, thus allowing the agents to adapt their modularity for better performing their mission. The main achievements attained in this thesis can be divided in two blocks: on the one hand, we have developed a support program using an open source Java-based platform that executes simulations upon a large set of possible scenarios (Turtlekit ), displays what agents are doing in real-time in a particular simulated case and stores their performance data for its posterior analysis; on the other hand, a widely known decentralized task allocation algorithm (Coupled Constraint Consensus Bundle-Based Algorithm, CCBBA) has been implemented and modiffied as necessary for handling speciffic constraints regarding adaptive modularity.

El objetivo de este proyecto es el diseño de un sistema SCADA que de manera centralizada permita supervisar varios procesos a la vez. Permitiendo enviar órdenes en todo momento a cada proceso y recibiendo en tiempo real la información que cada proceso envía.

Tener datos y saber hacer un correcto análisis e interpretación de estos es una de las claves en el éxito empresarial. Las empresas son conscientes de ello e invierten grandes recursos, ya sean de personal o material, pero cada vez el volumen de datos a tratar es mayor y se necesitan sistemas más eficientes. El proyecto que se explica a continuación en detalle, soluciona esta problemática con un sistema capaz de obtener, estructurar, distribuir y analizar los datos de la interacción de un usuario por un portal web. Tenir dades i saber fer un correcte anàlisi i interpretació d’aquestes és una de les claus en l’èxit empresarial. Les empreses són conscients d’això i inverteixen grans recursos, ja siguin de personal o material, però cada cop el volum de dades a tractar és més gran i es necessiten sistemes més eficients. El projecte que s’explica a continuació en detall, soluciona aquesta problemàtica amb un sistema capaç d’obtenir, estructurar, distribuir i analitzar les dades de la interacció d’un usuari per un portal web. Collecting data and knowing how to make a proper analysis and interpretation of it is the key to the business success. Corporations are aware of this and invest big amount of resources, personal and material, but as the amount of data to analyse increases, more efficient systems are needed. The following project, solves these problems with a system that is able to obtain, structure, distribute and analyse the data provided by the interaction of a user in a website.

Roles and positions are structural components in complex social systems which group actors based on how similarly they are connected to the rest of the actors. Role and position detection methods have been successfully used to evaluate and understand the dynamics of social networks and the behavior of their members. However, actor similarities used to detect positions have been based on pairwise comparisons so far: e.g., structural equivalence states that Alice and Bob are in the same position if they are both connected or not to the same other actors in the network, one by one. In this work we present a new framework to find positions and roles using comparisons between actors and sets of actors instead of just using pairwise comparisons. In this way we enable the usage of many more measures of similarity inside position and role detection methods, e.g., based on distances, community structure, triangles and cliques. As a result, we can identify new types of easily interpretable positions. Additionally, the proposed idea can be adapted to more complex models like hypergraphs or multiplex/multi-relational networks. We have evaluated our work on both synthetic and real data, using several existing and new similarity measures and providing both qualitative and quantitative evidence of the new possibilities enabled by our approach.

Neuro-oncologists must ultimately rely on their acquired knowledge and accumulated experience to undertake the sensitive task of brain tumour diagnosis. This task strongly depends on indirect, non-invasive measurements, which are the source of valuable data in the form of signals and images. Expert radiologists should benefit from their use as part of an at least partially automated computer-based medical decision support system. This paper focuses on Magnetic Resonance Spectroscopy signal analysis and illustrates a method that combines Gaussian Decomposition, dimensionality reduction by Moving Window with Variance Analysis and classification using adaptively regularized Artificial Neural Networks. The method yields encouraging results in the task of binary classification of human brain tumours, even for tumour types that have seldom been analyzed from this viewpoint.

The project goal is the study and development of specific techniques for obtaining user preferences from several interaction mechanisms with the objective of making recommendations in a conversational group recommender system. In this project, we capture and model user's preferences.

Nowadays, a huge amount of high resolution satellite images are freely available. Such images allow researchers in environmental sciences to study the different natural habitats and farming practices in a remote way. However, satellite images content strongly depends on the season of the acquisition. Due to the periodicity of natural and agricultural dynamics throughout seasons, sequential patterns arise as a new opportunity to model the behaviour of these environments. In this paper, we describe some preliminary results obtained with a new framework for studying spatiotemporal evolutions over natural and agricultural areas using k-partite graphs and sequential patterns extracted from segmented Landsat images.

Treballs Finals de Grau de Matemàtiques, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2014, Director: Maria Salamó Llorente, Case-Based Reasoning, CBR, is a kind of reasoning based on solving new problems using solutions obtained of previous problems. As something natural, human beings use this case-based reasoning to face problems that arise every day. For instance: the mechanic that repairs a car remembering another car having the same problems is using case-based reasoning. That is why CBR is called an expert system since is a system that tries to imitate the behavior of an expert human being, that is to say, it follows the same steps that a human being would take trying to solve different problems. Continuing with the example of the mechanic, an expert system in charge of diagnosing mechanical car problems would follow the same steps that a mechanic does. In this way, the advantages of creating a CBR are clear: to make easier experts’ task use CBR as a help system on decision making. In addition, human beings do something more: we update our knowledge; that is to say, we forget useless information and we learn and remember useful one. This is as important to experts systems as it is to human beings. On the one hand, an expert system that does not get new information is a high quality system nowadays but it will become obsolete in the future. On the other hand, an expert system that does not forget information and keeps on getting more and more will produce a huge database, excessively big to be processed in order to solve new problems or a mediocre expert system by means of cutting back the amount of information used.

Explicacions per a les prediccions dels sistemes de recomanació. Machine learning. [ANGLÈS] An explanation system helps user to identify those items from a variety of choices that best match their needs and preferences. Moreover, it has the ability to establish trust and confidence with the user providing some benefits. In this project developed in Sony Deutschland, I present the results of my work designing and developing an explanation engine for a real recommender system, in concrete for the PlayStation Store. [CASTELLÀ] Un sistema de explicaciones ayuda a los usuarios a identificar aquellos productos que mejor satisfacen sus necesidades entre una amplia variedad. Además tienen la habilidad de establecer una relación con el usuario y transmitirle un cierto grado de confianza con el sistema aportando una serie de beneficios. En este proyecto desarrollado a Sony Deutschland, expongo los resultados de mi trabajo diseñando y desarrollando un motor de explicaciones para sistemas el sistema de recomendación en un entorno de comercio electrónico real, concretamente para la PlayStation Store. [CATALÀ] Un sistema d'explicacions ajuda als usuaris a identificar aquest productes que millor satisfan les seves necessitats entre una amplia varietat. A més, tenen la habilitat de establir una relació amb l?usuari i transmetre-li un grau de confiança amb el sistema aportant una sèrie de beneficis. En aquest projecte realitzat a Sony Deutschland, exposo els resultats del meu treball dissenyant i desenvolupant un motor de explicacions per a sistemes de recomanació en un entorn de comerç electrònic real, concretament per a la PlayStation Store.

[ANGLÈS] The space exploration endeavor started in 1957 with the launch and operation of the first manmade satellite, the URSS Sputnik 1. Since then, multiple space programs have been developed, pushing the limits of technology and science but foremost unveiling the mysteries of the universe. In all these cases, the need for flexible and reliable communication systems has been primordial, allowing the return of collected science data and, when necessary, ensuring the well-being and safety of astronauts. To that end, multiple space communication networks have been globally deployed, be it through geographically distributed ground assets or through space relay satellites. Until now most of these systems have relied upon mature technology standards that have been adapted to the specific needs of particular missions and customers. Nevertheless, current trends in the space programs suggest that a shift of paradigm is needed: an Internet-like space network would increase the capacity and reliability of an interplanetary network while dramatically reducing its overall costs. In this context, the System Architecting Paradigm can be a good starting point. Through its formal decomposition of the system, it can help determine the architecturally distinguishing decisions and identify potential areas of commonality and cost reduction. This thesis presents a general framework to evaluate space communication relay systems for the near Earth domain. It indicates the sources of complexity in the modeling process, and discusses the validity and appropriateness of past approaches to the problem. In particular, it proposes a discussion of current models vis-à-vis the System Architecting Paradigm and how they fit into tradespace exploration studies. Next, the thesis introduces a computational performance model for the analysis and fast simulation of space relay satellite systems. The tool takes advantage of a specifically built-in rule-based expert system for storing the constitutive elements of the architecture and perform logical interactions between them. Analogously, it uses numerical models to assess the network topology over a given timeframe, perform physical layer computations and calculate plausible schedules for the overall system. In particular, it presents a newly developed heuristic scheduler that guarantees prioritization of specific missions and services while ensuring manageable computational times. [CASTELLÀ] El inicio de la carrera espacial se inició en 1957 con el lanzamiento y operación del primer satélite artificial, el Sputnik 1 de la URSS. Desde entonces se han desarrollado múltiples programas espaciales que han llevado al límite tanto la tecnología como la ciencia y han permitido desvelar los misterios del universo. En todos estos casos, la necesidad de sistemas de comunicación flexibles y fiables ha sido primordial con el fin de asegurar el retorno de los datos científicos recopilados y, en ciertos casos, garantizar la seguridad de los astronautas. Como consecuencia, múltiples redes de comunicaciones espaciales han sido desplegadas, ya sea a través de antenas globalmente distribuidas a través de la superficie terrestre o mediante satélites repetidores. Hasta ahora la mayoría de estos sistemas se ha basado en estándares tecnológicos maduros y testeados, los cuales se han adaptado con el fin de satisfacer las necesidades específicas de cada misión y cliente. Sin embargo, las tendencias actuales en el diseño de los nuevos programas espaciales indica que un cambio de paradigma es necesario: una red espacial a imagen de Internet permitiría incrementar la capacidad y fiabilidad de las comunicaciones interplanetarias y, a la vez, reducir dramáticamente sus costes. En este contexto, el paradigma de arquitectura de sistemas puede ser un buen punto de partida. Mediante la descomposición formal del sistema, puede ayudar a determinar las decisiones que tienen un impacto cabal en el diseño de la arquitectura así como identificar las áreas con tecnologías similares y de menor coste. Esta tesis presenta un marco teórico general para evaluar sistemas de comunicaciones espaciales para misiones que orbitan la Tierra. Adicionalmente, la tesis indica los principales orígenes de complejidad durante el modelado del sistema y presenta una discusión sobre la validez de anteriores estrategias para analizar el problema. En concreto, propone una comparación de anteriores modelos respecto el paradigma de arquitectura de sistemas y su grado de adecuación para evaluar y comprar arquitecturas. A continuación, la tesis introduce un modelo computacional para simular y evaluar el rendimiento de sistemas de repetidores por satélite. La herramienta utiliza un rule-based expert system específicamente diseñado con el fin de almacenar los principales elementos constitutivos de la arquitectura y comprender las interacciones lógicas entre ellos. Análogamente, el modelo usa métodos numéricos con el fin de calcular la evolución temporal de la topología de la red en un determinado intervalo de tiempo, así como su capa física y un posible programa de contactos. En concreto, presenta un nuevo scheduler heurístico que garantiza la correcta ordenación de las misiones y servicios a la vez que asegura un tiempo computacional aceptable. [CATALÀ] L'inici de la cursa espacial va iniciar-se l'any 1957 amb el llançament i operació del primer satèl·lit artificial, l'Sputnik 1 de la URSS. Des d'aleshores s'han dut a terme múltiples programes espacials que han portat al límit tant la tecnologia com la ciència i han permès desvelar els misteris de l'univers. En tots aquests casos, la necessitat de sistemes de comunicació flexibles i fiables ha sigut primordial per tal d'assegurar el retorn de les dades científiques recopilades i, en certs casos, garantir el benestar i seguretat dels astronautes. Com a conseqüència, múltiples xarxes de comunicacions espacials han sigut desplegades, ja sigui a través d'antenes globalment distribuïdes a través de la superfície terrestre o mitjançant satèl·lits repetidors. Fins ara la majoria d'aquests sistemes s'han basat en estàndards tecnològics madurs i testats, els quals s'han adaptat per tal de satisfer les necessitats específiques de cada missió i client. Això no obstant, les tendències actuals en el disseny dels nous programes espacials indica que un canvi de paradigma és necessari: una xarxa espacial a imatge d'Internet permetria incrementar la capacitat i fiabilitat de les comunicacions interplanetàries i, alhora, reduir dramàticament els seu costs. En aquest context, el paradigma d'arquitectura de sistemes pot ser un bon punt de partida. Mitjançant la descomposició formal del sistema, pot ajudar a determinar les decisions que tenen un impacte cabdal en el disseny de l'arquitectura així com permetre identificar àrees amb tecnologies similars i de menor cost. Aquesta tesi presenta un marc teòric general per avaluar sistemes de comunicacions espacials per missions orbitant la Terra. Addicionalment, la tesi indica els principals orígens de complexitat durant el modelatge del sistema i presenta una discussió sobre la validesa d'anteriors estratègies per analitzar el problema. En concret, proposa una comparació d'anteriors models respecte el paradigma d'arquitectura de sistemes i el seu grau d'adequació per avaluar i comparar arquitectures. A continuació, la tesi introdueix un model computacional per simular i avaluar el rendiment de sistemes de repetidors per satèl·lit. L'eina empra un rule-based expert system específicament dissenyat per tal d'emmagatzemar els principals elements constitutius de l'arquitectura i comprendre les interaccions lògiques entre ells. Anàlogament, el model utilitza mètodes numèrics per tal de calcular l'evolució temporal de la topologia de la xarxa en un determinat interval de temps, així com calcular la seva capa física i un possible programa de contactes. En concret, presenta un nou scheduler heurístic que garanteix la correcte ordenació de les missions i serveis tot assegurant un temps de computació acceptable.

Recommender engines (REs) also known as recommender systems are software tools and techniques providing suggestions to a user. The suggestions provided are aimed at supporting their users in various decision making processes such as what items to buy, what music to listen, what profiles to browse, or what news to read. This thesis studies the feasibility of the integration of a recommender engine as a module in a Liferay portal, and shows the process of its design and implementation using the Apache Mahout library. As such our work tackles two major problems which are: (1) the implementation of the recommender engine using the Apache Mahout library, and (2) the integration of the recommender in Liferay portal. Prior to the design of the application and the decisions made at the design stage, are shown the analysis of the requirements for the recommender system in the context of the business. The decisions made at design time are explained and the risks involved are analyzed and mitigated. Finally conclusions regarding the integration of the recommender engine in Liferay portal are made. This study was performed as part of an internship at Everis (www.everis.com).