TECHNOLOGICAL SCIENCES (5)

Author: MARTÍNEZ VELANDÍA DIEGO ALFONSO.
Year: 2005.
University: POLITÉCNICA DE VALENCIA [www.upv.es].
Place of defense: UNIVERSIDAD POLITÉCNICA DE VALENCIA.
Place of preparation: UNIVERSIDAD POLITÉCNICA DE VALENCIA.

Summary: Smoke silica puzolana very reactive, is a byproduct of the manufacture of elemental silicon (Si) and alloys ferrosilíceas in electric arc furnaces. The smoke was subjected to silica densification processes to be transported more easily and thus reduce the overall costs of distribution, and also to reduce occupational hazards through inhalation of the material in its natural state densificado not. Because of this treatment of smoke silica (HS) has been reduced its reactivity puzolánica therefore, in this paper we have the activation of smoke silica densificado (HSD) by different physical therapies, which aim to achieve essentially the material , to be incorporated into cement pastes and mortars, increase their performance with respect to the smoke densificado commercial silica (HSD). This work is done in five phases whose approaches and results described briefly below: * First phase: Characterization of different samples of silica fume, such as its chemical composition, mineralogical composition, morphology, density, particle size and surface specific. It was noted that all silica fume densificados studied show similar specific surfaces by BET, however, the size of each varies considerably, due to the presence of agglomerated microparticles and submicropartículas SM. * Second phase: They apply different physical treatments smoke silica to achieve the unbundling of large particles densificadas. The first of these, milling of materials in small and large quantities, the second milling joint smoke silica densificado (HSD) with fly ash (HP), which can act as an agent of the divisive agglomerates HS; and finally, the application of an ultrasonic bath samples of silica fume densificado (HSD) prepared in the form of slurry. It was generally observed a low effectiveness of treatments performed. * Third phase: Use an ultrasonic probe commercial laboratory scale, for the breakdown of the material, which was seen generating high quantities of particles submicrométricas SM. At this stage are conducted several studies to define the optimal parameters of sonication, power, time, number of shows with which to achieve a material thinner and higher reactivity. * Fourth stage: Comparison of the results obtendios with smoke silica sonicado (HSS) and smoke silica densificado (HSD). It was noted that the mortar with HSS trabajabilidad and increases its fluidity with respect to the mortar with HSD. There is also a very important increase of the mechanical resistance of mortar with HSS over the mortar with HSD, thanks to the greater reactivity puzolánica of HSS, as evidenced by the high levels of fixing protlandita. * Fifth stage: Calculation factor effectiveness cementante factor K.del HSS and HSD. The study shows that the effectiveness of HSS regarding the HSD is far superior, justifying treatment with the ultrasonic probe.

Author: GARZON SERRANO OLGA PATRICIA.
Year: 2005.
University: POLITÉCNICA DE VALENCIA [www.upv.es].
Place of defense: UNIVERSIDAD POLITÉCNICA DE VALENCIA.
Place of preparation: UNIVERSIDAD POLITÉCNICA DE VALENCIA.

Summary: The investigation into the construction industry has been directed basically to constructive techniques, construction materials and to a lesser extent, the management of enterprises. However, there has been little emphasis on the knowledge of other aspects to learn more about the construction market, and more specifically, the developer. This line of research can offer companies promoter analysis tools, prospecting and simulation of the various scenarios that could occur in the long process from the beginning of the promotion and sale of housing. Such knowledge would have the aim of reducing, as far as possible, the risks inherent in this type of activity. Within this general framework, this thesis aims to analyze, describe and delineate the housing market and the performance of real estate companies, in their various components, as well as the players in it. It develops a model for prediction of prices and other indicators, which can obtain further information. Likewise, he examines the role of production companies promoters. With the foregoing finally arose, designs and develops a model of integrative simulation and prediction of housing prices, and heuristic type of computerized, and multiuser operating in real time. This model is capable of providing sufficient information so that management can make the most appropriate decision for their own businesses, according to the different scenarios designed and rated before deciding promotion to be done. The model is conceived as a system to assist management by providing the opportunity to work simultaneously with different scenarios prior to finalizing the development of advocacy. No go on the planning and monitoring of a property development.

Author: TABLA SEVILLANO RAFAEL.
Year: 2005.
University: EXTREMADURA [www.unex.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: FACULTAD DE CIENCIAS.

Summary: In this work we have isolated strains of lactic acid bacteria indigenous i from milk and cheese ibores. These strains were genetically and biochemically characterized. Elaborándose ferments with the most suitable from the point of view cheese. With ferments well prepared were used for the production of cheese from raw milk in pilot plants and semi-industrial level, resulting in selecting the most appropriate ferment in the manufacture of cheese ibores addition, the strains of Lactococcus Lactis subsp. Lactis isolated from cheese ibores were characterized using different methods electorforéticos determined to turn various activities of technological interest (acidification, proteolysis, autolisis production bacteriocinasâ |).

Author: DUXANS BARROBES HELENCA.
Year: 2005.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: Aula Màster-Edifici A3-Campus Nord.
Place of preparation: D3 111 Nord.

Author: Ibáñez Rivas Juan Nicolás.
Year: 2005.
University: SEVILLA [www.us.es].
Place of defense: Escuela Superior de Ingenieros.
Place of preparation: Escuela Superior de Ingenieros.

Summary: The thesis abounds in the process of constructing questionnaires and subsequent treatment of preferences collected through them to the users of transport systems. This will jointly develop the next two disciplines, modeling of individual preferences and the design of experiments, providing adequate reproduction of actual behavior of individuals in their interaction with the transportation systems by those affected, and quantifying the characteristics most of these systems influence the population under study. In Chapter 2 presents the state of the art and practice with regard to studies of preferences, particularizándose in the case of decisions relating to the carriage of passengers and goods. It outlines the different approaches under which the transport is discussed in the literature, it is argued the need for introduction of modeling preferences, differs among the different contributions of this type linear model (regression) and nonlinear (discrete choice ), and finally presents the temporal evolution of the methodology more relevant. In chapter 3 outlines the similarities and differences between actors of different models of preferences, with special emphasis on introducing the statistical component of each model independently to its approach as optimization problem without restrictions. In chapter 4 presents models of linear preferences as a tool initial approach to the study of the decisions of the officers involved in the transport, using the optical nonlinear optimization unrestricted and optics of a geometric problem. Chapter 5 is devoted to non-linear models of preferences, exposing the theory associated with them and how reproduce the behavior of the agents involved in the transport of linear differently to the case. Once you develop the linear and nonlinear modeling, Chapter 6 presents a practical application that combines elements of both to address the case of availability of a limited number of observations in the study of the factors that determine demand for transportation freight. The practical application relates to the British market intercity freight. In Chapter 7 addresses models relaxing some of the assumptions on which lie the nonlinear models of preferences, including hierarchical models. Much of the chapter is devoted to the following three tasks: first, to relax the permissible range proposed in the literature for the values of a quotient of special interest in these hierarchical models, the so-called values of inclusion, secondly, the proposal new non-linear models in literature and who come to relax some restrictions associated with hierarchical models and thirdly, to the correct identification of the models examined processes to allow estimation of the parameters of the model were conducted in a consistent manner . Some of the results that are obtained are used in the construction of mobility patterns intercity referring to the Australian continent. In chapter 8 elaborates the principles generally accepted in the literature to define the theory of behavior of individuals (postulates of the theory RUM or maximize the use of random). In particular, it places special emphasis on the comparison of different implementations that exist in RUM, introducing conditions of equivalence including clarifying what for 8 mulación 693 of RUM is more general and how it translates into fewer restrictions to abide by the parameters of the models. In chapter 9 is the exhibition of the design process of questionnaires which are obtained by the preferences of individuals, which are processed using any of the models presented in the previous chapters. Results are presented which demonstrate the goodness of integrating the two phases of design and estimate together. Finally, using the concepts of building on previous questionnaires and estimating models preferences and on the interaction between the two takes place in this chapter a practical application to measure the acceptability that a sample of carriers Seville attached to a series measures proposed as urban logistics solutions to improve mobility in the city center. Qualification thesis: Excellent Cum Laude and Mention of the European Doctorate

Author: Resina Rodriguez David.
Year: 2005.
University: AUTÓNOMA DE BARCELONA [www.uab.es].
Place of defense: E.T.Superior d'Enginyeria.
Place of preparation: Universidad Autonoma de Barcelona.

Summary: During the course of this work, the expression of a Rhizopus oryze lipase (ROL) under the control of the FLD1 promoter in the methylotrophic yeast Pichia pastoris, as well as the development of methanol-free cultivation processes and metabolic engineering strategies for process improvement has been carried out. First, an expression vector containing the ROL gene and the FLD1 promoter from P. pastoris was constructed and subsequently transformed in P. pastoris. Subsequent studies in shake flask and bioreactor batch cultivations revealed the synergetic effect of methanol and methylamine on the FLD1 promoter induction. Moreover, the use of sorbitol as carbon source, combined with methylamine as nitrogen source, was found to be an attractive alternative to the use of methanol for high cell density cultivations. Second, a new fed-batch fermentation strategy was developed using sorbitol and methylamine as carbon and nitrogen source, respectively. Three fed-batches were performed at different specific growth rates, 0.005 h-1, 0.01 h-1, and 0.02 h-1. The specific growth rate was controlled during induction phase in the 0.005 h-1, 0.01 h-1 fermentations by implementing an exponential feeding rate at limiting substrate concentration. In the near-to-maximum specific growth rate fermentation performed at about 0.02 h-1, sorbitol concentration was keep into the medium at about 8 g·L-1 throughout the induction phase. The highest lipase productivity was obtained in the fed-batch performed at the 0.02 h-1 specific growth rate. However, product production rate of all three fermentations dropped soon after reaching a maximum at about 30 hours of culture. Third, in order to analyze potential bottlenecks in the protein folding or secretion pathway and, in particular, the potential activation of the unfolded protein response (UPR) during ROL overexpression and intracellular accumulation, the levels of a reticle endoplasmatic-resident chaperone (BiP), its induction has been related to the activation of the UPR, as well as the intracellular levels of ROL were analyzed by immunofluorescence staining combined with flow cytometry. The increase of BiP and ROL levels during the induction phase of fed-batch cultivations indicated the possible activation of the UPR. Fourth, based on these analytical studies, a metabolic engineering strategy was proposed to alleviate the supposed bottleneck in the folding and secretion pathway. The Saccharomyces cerevisiae HAC1 gene was cloned and constitutively expressed in P. pastoris co-expressing ROL. The HAC1 gene codifies for a transcriptional factor which activates the UPR related genes expression. Thus, the constitutively activation of the UPR was aimed at sustaining a better protein folding and secretion efficiency. Results from fed-batch cultivations with the HAC1-engineered strain showed an increase in the heterologous product expression, demonstrating the positive effect of such modification on protein folding and secretion. Besides, a transcriptional analysis of some bioprocess-relevant genes of the ROL-expressing P. pastoris strains was carried out. In particular, the alcohol oxidase 1 (AOX1), formaldehyde dehydrogenase (FLD1), protein disulfide isomerase (PDI), binding protein (KAR2), 26S rRNA and the recombinant product (ROL) mRNA levels were monitored in both shake flask and fed-batch cultivations. The mRNA levels of these genes were quantified by using a sandwich hybridization assay based on immobilization of the desired mRNA on magnetic beads and subsequent detection by a fluorescence-based technique. The specific growth rate showed to be a key factor in the transcriptional level of the analyzed genes; moreover, the results gave additional evidence to the hypothesis that ROL triggers the UPR. Fifth, a second genetic modification was carried out in order to improve the secretion efficiency of ROL. The P. pastoris GAS1 gene, which codifies for a glycoprotein anchored to the outer layer of the plasma membrane performing cross links between 8 the cel a5b l wall -glucans, was disrupted. The deletion of this gene provoked a change in the cell wall conformation, causing an increase of the cell wall porosity and permeability, i.e. allowing for a better release of the recombine product. Although the mutant strain experienced a reduction of the maximum specific growth rate, the specific productivity, compared to the control strain, was doubled. Finally, in order to gain further insights on the potential (subcellular) accumulation of the recombinant product, the ROL gene was fused with the Aequorea victoria green fluorescence protein (GFP) gene and expressed in P. pastoris. The GFP was fused in N-terminus and C-terminus with ROL. However only the GFP-ROL fusion transformants showed detectable lipase activity. A batch culture was performed with the GFP-ROL fusion expressing strain, and compared with the ROL expressing control strain. The expression of the ROL as a fusion protein caused an important decrease in the extracellular levels of the recombinant lipase activity when compared with the results obtained with the control strain. Moreover, the quantification of the extracellular GFP fluorescence levels by direct fluorimetry was not possible due to the presence in the culture broth of a fluorescent by-product produced by P. pastoris during cultivation. The removal of this product (probably riboflavin) from culture supernatants allowed the detection of GFP fluorescent signal. On the other hand, the intracellular accumulation of the fusion product was followed by confocal microscopy and flow cytometry. The GFP-fusion protein appeared to accumulate in the cell wall periplasm as well as intracellularly into large, vacuole-likely, compartments. Overall, this study has allowed the development and improvement of the P. pastoris expression system based on the formaldehyde deshydrogenase promoter as an alternative to the classic methanol inducible alcohol oxidase promoter.

Author: ARAQUE CUENCA FRANCISCO.
Year: 2005.
University: GRANADA [www.ugr.es].
Place of defense: E.T.S. INGENIERÍA INFORMÁTICA.
Place of preparation: ETS INGENIERÍA INFORMÁTICA.

Summary: A data warehouse contains a copy of data from operational systems, and provides a historical overview of the data structure with a particularly appropriate for consultations and analysis. The diagram of a data warehouse should be directed to the storage itself, without taking into account possible consultations that will ever be made on him. The consultations will be carried out mainly on multidimensional databases designed for this purpose and whose cough is obtained from the warehouse. At this point, it is very important to take into account the temporal dimension, which is vital in any task analysis. The outline of the data warehouse must be able to reflect the temporal characteristics of the data. In the same way, it will be equally important to study the mechanisms of former draft operational systems components for this type of data. Moreover, it is also necessary to move the cough gives operational systems to the model's data warehouse, as well as test whether the data from the data sources have changed since the last access to, if necessary, update the data warehouse . To make this integration process is suitable to use a model object-oriented (OO) as a canonical data model. The standard ODMG defines a set of specifications that establishes how to identify and consult databases object oriented. Currently, that standard has not been recognized as such by any committee of standardization, although today is the de facto standard. However, this standard does not include any feature that allows temporary defining elements. The inclusion of a feature of this would represent elements and temporal properties in the schedules of both sources of data as one's own data warehouse. The databases of real-time and temporary databases using concepts can be applied to the task of enriching the OOP model to facilitate the process of defining the model of the warehouse and updating the same. Take into account the temporal characteristics of data sources, may help when the design schema data warehouse, as the designer may have information not only on the data that different sources can offer, but also for their temporary features, which are both components schema data warehouse. The integration of data sources for incorporation into the data warehouse is a field that has been studied from the point of view of solving various problems, including: Integration of the semantic properties of the data, integrating security levels, and so on. However, so far have not taken into account the temporal properties of the data. These properties can be inferred from the very source of the data is, as the extraction method used to obtain it. The integration of the temporal properties of the data implies that we have algorithms to solve the problem if possible, taking into account the temporal properties of the data, the integration of data from different sources to be incorporated into the warehouse. This thesis explores the problem of defining the model and pattern of the data warehouse in terms of the temporary operational systems components. Based on the object model ODMG as canonical data model, an enrichment to incorporate elements of the same temporal and extend metadata to contemplate the elements added to the model. It also proposes algorithms for the integration of the temporal properties of the data before being incorporated into the data warehouse and the requirements of the manager's own warehouse.

Author: MARTÍNEZ ÁLVAREZ ANTONIO.
Year: 2005.
University: GRANADA [www.ugr.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA INFORMÁTICA.

Summary: This dissertation addresses the design hardware / software processing systems vision. The work has been motivated by the European Project CORTIVIS (QLK6-CT-2001-00279), whose main purpose is to design a neuroprótesis cortical capable of producing automatically flows nerve in the visual cortex of blind people. The solution that brings conducts a design descending (top-down) for a model of retina, taking into account different levels of abstraction. The result is a system designed to the specification, simulation, verification and synthesis of computational models of vision that makes special emphasis on a particular optimization strategy. The main tasks carried out in the fall doctoral work in the following items: * Definition of a model of retina. * Designing a functional simulator modeling vision bioinspirados with the possibility of the results biological and synthetic. Developing a system capable of carrying out the automatic generation of systems bioinspirados vision (such as a retina) in hardware, from high-level specifications of the system given by the previous simulator. This specification of the model is transported to reconfigurable hardware considering certain processes optimization. The platform simulation and synthesis generates optimal digital circuit descriptions, using the standard IEEE VHDL language, which have been evaluated in a number of successful technologies and proven FPGA synthesis tools with different logic.

Author: SANCHEZ CALVO RAÚL.
Year: 2005.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: ETSI CAMINOS, CANALES Y PUERTOS.
Place of preparation: ETSI CAMINOS, CANALES Y PUERTOS.

Summary: The keys hydraulic machines are used in the distribution network hydraulic pressure as elements of automatic control and maneuver. Among its automatic functions more frequent irrigation networks are regulatory pressure and restriction of flow. In addition, these keys are carried out operations of opening and closing. Previous work can be found on various ways to model the operation of which have ideal role for regulating pressure to be introduced into the study of hydraulic problems associated with the network. However, there has been no regulation on the real pressure, nor on the influence of various elements attached to the key, despite the fact that the regulation dela pressure is related to the results of irrigation. Thus, the thesis deals with the prediction of its actual operation, both permanent regime as a variable. Based on the analysis has been conducted modulation mathematics of the physical phenomena occurring in the different elements that compose it. Through experiments in the laboratory have been contrasting the methods and assumptions, and are calibrated parameters of the models developed. With the latter, has predicted the actual operation mentioned. The proposed parameters can be taken to characterize those keys in your technical information. The regulatory functions of the keys is based on the ability to modify losses burden by increasing or decreasing its section of way. In the vicinity of this section, as a result of increased speed, there is a drop in pressure, so that if the latter large enough, cavitation occurs. For this reason, the thesis, in addition to the loss of load depending on the position of the operculum of the key has been studied relate these to the cavitation, depending on the pressures imposed by the terms of the limit set in you work that key. This relationship allows eliminate uncertainty in drop tests, whether cavitation has influenced such losses, or not, in addition to allowing his prediction. By the models mentioned, he examines the impact on the functioning of these keys, the adjustment of the different elements that compose it. These tests are a useful tool both for design and for his selection. Also, for purposes of the project, considers the operating point for which a key equipped with a pilot hydraulic pressure regulator presents the maximum aperture, and also provides the pressure under the onslaught of the unit irrigation. This will reduce the losses of cargo, which carry positive implications from the economic point of view. For an automatic key regulatory pressure, installed in the field, to provide the proposed operation, establishing an operating procedure, based on the extent of pressures on the spot, to adjust the screw compressor spring's pilot regulatory depression of the key. With regard to the application of this procedure may even be ignored the technical characteristics of the key and the pilot. As regards the regime variable, it predicts the response of the regulatory keys to the same pressure. It was also modeled, in a relatively simplified, the maneuver of closing a key hydraulics, providing a closing act that can be useful in trouble regime variable distribution networks.

Author: MARTINEZ PUENTES MONTSERRAT.
Year: 2005.
University: GIRONA [www.udg.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: FACULTAD DE CIENCIAS.

Author: ORGHIDAN RADU.
Year: 2005.
University: GIRONA [www.udg.es].
Place of defense: Universidad de Girona.
Place of preparation: Universidad de Girona.

Summary: The perception vision is improved when you can use a wide field of vision. This thesis focuses on the visual perception of depth with the help of cameras omnidireccionales. The perception 3D is usually obtained by the vision computer using stereo configurations with the disadvantage of high cost computer to find common elements between the visual images. The solution offered by the use light thesis is structured to address the problem of relating the correspondencias.Se has conducted a study on the omnidirectional vision systems. We have evaluated several configurations stereo and has been chosen better. The model parameters are difficult to measure directly, and therefore have developed a series of methods calibración.Los results are promising and show that the sensor can be used in applications for the perception of depth and serious: a model the scene, inspecting pipes, and so on.

Author: GUZMÁN SÁNCHEZ JOSÉ LUIS.
Year: 2005.
University: ALMERÍA [www.ual.es].
Place of defense: DEPARTAMENTO DE LENGUAJES Y COMPUTACIÓN.
Place of preparation: UNIVERSIDAD DE ALMERÍA.

Summary: The initial approach of the doctoral thesis was primarily aimed to develop methods and tools to support teaching in Automatic Control, focused on process control techniques widely used in industry, as drivers Proporcional-Integral-Derivativo ( PID) and the Generalized Predictive Control (Generalized Predictive Control GPC). The methods and tools designed to make use of recent advances in the New Information Technologies and Communications NTIC), and intended to respond to needs identified in the field of various committees and forums. In this objective teaching underlies the need to design computationally efficient algorithms, to support one of the key features that require the tools teachers, Interactivity, which is treated in detail in this thesis. Once implemented tools related to PID controllers and GPC, including treatment of restrictions in both cases and to be one of the major contributions of the thesis, interest is focused on exploiting the power of the tool in the analysis and treatment of uncertainty, especially in the schemes based control GPC. Since the initial studies appreciated the difficulty inherent in reconciling the efficiency required by the interactive tools with the great burden that characterizes computational algorithms robust predictive control, and the investigation led to the design and implementation techniques sub-óptimas that represent a compromise between Optimal and efficient computational tools required by the teachers. In this area comes the second major contribution of the thesis, which consists of a robust predictive control scheme which seeks to combine the advantages of the ealimentación Quantitative Theory (Quantitative Feedback Theory, QFT) in terms of efficiency and in mild conservatism the treatment of uncertainty, and control GPC in the treatment field restrictions and predictability. The thesis presents the control scheme and methodologies for the design and analysis of stability in both cases without restrictions as subjects. In order to demonstrate the efficiency of some of the computational algorithms proposed, has developed an interactive tool that will be described in the same way along the thesis. In the same way, a solution has been obtained using the proposed scheme and applying Linear Matrix Inequalities (LMI) in order to obtain solutions robustly stable in the presence of constraints. Of course, another added advantage of the techniques developed is that, being computationally efficient, are susceptible of being used in industrial applications, an element that seeks to exploit natural as implementation of the outcome of the thesis.

Author: GARCÍA MORALES LINO PEDRO.
Year: 2005.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: E.T.S.I.T..
Place of preparation: E.T.S.I.T..

Summary: The basic idea of the cancellation of echoes is blocking the signal from the remote room that is not realimente. This places an adaptive system in the middle that generates âidealmenteâ the same signal that comes from the local chamber and sent back to the courtroom remote difference between the signal and the estimated returns of the local chamber, rather than the latter . The introduction of multiple channels for capturing the spatiality of the speakers but introduces a high correlation between signals that affects the behavior of cancelador adaptive. This disadvantage of poor conditioning system, known as the problem of no-unicidad, can make even the adaptive system converges to a solution is not unique. In canceling echoes multichannel these solutions depend not only on responses to the momentum of the local chamber, but also the responses to the momentum of the room remotely. Responses to the momentum of the Boards typical hands-free applications, which require the cancellation of echoes, are extremely large (in the order of thousands of coefficients). The enormous length of the replies to boost not only results in a large volume of mathematical operations, but also in a delay unacceptable perceptualmente. In the first part of this thesis explores the problem of multichannel acoustic echo cancellation. Here are explored and comparing different multichannel adaptive filtering techniques in the search for suitability for the application of the cancellation of echoes. This provides for eight experiments: the first and fourth, a single channel (simplest case), the second, fifth and seventh, two-channel (stereo case: the simplest case multichannel overall), third, sixth and eighth, five channels (multichannel typical case in domestic applications and videoconferencing). The experiments were elected in a manner constituting critical cases with a very high correlation inter-canal to estimate the behavior of algorithms in a very critical situation. These adaptive filtering techniques can not be applied directly to a system of multi-channel acoustic echo cancellation adaptive by the delay and the computational load imposed lengthy responses to boost acoustic involved. Therefore, in the second part of the thesis discusses adaptive filtering architectures multirresolución to address the problem in the time domain and frequency: decomposition into subbands and adaptive filtering in the frequency domain partitioning blocks. Finally there is a study for the decorrelación inter-canal looking for a better condition of the problem: the decorrelación by transforming adaptive Karhunen-Loève and the introduction of broad-band noise decorrelado. The latter technique allows to tackle the problem of canceling echoes multichannel detection without double locution. Finally, we sought a subjective assessment of resultados.En appendices addresses two issues very important for the development of this thesis. The first deals with the measurement and simulation rooms. Both techniques are very important for the cancellation of echoes multichannel because life provide answers to the momentum in different conditions, correlation, etc.. Based on those simulations of algorithms developed and analyzed in the thesis. The second deals with the conjugate gradient techniques that although an optimization algorithm for minimizing features, and its importance in this research deserves detailed study. Using techniques conjugate gradient in the multichannel acoustic echo cancellation is one of the fundamental contributions of this investigation and this causes different adaptive algorithms.

Author: MARTÍNEZ SANTA-MARÍA CAROLINA.
Year: 2005.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: E.T.S. DE INGENIEROS DE MONTES.
Place of preparation: E.T.S. DE INGENIEROS DE MONTES.

Summary: THIS THESIS PROPOSES TWO NEW APPROACHES TO: 1 .- ALTERATION TO EVALUATE THE TERMS HIDROLÓGICOS OCCURS IN AS A RESULT OF THE PRESENCE OF AN PRESA: AHI METHODOLOGY SO-CALLED OR RATES TO PROPOSE ALTERATION HIDROLÓGICA 2 .- PERTAINING TO ENVIRONMENTAL CAUDALES, UNDERSTANDING FOR AS, AQUEL SYSTEM THAT CIRCULANDO WATER TO THE LOW IS PRESA CAPAZ ENSURE THE FUNCTIONALITY OF ECOSYSTEM FLUVIAL CONCERNED IN SOME ACCEPTABLE LEVELS: METHODOLOGY RAC. 3 .- PROPOSED BY LAST ONE SET OF PROTOCOLS AND INDICATORS TO INTEGRATE THE RESULTS OBTAINED AT THE DECISION-MAKING.

Author: MENENDEZ PIDAL DE NAVASCUES IGNACIO.
Year: 2005.
University: POLITÉCNICA DE MADRID [www.upm.es].
Place of defense: ETSI DE CAMINOS, CANALES Y PUERTOS.
Place of preparation: ETSI INGENIEROS DE CAMINOS, CANALES Y PUERTOS.

Summary: The materials that make up the Facies Utrillas, called defined and in many ways, routine nature of sandy, often white, accepted Lower Cretaceous age usually Albense- without any apparent hardness and complication, with a specific intrinsic characteristics, and to be referring who was confronted by them or overcome them are qualified deservedly with adjectives such as' fearsome and avoidable as far as possible ', with a particular entity itself, and are themselves a source of sufficient study independently. Accordingly, this study covers a doctoral dual purpose. On the one hand, provide a possible answer and explanation to the continuing difficulties and problems that emerged in implementing the various civil engineering works, basically underground water facilities and other types of linear as rail, conducted between 1935-1981 mainly in the east of the peninsula, is prone to outcrop Facies Utrillas .. Moreover, as a second aim, is analyzed in terms of geological and geotechnical the adverse rocky material that was intercepted in the construction of infrastructure identified

Author: Pegalajar Cuéllar Manuel.
Year: 2005.
University: GRANADA [www.ugr.es].
Place of defense: E.T.S. Ingeni. Inform. y de Teleco..
Place of preparation: E.T.S. Ingeniería Informática y de Telecomunicación.

Summary: The objective of this thesis is the design of intelligent systems for modeling and forecasting data sets to provide some temporary units. Our proposal will focus on the use of models bioinspirados (neural networks, algorithms bioinspirados, etc..), Which adapt to carry out our goal. The purpose of the analysis of data sets is the knowledge of his pattern of behavior so that they can predict the evolution of the same, under the assumption that the conditions affecting the series does not vary with respect to past and current data , over time. However, if knowing past and present values of the series could be predict future values without error, we would be facing a deterministic process, and his study would have no interest. The fact that motivates the study of the time series is the emergence of phenomena in nature random, partial or total, which makes the prediction of future values of the series can only be accomplished through probabilistic methods. Most of the tools used so far for analysis and forecasting data sets are based on statistical linear regression techniques (Box-Jenkins method), methods for smoothing, and some non-linear methods, mainly regressions models Markov, neural networks, and so on. Much of the current literature uses neural networks to forecast data series, as networks RBF and Perceptrón Multicapa some of the most commonly used. The advantages brought about by the use of neural models for predicting time series are numerous: robustness against noise, adaptability, condition aproximadores universal, non-linearity, and so on. However, there are certain drawbacks in terms of a time series model using neural networks feedforward: Finding the proper structure of entry for the network is a complex task, because of the ignorance of existing relationships in the data set. This task requires multiple experiments and tests until the appropriate structure for the network. Moreover, if the time series presents a dynamic behavior, ie have different time-dependent behavior, the use of neural networks feedforward becomes ineffective, so it should be a model for each type of network behavior than it is filed. In our work will address the problem of modeling and prediction data sets using neural networks recurring dynamics (RNRD onwards). The RNRD are models of artificial neural networks, which incorporates some feedback on its structure. This characteristic makes them a suitable tool for data processing presenting some structure over time. However, the training of this type of model is a complex task because of the difficulty introduced by the treatment of relapse. The classical training algorithms for such networks, primarily based on gradient, have certain limitations because they have a tendency to get caught in local optimum, thereby hampering the correct learning of the neural network. Our work will focus in part on the development of techniques to help solve these problems, addressing from various perspectives: On the one hand, evolutionary algorithms are search algorithms, optimization and learning, which have been applied to a large number of problems getting adequate results. Specifically, in the field of training of neural networks, the use of genetic algorithms has produced better results than those obtained by conventional training algorithms based on the gradient. The adaptation and study div 8 ersas tea a8f cnicas evolutionary applied to the training of RNRD in trouble modeling and forecasting data series, will form the basis for development of the thesis. Starting from this point, we will develop new techniques that can compete with evolutionary algorithms existing training, so as to improve the learning of RNRD. Secondly, some classical techniques that enhance the power of the algorithms based on the gradient, such as certain methods Quasi-Newton and other non-linear programming techniques, have proven to be effective tools to solve optimization problems. While the evolutionary algorithms do a heuristic search along the solution space, trying to find optimal solutions, algorithms Quasi-Newton directed the search based on information provided by the gradient. The main drawback of evolutionary algorithms is the high runtime need to find appropriate solutions. Therefore, the use of classical techniques of non-linear optimization, greater power that the algorithms based on the gradient can be adapted so that the training of RNRD is conducted in an appropriate manner. Additionally, complementing the effectiveness of such techniques with evolutionary algorithms, we intend to obtain new hybrid techniques to improve current results in terms of learning neural networks. Finally, another key aspect in the design of neural networks, is to find the structure of the same to enable a proper apprenticeship, without losing the ability to spread the network by sobreentrenamiento. The development of this phase requires a large number of experiments and tests, being primarily addressed at the present time by trial and error methods. In our work, the development of optimization techniques multiobjetivo will allow us to train a neural network so as to achieve the optimal structure of the same.

Author: PÉREZ PÉREZ RAMÓN.
Year: 2005.
University: GRANADA [www.ugr.es].
Place of defense: E.T.S. INGE. INFORMÁTICA Y DE TELECO..
Place of preparation: E.T.S. INGENIERÍA INFORMÁTICA Y DE TELECOMUNICACIÓN.

Summary: The argument made by D. Ramon Perez-Perez presents an important contribution to the scientific community. Specifically, in the field of bioinformatics brings a remarkable breakthrough in the management of collections. Given the high potential discovered from the application of advanced computational techniques in the study of biodiversity, it has surpassed the stage at which the study of the problems could be done with software that targets without further consultation or presencia statistical study elementary. At present, due to the environment and the preservation of biodiversity are global problems and as such must be addressed, the necessary software is necessarily complex, the amount of data to be processed for answers and solutions to problems of enormous magnitude space. The contribution made in the thesis can provide new tools for studies of biodiversity, using Systems Multiagentes, incorporating significant improvements over existing methods. In addition, the framework of a research area in which the thesis has been developed is very promising, considering the potential for expansion of the research done in future work.

Author: GALINDO ANDRADES CIPRIANO.
Year: 2005.
University: MÁLAGA [www.uma.es].
Place of defense: UNIVERSIDAD DE MÁLAGA.
Place of preparation: UNIVERSIDAD DE MÁLAGA.

Summary: The work carried out in this thesis focuses on the study and application of structures multijerárquicas, representing around one mobile robot, with the aim of improving its ability to perform complex tasks in human scenarios. A mobile robot must hold a symbolic representation of their environment in order to carry out operations deliberative, such planning tasks. But in the real environments represented symbolically, given its complexity, it is essential to have mechanisms capable of organizing and facilitating access to the enormous amount of information that is derived from them. Besides the difficulty of dealing with large amounts of information, there are other problems underlying the symbolic representation of actual environments, which have not yet been fully resolved in the scientific literature. One is the maintenance of symbolic representation optimized with respect to the tasks that the robot must perform consistent with the environment in which it operates. Another problem, related concern is the creation / modification of the symbolic information from purely sensory information (this problem is known as symbol-grounding). This thesis examines these problems and providing solutions through structures multijerárquicas. These symbolic structures, based on the notion of abstraction, imitate the way in which humans organize space and enables a mobile robot to improve their skills in complex environments. The main contributions of this work are: * It has been mathematically formalized a model based on multiple symbolic abstractions (multijerarquías) by Theory categories. * It has developed an efficient scheduler task that is capable of utilizing the hierarchical organization of the model symbolic entonor. Our method has been validated mathematically and have been implemented and purchased two variants thereof. * A particular instance of the model multijerárquico been studied and implemented in organizing symbolic information with the aim of improving different tasks to be performed simultaneously by a mobile robot. * It has developed a procedure: 1-Build a hierarchical model of a robot's environment. 2-What remains consistent and updated. 3, - The optimized to enhance the tasks performed by a mobile robot. * Finally, we have implemented an architecture robotics that encompasses all the issues mentioned above. Actual tests have been conducted with a robotic wheelchair that demonstrate the utility of using structures multijerárquicas.

Author: MARTÍN EZQUERRA ANA BELÉN.
Year: 2005.
University: COMPLUTENSE DE MADRID [www.ucm.es].
Place of defense: FACULTAD DE CIENCIAS QUÍMICAS.
Place of preparation: FACULTAD DE CIENCIAS QUÍMICAS.

Summary: The purpose of this work is to study the biodesulfuración of dibenzotiofeno (DBT) using Pseudomonas putida CEC5279 as biocatalizador process. The DBT has been chosen as a representative model composed of sulfur compounds remaining after some desulfurization, hidrodesulfuración (HDS), oil fractions, as gas-óleo. The study was conducted in aqueous phase with DBT dissolved in the middle, regardless of the presence of an organic phase, a problem that is addressed in other parallel work. The chosen form of operation is known as resting cells because it is living cells in a latent state in which they have been detained growth in a particular item, by eliminating the nutrieres of culture medium, so the process includes two stages: growth, during which it was obtained bacterial cells capable of desulfurizing and desulfurization. Before embarking on any part of the study, it was necessary to establish a working protocol standard, which ensures the repeatability of the results, to be able, in this way, analyze and compare the results obtained in different experiments. To this end a series of preliminary studies primarily focused on obtaining the inoculum and in the maintenance of cell viability, to establish specific experimental procedures parala conservation of the microorganism, and for the study of the stages of growth and biodesulfuración DBT. Likewise, it is necessary to the development of analytical methods that provide information about the evolution of the system for measuring bacterial growth, the consumption of substrate (carbon source, a source of nitrogen and dissolved oxygen), as well as the ability desuflurante developed it during the same and then apply at the stage of BDS. As for the production of biocatalizador, the study will focus on the factors promoting growth and development of the capacity desulfurizing of P.putida. Specifically, we examine the composition of the growth medium BSM, and the conditions of operation. On the other hand, at the stage of Biodesulfuración DBT will study the operating conditions (temperature and dissolved oxygen). In addition, it intends to conduct a study of the most basic enzymatic activities of enzymes of the route 4S, using cells obtained under optimal conditions for the development of desulfurizing capacity, thus beginning the study of certain phenomena such as transport of intermediate delos dela route through cell membrane in a macroscopic, and the influence of the intracellular concentration of certain compounds, such as reduced equivalents (NADH and FMNH2) and the enzymatic activities of the enzymes DsxAm, DszB, DszC and HpaC. Once known the conditions for carrying out the process of BDS, will be the kinetic description of the various phases and events occurring during the same: growth, capacity building desulfurizing and consumption and transport of dissolved oxygen. In principle, will be formulated mathematically simple models to describe the evolution of the concentration of biomass, the capacity desulfurizing, of the concentration of dissolved oxygen and different genetic compounds dela route 4S checking the capacity of these descriptive models.

Author: MENDES BARBOSA ÁLVARO MANUEL.
Year: 2005.
University: POMPEU FABRA [www.upf.edu].
Place of defense: DEPARTAMENTO DE TECNOLOGÍA.
Place of preparation: UNIVERSIDAD POMPEU FABRA.

Summary: This thesis is the investigation into practices musical based on computer networks in the Music Technology Group at the Universitat Pompeu Fabra in Barcelona between the year 2001 and 2005. Part of the work is based on developments of the last decade in the field of Cooperative Working with Computers (Computer-Supported Cooperative Work, CSCW), which provides collaborative mechanisms which, from a musical point of view, they can be studied in various scenarios: composition, interpretation, improvisation and education. The first contribution of this paper is a comprehensive analysis and a systematic classification of Cooperative Working with ordinations for Applications Musical. This analysis will focus on identifying innovative proposals, models and applications, with a special emphasis on the shared nature of communication via the Internet. The concept of Shared Environment Sonoro is presented and implemented in a prototype application nominated Public Sound Objects (PSOs). The second largest contribution of the thesis is to study possible methods to reduce interruptions due to the delays inherent in the communication networks between music very displaced in the geographical area. From experimentation and evaluation laboratory techniques Adaptive Network Latency Tempo and Single Delayed Feed-Back are defined and implemented in the prototype PSOs. As long development project PSOs other problems have arisen, such as the design of interfaces depending on the behavior of the applications, overcoming the various security systems computer networks and the potential for scalability for a variety of audio applications the web. During the elaboration of this thesis is to discuss different perspectives to solve problems related to practice music for computer networks, using different viewpoints provinentes psychosocial study of the processes of musical collaboration in the world of computers and technology music.