Regularized model learning in EDAs for continuous and multi-objective optimization

Probabilistic modeling is the de���ning characteristic of estimation of distribution algorithms (EDAs) which determines their behavior and performance in optimization. Regularization is a well-known statistical technique used for obtaining an improved model by reducing the generalization error of estimation, especially in high-dimensional problems. `1-regularization is a type of this technique with the appealing variable selection property which results in sparse model estimations. In this thesis, we study the use of regularization techniques for model learning in EDAs. Several methods for regularized model estimation in continuous domains based on a Gaussian distribution assumption are presented, and analyzed from di���erent aspects when used for optimization in a high-dimensional setting, where the population size of EDA has a logarithmic scale with respect to the number of variables. The optimization results obtained for a number of continuous problems with an increasing number of variables show that the proposed EDA based on regularized model estimation performs a more robust optimization, and is able to achieve signi���cantly better results for larger dimensions than other Gaussian-based EDAs. We also propose a method for learning a marginally factorized Gaussian Markov random ���eld model using regularization techniques and a clustering algorithm. The experimental results show notable optimization performance on continuous additively decomposable problems when using this model estimation method. Our study also covers multi-objective optimization and we propose joint probabilistic modeling of variables and objectives in EDAs based on Bayesian networks, speci���cally models inspired from multi-dimensional Bayesian network classi���ers. It is shown that with this approach to modeling, two new types of relationships are encoded in the estimated models in addition to the variable relationships captured in other EDAs: objectivevariable and objective-objective relationships. An extensive experimental study shows the e���ectiveness of this approach for multi- and many-objective optimization. With the proposed joint variable-objective modeling, in addition to the Pareto set approximation, the algorithm is also able to obtain an estimation of the multi-objective problem structure. Finally, the study of multi-objective optimization based on joint probabilistic modeling is extended to noisy domains, where the noise in objective values is represented by intervals. A new version of the Pareto dominance relation for ordering the solutions in these problems, namely ���-degree Pareto dominance, is introduced and its properties are analyzed. We show that the ranking methods based on this dominance relation can result in competitive performance of EDAs with respect to the quality of the approximated Pareto sets. This dominance relation is then used together with a method for joint probabilistic modeling based on `1-regularization for multi-objective feature subset selection in classi���cation, where six di���erent measures of accuracy are considered as objectives with interval values. The individual assessment of the proposed joint probabilistic modeling and solution ranking methods on datasets with small-medium dimensionality, when using two di���erent Bayesian classi���ers, shows that comparable or better Pareto sets of feature subsets are approximated in comparison to standard methods.

Maximum Likelihood Estimation of Modal Parameters in Structures Using the Expectation Maximization Algorithm

This paper presents a time-domain stochastic system identification method based on maximum likelihood estimation (MLE) with the expectation maximization (EM) algorithm. The effectiveness of this structural identification method is evaluated through numerical simulation in the context of the ASCE benchmark problem on structural health monitoring. The benchmark structure is a four-story, two-bay by two-bay steel-frame scale model structure built in the Earthquake Engineering Research Laboratory at the University of British Columbia, Canada. This paper focuses on Phase I of the analytical benchmark studies. A MATLAB-based finite element analysis code obtained from the IASC-ASCE SHM Task Group web site is used to calculate the dynamic response of the prototype structure. A number of 100 simulations have been made using this MATLAB-based finite element analysis code in order to evaluate the proposed identification method. There are several techniques to realize system identification. In this work, stochastic subspace identification (SSI)method has been used for comparison. SSI identification method is a well known method and computes accurate estimates of the modal parameters. The principles of the SSI identification method has been introduced in the paper and next the proposed MLE with EM algorithm has been explained in detail. The advantages of the proposed structural identification method can be summarized as follows: (i) the method is based on maximum likelihood, that implies minimum variance estimates; (ii) EM is a computational simpler estimation procedure than other optimization algorithms; (iii) estimate more parameters than SSI, and these estimates are accurate. On the contrary, the main disadvantages of the method are: (i) EM algorithm is an iterative procedure and it consumes time until convergence is reached; and (ii) this method needs starting values for the parameters. Modal parameters (eigenfrequencies, damping ratios and mode shapes) of the benchmark structure have been estimated using both the SSI method and the proposed MLE + EM method. The numerical results show that the proposed method identifies eigenfrequencies, damping ratios and mode shapes reasonably well even in the presence of 10% measurement noises. These modal parameters are more accurate than the SSI estimated modal parameters.

An evolutionary algorithm for the surface structure problem

Many macroscopic properties: hardness, corrosion, catalytic activity, etc. are directly related to the surface structure, that is, to the position and chemical identity of the outermost atoms of the material. Current experimental techniques for its determination produce a ���signature��� from which the structure must be inferred by solving an inverse problem: a solution is proposed, its corresponding signature computed and then compared to the experiment. This is a challenging optimization problem where the search space and the number of local minima grows exponentially with the number of atoms, hence its solution cannot be achieved for arbitrarily large structures. Nowadays, it is solved by using a mixture of human knowledge and local search techniques: an expert proposes a solution that is refined using a local minimizer. If the outcome does not fit the experiment, a new solution must be proposed again. Solving a small surface can take from days to weeks of this trial and error method. Here we describe our ongoing work in its solution. We use an hybrid algorithm that mixes evolutionary techniques with trusted region methods and reuses knowledge gained during the execution to avoid repeated search of structures. Its parallelization produces good results even when not requiring the gathering of the full population, hence it can be used in loosely coupled environments such as grids. With this algorithm, the solution of test cases that previously took weeks of expert time can be automatically solved in a day or two of uniprocessor time.

Approximations on minimum weight pseudo-triangulation problem using ant colony optimization

In this work, we consider the Minimum Weight Pseudo-Triangulation (MWPT) problem of a given set of n points in the plane. Globally optimal pseudo-triangulations with respect to the weight, as optimization criteria, are difficult to be found by deterministic methods, since no polynomial algorithm is known. We show how the Ant Colony Optimization (ACO) metaheuristic can be used to find high quality pseudo-triangulations of minimum weight. We present the experimental and statistical study based on our own set of instances since no reference to benchmarks for these problems were found in the literature. Throughout the experimental evaluation, we appraise the ACO metaheuristic performance for MWPT problem.

Decentralized multi-tasks distribution in heterogeneous robot teams by means of ant colony optimization and learning automata

This paper focuses on the general problem of coordinating multiple robots. More specifically, it addresses the self-election of heterogeneous specialized tasks by autonomous robots. In this paper we focus on a specifically distributed or decentralized approach as we are particularly interested on decentralized solution where the robots themselves autonomously and in an individual manner, are responsible of selecting a particular task so that all the existing tasks are optimally distributed and executed. In this regard, we have established an experimental scenario to solve the corresponding multi-tasks distribution problem and we propose a solution using two different approaches by applying Ant Colony Optimization-based deterministic algorithms as well as Learning Automata-based probabilistic algorithms. We have evaluated the robustness of the algorithm, perturbing the number of pending loads to simulate the robot���s error in estimating the real number of pending tasks and also the dynamic generation of loads through time. The paper ends with a critical discussion of experimental results.

Auto-localization algorithm for local positioning systems

This paper studies the problem of determining the position of beacon nodes in Local Positioning Systems (LPSs), for which there are no inter-beacon distance measurements available and neither the mobile node nor any of the stationary nodes have positioning or odometry information. The common solution is implemented using a mobile node capable of measuring its distance to the stationary beacon nodes within a sensing radius. Many authors have implemented heuristic methods based on optimization algorithms to solve the problem. However, such methods require a good initial estimation of the node positions in order to find the correct solution. In this paper we present a new method to calculate the inter-beacon distances, and hence the beacons positions, based in the linearization of the trilateration equations into a closed-form solution which does not require any approximate initial estimation. The simulations and field evaluations show a good estimation of the beacon node positions.

Aerodynamic Optimization of the Nose Shape of a High-speed Train

La influencia de la aerodin��mica en el dise��o de los trenes de alta velocidad, unida a la necesidad de resolver nuevos problemas surgidos con el aumento de la velocidad de circulaci��n y la reducci��n de peso del veh��culo, hace evidente el inter��s de plantear un estudio de optimizaci��n que aborde tales puntos. En este contexto, se presenta en esta tesis la optimizaci��n aerodin��mica del testero de un tren de alta velocidad, llevada a cabo mediante el uso de m��todos de optimizaci��n avanzados. Entre estos m��todos, se ha elegido aqu�� a los algoritmos gen��ticos y al m��todo adjunto como las herramientas para llevar a cabo dicha optimizaci��n. La base conceptual, las caracter��sticas y la implementaci��n de los mismos se detalla a lo largo de la tesis, permitiendo entender los motivos de su elecci��n, y las consecuencias, en t��rminos de ventajas y desventajas que cada uno de ellos implican. El uso de los algorimos gen��ticos implica a su vez la necesidad de una parametrizaci��n geom��trica de los candidatos a ��ptimo y la generaci��n de un modelo aproximado que complementa al m��todo de optimizaci��n. Estos puntos se describen de modo particular en el primer bloque de la tesis, enfocada a la metodolog��a seguida en este estudio. El segundo bloque se centra en la aplicaci��n de los m��todos a fin de optimizar el comportamiento aerodin��mico del tren en distintos escenarios. Estos escenarios engloban los casos m��s comunes y tambi��n algunos de los m��s exigentes a los que hace frente un tren de alta velocidad: circulaci��n en campo abierto con viento frontal o viento lateral, y entrada en t��nel. Considerando el caso de viento frontal en campo abierto, los dos m��todos han sido aplicados, permitiendo una comparaci��n de las diferentes metodolog��as, as�� como el coste computacional asociado a cada uno, y la minimizaci��n de la resistencia aerodin��mica conseguida en esa optimizaci��n. La posibilidad de evitar parametrizar la geometr��a y, por tanto, reducir el coste computacional del proceso de optimizaci��n es la caracter��stica m��s significativa de los m��todos adjuntos, mientras que en el caso de los algoritmos gen��ticos se destaca la simplicidad y capacidad de encontrar un ��ptimo global en un espacio de dise��o multi-modal o de resolver problemas multi-objetivo. El caso de viento lateral en campo abierto considera nuevamente los dos m��toxi dos de optimizaci��n anteriores. La parametrizaci��n se ha simplificado en este estudio, lo que notablemente reduce el coste num��rico de todo el estudio de optimizaci��n, a la vez que a��n recoge las caracter��sticas geom��tricas m��s relevantes en un tren de alta velocidad. Este an��lisis ha permitido identificar y cuantificar la influencia de cada uno de los par��metros geom��tricos inclu��dos en la parametrizaci��n, y se ha observado que el dise��o de la arista superior a barlovento es fundamental, siendo su influencia mayor que la longitud del testero o que la secci��n frontal del mismo. Finalmente, se ha considerado un escenario m��s a fin de validar estos m��todos y su capacidad de encontrar un ��ptimo global. La entrada de un tren de alta velocidad en un t��nel es uno de los casos m��s exigentes para un tren por el pico de sobrepresi��n generado, el cual afecta a la confortabilidad del pasajero, as�� como a la estabilidad del veh��culo y al entorno pr��ximo a la salida del t��nel. Adem��s de este problema, otro objetivo a minimizar es la resistencia aerodin��mica, notablemente superior al caso de campo abierto. Este problema se resuelve usando algoritmos gen��ticos. Dicho m��todo permite obtener un frente de Pareto donde se incluyen el conjunto de ��ptimos que minimizan ambos objetivos. ABSTRACT Aerodynamic design of trains influences several aspects of high-speed trains performance in a very significant level. In this situation, considering also that new aerodynamic problems have arisen due to the increase of the cruise speed and lightness of the vehicle, it is evident the necessity of proposing an optimization study concerning the train aerodynamics. Thus, the aerodynamic optimization of the nose shape of a high-speed train is presented in this thesis. This optimization is based on advanced optimization methods. Among these methods, genetic algorithms and the adjoint method have been selected. A theoretical description of their bases, the characteristics and the implementation of each method is detailed in this thesis. This introduction permits understanding the causes of their selection, and the advantages and drawbacks of their application. The genetic algorithms requirethe geometrical parameterization of any optimal candidate and the generation of a metamodel or surrogate model that complete the optimization process. These points are addressed with a special attention in the first block of the thesis, focused on the methodology considered in this study. The second block is referred to the use of these methods with the purpose of optimizing the aerodynamic performance of a high-speed train in several scenarios. These scenarios englobe the most representative operating conditions of high-speed trains, and also some of the most exigent train aerodynamic problems: front wind and cross-wind situations in open air, and the entrance of a high-speed train in a tunnel. The genetic algorithms and the adjoint method have been applied in the minimization of the aerodynamic drag on the train with front wind in open air. The comparison of these methods allows to evaluate the methdology and computational cost of each one, as well as the resulting minimization of the aerodynamic drag. Simplicity and robustness, the straightforward realization of a multi-objective optimization, and the capability of searching a global optimum are the main attributes of genetic algorithm. However, the requirement of geometrically parameterize any optimal candidate is a significant drawback that is avoided with the use of the adjoint method. This independence of the number of design variables leads to a relevant reduction of the pre-processing and computational cost. Considering the cross-wind stability, both methods are used again for the minimization of the side force. In this case, a simplification of the geometric parameterization of the train nose is adopted, what dramatically reduces the computational cost of the optimization process. Nevertheless, some of the most important geometrical characteristics are still described with this simplified parameterization. This analysis identifies and quantifies the influence of each design variable on the side force on the train. It is observed that the A-pillar roundness is the most demanding design parameter, with a more important effect than the nose length or the train cross-section area. Finally, a third scenario is considered for the validation of these methods in the aerodynamic optimization of a high-speed train. The entrance of a train in a tunnel is one of the most exigent train aerodynamic problems. The aerodynamic consequences of high-speed trains running in a tunnel are basically resumed in two correlated phenomena, the generation of pressure waves and an increase in aerodynamic drag. This multi-objective optimization problem is solved with genetic algorithms. The result is a Pareto front where a set of optimal solutions that minimize both objectives.

A multi-criteria optimization and decision-making approach for improvement of food engineering processes.

The objective of this study was to propose a multi-criteria optimization and decision-making technique to solve food engineering problems. This technique was demostrated using experimental data obtained on osmotic dehydratation of carrot cubes in a sodium chloride solution. The Aggregating Functions Approach, the Adaptive Random Search Algorithm, and the Penalty Functions Approach were used in this study to compute the initial set of non-dominated or Pareto-optimal solutions. Multiple non-linear regression analysis was performed on a set of experimental data in order to obtain particular multi-objective functions (responses), namely water loss, solute gain, rehydration ratio, three different colour criteria of rehydrated product, and sensory evaluation (organoleptic quality). Two multi-criteria decision-making approaches, the Analytic Hierarchy Process (AHP) and the Tabular Method (TM), were used simultaneously to choose the best alternative among the set of non-dominated solutions. The multi-criteria optimization and decision-making technique proposed in this study can facilitate the assessment of criteria weights, giving rise to a fairer, more consistent, and adequate final compromised solution or food process. This technique can be useful to food scientists in research and education, as well as to engineers involved in the improvement of a variety of food engineering processes.

Global optimization of solar thermophotovoltaic systems

n this paper, we present a theoretical model based on the detailed balance theory of solar thermophotovoltaic systems comprising multijunction photovoltaic cells, a sunlight concentrator and spectrally selective surfaces. The full system has been defined by means of 2n���+���8 variables (being n the number of sub-cells of the multijunction cell). These variables are as follows: the sunlight concentration factor, the absorber cut-off energy, the emitter-to-absorber area ratio, the emitter cut-off energy, the band-gap energy(ies) and voltage(s) of the sub-cells, the reflectivity of the cells' back-side reflector, the emitter-to-cell and cell-to-cell view factors and the emitter-to-cell area ratio. We have used this model for carrying out a multi-variable system optimization by means of a multidimensional direct-search algorithm. This analysis allows to find the set of system variables whose combined effects results in the maximum overall system efficiency. From this analysis, we have seen that multijunction cells are excellent candidates to enhance the system efficiency and the electrical power density. Particularly, multijunction cells report great benefits for systems with a notable presence of optical losses, which are unavoidable in practical systems. Also, we have seen that the use of spectrally selective absorbers, rather than black-body absorbers, allows to achieve higher system efficiencies for both lower concentration and lower emitter-to-absorber area ratio. Finally, we have seen that sun-to-electricity conversion efficiencies above 30% and electrical power densities above 50���W/cm2 are achievable for this kind of systems.

Multi-objective Aerodynamic Optimization of High-speed Trains in Tunnels

A genetic algorithm (GA) is employed for the multi-objective shape optimization of the nose of a high-speed train. Aerodynamic problems observed at high speeds become still more relevant when traveling along a tunnel. The objective is to minimize both the aerodynamic drag and the amplitude of the pressure gradient of the compression wave when a train enters a tunnel. The main drawback of GA is the large number of evaluations need in the optimization process. Metamodels-based optimization is considered to overcome such problem. As a result, an explicit relationship between pressure gradient and geometrical parameters is obtained.

A framework for the analysis and optimization of encoding latency for multiview video

We present a novel framework for the analysis and optimization of encoding latency for multiview video. Firstly, we characterize the elements that have an influence in the encoding latency performance: (i) the multiview prediction structure and (ii) the hardware encoder model. Then, we provide algorithms to find the encoding latency of any arbitrary multiview prediction structure. The proposed framework relies on the directed acyclic graph encoder latency (DAGEL) model, which provides an abstraction of the processing capacity of the encoder by considering an unbounded number of processors. Using graph theoretic algorithms, the DAGEL model allows us to compute the encoding latency of a given prediction structure, and determine the contribution of the prediction dependencies to it. As an example of DAGEL application, we propose an algorithm to reduce the encoding latency of a given multiview prediction structure up to a target value. In our approach, a minimum number of frame dependencies are pruned, until the latency target value is achieved, thus minimizing the degradation of the rate-distortion performance due to the removal of the prediction dependencies. Finally, we analyze the latency performance of the DAGEL derived prediction structures in multiview encoders with limited processing capacity.

Multidisciplinary Design Optimization Application to Conceptual Design of University-class Microsatellite Projects

Esta tesis se ha realizado en el contexto del proyecto UPMSat-2, que es un microsat��lite dise��ado, construido y operado por el Instituto Universitario de Microgravedad "Ignacio Da Riva" (IDR / UPM) de la Universidad Polit��cnica de Madrid. Aplicaci��n de la metodolog��a Ingenier��a Concurrente (Concurrent Engineering: CE) en el marco de la aplicaci��n de dise��o multidisciplinar (Multidisciplinary Design Optimization: MDO) es uno de los principales objetivos del presente trabajo. En los ��ltimos a��os, ha habido un inter��s continuo en la participaci��n de los grupos de investigaci��n de las universidades en los estudios de la tecnolog��a espacial a trav��s de sus propios microsat��lites. La participaci��n en este tipo de proyectos tiene algunos desaf��os inherentes, tales como presupuestos y servicios limitados. Adem��s, debido al hecho de que el objetivo principal de estos proyectos es fundamentalmente educativo, por lo general hay incertidumbres en cuanto a su misi��n en ��rbita y cargas ��tiles en las primeras fases del proyecto. Por otro lado, existen limitaciones predeterminadas para sus presupuestos de masa, volumen y energ��a, debido al hecho de que la mayor��a de ellos est��n considerados como una carga ��til auxiliar para el lanzamiento. De este modo, el costo de lanzamiento se reduce considerablemente. En este contexto, el subsistema estructural del sat��lite es uno de los m��s afectados por las restricciones que impone el lanzador. Esto puede afectar a diferentes aspectos, incluyendo las dimensiones, la resistencia y los requisitos de frecuencia. En la primera parte de esta tesis, la atenci��n se centra en el desarrollo de una herramienta de dise��o del subsistema estructural que eval��a, no s��lo las propiedades de la estructura primaria como variables, sino tambi��n algunas variables de nivel de sistema del sat��lite, como la masa de la carga ��til y la masa y las dimensiones extremas de sat��lite. Este enfoque permite que el equipo de dise��o obtenga una mejor visi��n del dise��o en un espacio de dise��o extendido. La herramienta de dise��o estructural se basa en las f��rmulas y los supuestos apropiados, incluyendo los modelos est��ticos y din��micos del sat��lite. Un algoritmo gen��tico (Genetic Algorithm: GA) se aplica al espacio de dise��o para optimizaciones de objetivo ��nico y tambi��n multiobjetivo. El resultado de la optimizaci��n multiobjetivo es un Pareto-optimal basado en dos objetivo, la masa total de sat��lites m��nimo y el m��ximo presupuesto de masa de carga ��til. Por otro lado, la aplicaci��n de los microsat��lites en misiones espaciales es de inter��s por su menor coste y tiempo de desarrollo. La gran necesidad de las aplicaciones de teledetecci��n es un fuerte impulsor de su popularidad en este tipo de misiones espaciales. Las misiones de tele-observaci��n por sat��lite son esenciales para la investigaci��n de los recursos de la tierra y el medio ambiente. En estas misiones existen interrelaciones estrechas entre diferentes requisitos como la altitud orbital, tiempo de revisita, el ciclo de vida y la resoluci��n. Adem��s, todos estos requisitos puede afectar a toda las caracter��sticas de dise��o. Durante los ��ltimos a��os la aplicaci��n de CE en las misiones espaciales ha demostrado una gran ventaja para llegar al dise��o ��ptimo, teniendo en cuenta tanto el rendimiento y el costo del proyecto. Un ejemplo bien conocido de la aplicaci��n de CE es la CDF (Facilidad Dise��o Concurrente) de la ESA (Agencia Espacial Europea). Est�� claro que para los proyectos de microsat��lites universitarios tener o desarrollar una instalaci��n de este tipo parece estar m��s all�� de las capacidades del proyecto. Sin embargo, la pr��ctica de la CE a cualquier escala puede ser beneficiosa para los microsat��lites universitarios tambi��n. En la segunda parte de esta tesis, la atenci��n se centra en el desarrollo de una estructura de optimizaci��n de dise��o multidisciplinar (Multidisciplinary Design Optimization: MDO) aplicable a la fase de dise��o conceptual de microsat��lites de teledetecci��n. Este enfoque permite que el equipo de dise��o conozca la interacci��n entre las diferentes variables de dise��o. El esquema MDO presentado no s��lo incluye variables de nivel de sistema, tales como la masa total del sat��lite y la potencia total, sino tambi��n los requisitos de la misi��n como la resoluci��n y tiempo de revisita. El proceso de dise��o de microsat��lites se divide en tres disciplinas; a) dise��o de ��rbita, b) dise��o de carga ��til y c) dise��o de plataforma. En primer lugar, se calculan diferentes par��metros de misi��n para un rango pr��ctico de ��rbitas helio-s��ncronas (sun-synchronous orbits: SS-Os). Luego, seg��n los par��metros orbitales y los datos de un instrumento como referencia, se calcula la masa y la potencia de la carga ��til. El dise��o de la plataforma del sat��lite se estima a partir de los datos de la masa y potencia de los diferentes subsistemas utilizando relaciones emp��ricas de dise��o. El dise��o del subsistema de potencia se realiza teniendo en cuenta variables de dise��o m��s detalladas, como el escenario de la misi��n y diferentes tipos de c��lulas solares y bater��as. El escenario se selecciona, de modo de obtener una banda de cobertura sobre la superficie terrestre paralelo al Ecuador despu��s de cada intervalo de revisita. Con el objetivo de evaluar las interrelaciones entre las diferentes variables en el espacio de dise��o, todas las disciplinas de dise��o mencionados se combinan en un c��digo unificado. Por ��ltimo, una forma b��sica de MDO se ajusta a la herramienta de dise��o de sistema de sat��lite. La optimizaci��n del dise��o se realiza por medio de un GA con el ��nico objetivo de minimizar la masa total de microsat��lite. Seg��n los resultados obtenidos de la aplicaci��n del MDO, existen diferentes puntos de dise��os ��ptimos, pero con diferentes variables de misi��n. Este an��lisis demuestra la aplicabilidad de MDO para los estudios de ingenier��a de sistema en la fase de dise��o conceptual en este tipo de proyectos. La principal conclusi��n de esta tesis, es que el dise��o cl��sico de los sat��lites que por lo general comienza con la definici��n de la misi��n y la carga ��til no es necesariamente la mejor metodolog��a para todos los proyectos de sat��lites. Un microsat��lite universitario, es un ejemplo de este tipo de proyectos. Por eso, se han desarrollado un conjunto de herramientas de dise��o para encarar los estudios de la fase inicial de dise��o. Este conjunto de herramientas incluye diferentes disciplinas de dise��o centrados en el subsistema estructural y teniendo en cuenta una carga ��til desconocida a priori. Los resultados demuestran que la m��nima masa total del sat��lite y la m��xima masa disponible para una carga ��til desconocida a priori, son objetivos conflictivos. En este contexto para encontrar un Pareto-optimal se ha aplicado una optimizaci��n multiobjetivo. Seg��n los resultados se concluye que la selecci��n de la masa total por sat��lite en el rango de 40-60 kg puede considerarse como ��ptima para un proyecto de microsat��lites universitario con carga ��til desconocida a priori. Tambi��n la metodolog��a CE se ha aplicado al proceso de dise��o conceptual de microsat��lites de teledetecci��n. Los resultados de la aplicaci��n del CE proporcionan una clara comprensi��n de la interacci��n entre los requisitos de dise��o de sistemas de sat��lites, tales como la masa total del microsat��lite y la potencia y los requisitos de la misi��n como la resoluci��n y el tiempo de revisita. La aplicaci��n de MDO se hace con la minimizaci��n de la masa total de microsat��lite. Los resultados de la aplicaci��n de MDO aclaran la relaci��n clara entre los diferentes requisitos de dise��o del sistema y de misi��n, as�� como que permiten seleccionar las l��neas de base para el dise��o ��ptimo con el objetivo seleccionado en las primeras fase de dise��o. ABSTRACT This thesis is done in the context of UPMSat-2 project, which is a microsatellite under design and manufacturing at the Instituto Universitario de Microgravedad ���Ignacio Da Riva��� (IDR/UPM) of the Universidad Polit��cnica de Madrid. Application of Concurrent Engineering (CE) methodology in the framework of Multidisciplinary Design application (MDO) is one of the main objectives of the present work. In recent years, there has been continuing interest in the participation of university research groups in space technology studies by means of their own microsatellites. The involvement in such projects has some inherent challenges, such as limited budget and facilities. Also, due to the fact that the main objective of these projects is for educational purposes, usually there are uncertainties regarding their in orbit mission and scientific payloads at the early phases of the project. On the other hand, there are predetermined limitations for their mass and volume budgets owing to the fact that most of them are launched as an auxiliary payload in which the launch cost is reduced considerably. The satellite structure subsystem is the one which is most affected by the launcher constraints. This can affect different aspects, including dimensions, strength and frequency requirements. In the first part of this thesis, the main focus is on developing a structural design sizing tool containing not only the primary structures properties as variables but also the satellite system level variables such as payload mass budget and satellite total mass and dimensions. This approach enables the design team to obtain better insight into the design in an extended design envelope. The structural design sizing tool is based on the analytical structural design formulas and appropriate assumptions including both static and dynamic models of the satellite. A Genetic Algorithm (GA) is applied to the design space for both single and multiobejective optimizations. The result of the multiobjective optimization is a Pareto-optimal based on two objectives, minimum satellite total mass and maximum payload mass budget. On the other hand, the application of the microsatellites is of interest for their less cost and response time. The high need for the remote sensing applications is a strong driver of their popularity in space missions. The satellite remote sensing missions are essential for long term research around the condition of the earth resources and environment. In remote sensing missions there are tight interrelations between different requirements such as orbital altitude, revisit time, mission cycle life and spatial resolution. Also, all of these requirements can affect the whole design characteristics. During the last years application of the CE in the space missions has demonstrated a great advantage to reach the optimum design base lines considering both the performance and the cost of the project. A well-known example of CE application is ESA (European Space Agency) CDF (Concurrent Design Facility). It is clear that for the university-class microsatellite projects having or developing such a facility seems beyond the project capabilities. Nevertheless practicing CE at any scale can be beneficiary for the university-class microsatellite projects. In the second part of this thesis, the main focus is on developing a MDO framework applicable to the conceptual design phase of the remote sensing microsatellites. This approach enables the design team to evaluate the interaction between the different system design variables. The presented MDO framework contains not only the system level variables such as the satellite total mass and total power, but also the mission requirements like the spatial resolution and the revisit time. The microsatellite sizing process is divided into the three major design disciplines; a) orbit design, b) payload sizing and c) bus sizing. First, different mission parameters for a practical range of sun-synchronous orbits (SS-Os) are calculated. Then, according to the orbital parameters and a reference remote sensing instrument, mass and power of the payload are calculated. Satellite bus sizing is done based on mass and power calculation of the different subsystems using design estimation relationships. In the satellite bus sizing, the power subsystem design is realized by considering more detailed design variables including a mission scenario and different types of solar cells and batteries. The mission scenario is selected in order to obtain a coverage belt on the earth surface parallel to the earth equatorial after each revisit time. In order to evaluate the interrelations between the different variables inside the design space all the mentioned design disciplines are combined in a unified code. The integrated satellite system sizing tool developed in this section is considered as an application of the CE to the conceptual design of the remote sensing microsatellite projects. Finally, in order to apply the MDO methodology to the design problem, a basic MDO framework is adjusted to the developed satellite system design tool. Design optimization is done by means of a GA single objective algorithm with the objective function as minimizing the microsatellite total mass. According to the results of MDO application, there exist different optimum design points all with the minimum satellite total mass but with different mission variables. This output demonstrates the successful applicability of MDO approach for system engineering trade-off studies at the conceptual design phase of the design in such projects. The main conclusion of this thesis is that the classical design approach for the satellite design which usually starts with the mission and payload definition is not necessarily the best approach for all of the satellite projects. The university-class microsatellite is an example for such projects. Due to this fact an integrated satellite sizing tool including different design disciplines focusing on the structural subsystem and considering unknown payload is developed. According to the results the satellite total mass and available mass for the unknown payload are conflictive objectives. In order to find the Pareto-optimal a multiobjective GA optimization is conducted. Based on the optimization results it is concluded that selecting the satellite total mass in the range of 40-60 kg can be considered as an optimum approach for a university-class microsatellite project with unknown payload(s). Also, the CE methodology is applied to the remote sensing microsatellites conceptual design process. The results of CE application provide a clear understanding of the interaction between satellite system design requirements such as satellite total mass and power and the satellite mission variables such as revisit time and spatial resolution. The MDO application is done with the total mass minimization of a remote sensing satellite. The results from the MDO application clarify the unclear relationship between different system and mission design variables as well as the optimum design base lines according to the selected objective during the initial design phases.

Automatic design of Ant Colony Optimization for permutation problems

In this study, we present a framework based on ant colony optimization (ACO) for tackling combinatorial problems. ACO algorithms have been applied to many diferent problems, focusing on algorithmic variants that obtain high-quality solutions. Usually, the implementations are re-done for various problem even if they maintain the same details of the ACO algorithm. However, our goal is to generate a sustainable framework for applications on permutation problems. We concentrate on understanding the behavior of pheromone trails and specific methods that can be combined. Eventually, we will propose an automatic offline configuration tool to build an efective algorithm. ---RESUMEN---En este trabajo vamos a presentar un framework basado en la familia de algoritmos ant colony optimization (ACO), los cuales est��n dise~nados para enfrentarse a problemas combinacionales. Los algoritmos ACO han sido aplicados a diversos problemas, centr��ndose los investigadores en diversas variantes que obtienen buenas soluciones. Normalmente, las implementaciones se tienen que rehacer, inclusos si se mantienen los mismos detalles para los algoritmos ACO. Sin embargo, nuestro objetivo es generar un framework sostenible para aplicaciones sobre problemas de permutaciones. Nos centraremos en comprender el comportamiento de la sendas de feromonas y ciertos m��todos con los que pueden ser combinados. Finalmente, propondremos una herramienta para la configuraron autom��tica offline para construir algoritmos eficientes.

Fully adaptive gaussian mixture metropolis-hastings algorithm

Markov Chain Monte Carlo methods are widely used in signal processing and communications for statistical inference and stochastic optimization. In this work, we introduce an efficient adaptive Metropolis-Hastings algorithm to draw samples from generic multimodal and multidimensional target distributions. The proposal density is a mixture of Gaussian densities with all parameters (weights, mean vectors and covariance matrices) updated using all the previously generated samples applying simple recursive rules. Numerical results for the one and two-dimensional cases are provided.

Efficient Monte Carlo optimization for multi-label classifier chains

Multi-label classification (MLC) is the supervised learning problem where an instance may be associated with multiple labels. Modeling dependencies between labels allows MLC methods to improve their performance at the expense of an increased computational cost. In this paper we focus on the classifier chains (CC) approach for modeling dependencies. On the one hand, the original CC algorithm makes a greedy approximation, and is fast but tends to propagate errors down the chain. On the other hand, a recent Bayes-optimal method improves the performance, but is computationally intractable in practice. Here we present a novel double-Monte Carlo scheme (M2CC), both for finding a good chain sequence and performing efficient inference. The M2CC algorithm remains tractable for high-dimensional data sets and obtains the best overall accuracy, as shown on several real data sets with input dimension as high as 1449 and up to 103 labels.