Orientador: Marcos Júlio Rider Flores / Resumo: Nesta tese são apresentados três modelos de programação matemática que abordam os problemas de otimização relacionados ao gerenciamento da energia nos sistemas de distribuição de energia elétrica (SDEE), como: 1) Programação ótima das entregas e carregamento dos veículos elétricos (VEs) durante a navegação em um mapa de cidade, 2) Gerenciamento ótimo pelo lado da demanda considerando um sistema fotovoltaico híbrido (SFH) em uma residência em baixa tensão (RBT) no SDEE, e 3) O melhoramento do fator de carga (FC) do SDEE através do controle da demanda. O primeiro problema visa minimizar os custos relacionados com a manutenção e geração de horas extra durante a operação de uma frota de VEs, levando em conta um conjunto de entregas pre-especificadas, assim como, pontos de carregamento alocados ao longo de cada via urbana (principal e/ou secundária) pertencente ao mapa da cidade. No segundo problema, para uma residência em baixa tensão é planejado um perfil ótimo de consumo para o dia seguinte. Este perfil de consumo é obtido através de um programa de gerenciamento pelo lado da demanda (GLD) que considera uma estrutura tarifária e um esquema de operação que otimiza os recursos energéticos vindos de um SFH e o SDEE. Para cada problema de otimização é apresentado o seu correspondente modelo de programação não linear inteiro misto (PNLIM). O terceiro problema visa minimizar os custos por compra de energia (consumo e perdas de potência ativa) da concessionária, levando em conta, o co... (Resumo completo, clicar acesso eletrônico abaixo) / Doutor

Depuis une dizaine d'année, l'évolution des machines de calcul tend vers des architectures parallèles et hétérogènes. Composées de plusieurs nœuds connectés via un réseau incluant chacun des unités de traitement hétérogènes, ces grilles offrent de grandes performances. Pour programmer ces architectures, l'utilisateur doit s'appuyer sur des modèles de programmation comme MPI, OpenMP, CUDA. Toutefois, il est toujours difficile d'obtenir à la fois une bonne productivité du programmeur, qui passe par une abstraction des spécificités de l'architecture et performances. Dans cette thèse, nous proposons d'exploiter l'idée qu'un modèle de programmation spécifique à un domaine applicatif particulier permet de concilier ces deux objectifs antagonistes. En effet, en caractérisant une famille d'applications, il est possible d'identifier des abstractions de haut niveau permettant de les modéliser. Nous proposons deux modèles spécifiques au traitement du signal et de l'image sur cluster hétérogène. Le premier modèle est statique. Nous lui apportons une fonctionnalité de migration de tâches. Le second est dynamique, basé sur le support exécutif StarPU. Les deux modèles offrent d'une part un haut niveau d'abstraction en modélisant les applications de traitement du signal et de l'image sous forme de graphe de flot de données et d'autre part, ils permettent d'exploiter efficacement les différents niveaux de parallélisme tâche, données, graphe. Ces deux modèles sont validés par plusieurs implémentations et comparaisons incluant deux applications de traitement de l'image du monde réel sur cluster CPU-GPU. / Since a decade, computing systems evolved to parallel and heterogeneous architectures. Composed of several nodes connected via a network and including heterogeneous processing units, clusters achieve high performances. To program these architectures, the user must rely on programming models such as MPI, OpenMP or CUDA. However, it is still difficult to conciliate productivity provided by abstracting the architectural specificities, and performances. In this thesis, we exploit the idea that a programming model specific to a particular domain of application can achieve these antagonist goals. In fact, by characterizing a family of application, it is possible to identify high level abstractions to efficiently model them. We propose two models specific to the implementation of signal and image processing applications on heterogeneous clusters. The first model is static. We enrich it with a task migration feature. The second model is dynamic, based on the StarPU runtime. Both models offer firstly a high level of abstraction by modeling image and signal applications as a data flow graph and secondly they efficiently exploit task, data and graph parallelisms. We validate these models with different implementations and comparisons including two real-world applications of images processing on a CPU-GPU cluster.

The effective use of GPUs for accelerating applications depends on a number of factors including effective asynchronous use of heterogeneous resources, reducing data transfer between CPU and GPU, increasing occupancy of GPU kernels, overlapping data transfers with computations, reducing GPU idling and kernel optimizations. Overcoming these challenges require considerable effort on the part of the application developers. Most optimization strategies are often proposed and tuned specifically for individual applications. Message-driven executions with over-decomposition of tasks constitute an important model for parallel programming and provide multiple benefits including communication-computation overlap and reduced idling on resources. Charm++ is one such message-driven language which employs over decomposition of tasks, computation-communication overlap and a measurement-based load balancer to achieve high CPU utilization. This research has developed an adaptive runtime framework for efficient executions of Charm++ message-driven parallel applications on GPU systems. In the first part of our research, we have developed a runtime framework, G-Charm with the focus primarily on optimizing regular applications. At runtime, G-Charm automatically combines multiple small GPU tasks into a single larger kernel which reduces the number of kernel invocations while improving CUDA occupancy. G-Charm also enables reuse of existing data in GPU global memory, performs GPU memory management and dynamic scheduling of tasks across CPU and GPU in order to reduce idle time. In order to combine the partial results obtained from the computations performed on CPU and GPU, G-Charm allows the user to specify an operator using which the partial results are combined at runtime. We also perform compile time code generation to reduce programming overhead. For Cholesky factorization, a regular parallel application, G-Charm provides 14% improvement over a highly tuned implementation. In the second part of our research, we extended our runtime to overcome the challenges presented by irregular applications such as a periodic generation of tasks, irregular memory access patterns and varying workloads during application execution. We developed models for deciding the number of tasks that can be combined into a kernel based on the rate of task generation, and the GPU occupancy of the tasks. For irregular applications, data reuse results in uncoalesced GPU memory access. We evaluated the effect of altering the global memory access pattern in improving coalesced access. We’ve also developed adaptive methods for hybrid execution on CPU and GPU wherein we consider the varying workloads while scheduling tasks across the CPU and GPU. We demonstrate that our dynamic strategies result in 8-38% reduction in execution times for an N-body simulation application and a molecular dynamics application over the corresponding static strategies that are amenable for regular applications.

Le calcul scientifique parallèle est un domaine en plein essor qui permet à la fois d’augmenter la vitesse des longs traitements, de traiter des problèmes de taille plus importante ou encore des problèmes plus précis. Ce domaine permet donc d’aller plus loin dans les calculs scientifiques, d’obtenir des résultats plus pertinents, car plus précis, ou d’étudier des problèmes plus volumineux qu’auparavant. Dans le monde plus particulier de la simulation numérique scientifique, la résolution d’équations aux dérivées partielles (EDP) est un calcul particulièrement demandeur de ressources parallèles. Si les ressources matérielles permettant le calcul parallèle sont de plus en plus présentes et disponibles pour les scientifiques, à l’inverse leur utilisation et la programmation parallèle se démocratisent difficilement. Pour cette raison, des modèles de programmation parallèle, des outils de développement et même des langages de programmation parallèle ont vu le jour et visent à simplifier l’utilisation de ces machines. Il est toutefois difficile, dans ce domaine dit du “parallélisme implicite”, de trouver le niveau d’abstraction idéal pour les scientifiques, tout en réduisant l’effort de programmation. Ce travail de thèse propose tout d’abord un modèle permettant de mettre en oeuvre des solutions de parallélisme implicite pour les simulations numériques et la résolution d’EDP. Ce modèle est appelé “Structured Implicit Parallelism for scientific SIMulations” (SIPSim), et propose une vision au croisement de plusieurs types d’abstraction, en tentant de conserver les avantages de chaque vision. Une première implémentation de ce modèle, sous la forme d’une librairie C++ appelée SkelGIS, est proposée pour les maillages cartésiens à deux dimensions. Par la suite, SkelGIS, et donc l’implémentation du modèle, est étendue à des simulations numériques sur les réseaux (permettant l’application de simulations représentant plusieurs phénomènes physiques). Les performances de ces deux implémentations sont évaluées et analysées sur des cas d’application réels et complexes et démontrent qu’il est possible d’obtenir de bonnes performances en implémentant le modèle SIPSim. / Parallel scientific computations is an expanding domain of computer science which increases the speed of calculations and offers a way to deal with heavier or more accurate calculations. Thus, the interest of scientific computations increases, with more precised results and bigger physical domains to study. In the particular case of scientific numerical simulations, solving partial differential equations (PDEs) is an especially heavy calculation and a perfect applicant to parallel computations. On one hand, it is more and more easy to get an access to very powerfull parallel machines and clusters, but on the other hand parallel programming is hard to democratize, and most scientists are not able to use these machines. As a result, high level programming models, framework, libraries, languages etc. have been proposed to hide technical details of parallel programming. However, in this “implicit parallelism” field, it is difficult to find the good abstraction level while keeping a low programming effort. This thesis proposes a model to write implicit parallelism solutions for numerical simulations such as mesh-based PDEs computations. This model is called “Structured Implicit Parallelism for scientific SIMulations” (SIPSim), and proposes an approach at the crossroads of existing solutions, taking advantage of each one. A first implementation of this model is proposed, as a C++ library called SkelGIS, for two dimensional Cartesian meshes. A second implementation of the model, and an extension of SkelGIS, proposes an implicit parallelism solution for network-simulations (which deals with simulations with multiple physical phenomenons), and is studied in details. A performance analysis of both these implementations is given on real case simulations, and it demonstrates that the SIPSim model can be implemented efficiently.

[ES] En esta tesis doctoral se desarrollan modelos de programación matemática para el diseño estratégico-táctico de una red de transporte intermodal que combina dos tipos de problemas de decisiones: la localización de instalaciones y el diseño de la red de transporte. Esta combinación se reconoce en la literatura como problemas combinados LI-DR. El problema combinado se estudia para una situación real y se analizan el comportamiento de la solución óptima, a partir de distintos aspectos como: la disponibilidad del presupuesto de inversión, capacidad de las instalaciones intermodales, múltiples periodos de decisiones, interdependencia en la priorización de las decisiones, múltiples fuentes de financiación y criterios de optimización para las tres dimensiones de sostenibilidad. Para la situación real se referencian las condiciones de acceso y conectividad de la Zona de Desarrollo Económico y Social (ZODES) Magdalena Medio del departamento de Bolívar en Colombia, y el potencial de la industria agroalimentaria para esta subregión. Con el propósito de identificar factores clave que perfilen la formulación respecto a la composición y funcionamiento de los sistemas intermodales y en los problemas combinados LI-DR se analiza la literatura desde estas dos perspectivas. El análisis de la literatura ha permitido aportar dos clasificaciones novedosas e identificar retos para la investigación futura. Para la formulación de los modelos se lleva a cabo la sistemática de pasos definidos para la aplicación de las técnicas de programación matemática. Con estos pasos se logra transformar el problema del mundo real a un problema manejable con estas técnicas. La transformación favorece la interpretación matemática del problema combinado LI-DR intermodal, la modelación de los datos y la definición de una estructura de red de entrada para indexar las decisiones estratégicas y tácticas. Los modelos de programación matemática se construyen de manera gradual. En concreto, se proponen 2 versiones que se representan en 5 variantes. Se comienza formulando un modelo de programación lineal entero-mixto (MPLEM) mono-periodo para analizar, desde un enfoque económico, la sensibilidad de las capacidades del sistema intermodal, la capacidad financiera de los tomadores de decisiones y la variación de la demanda. A continuación, sobre la base de este modelo se propone un MPLEM multi-periodo y dos variantes para validar las condiciones de interdependencia en la toma de decisiones estratégica y la participación de múltiples actores en la financiación de los proyectos de inversión. Finalmente, se formula un MPLEM multi-objetivo para optimizar simultáneamente las tres dimensiones de sostenibilidad. Para resolver y validar los modelos se implementaron dos esquemas de resolución. En los esquemas se utilizan los lenguajes de programación R y Python con el software de optimización matemática Gurobi Optimizer. Se realizan experimentos numéricos para distintos escenarios y se analiza el comportamiento de las soluciones considerando distintos valores a los parámetros. Los resultados obtenidos permiten comprobar la utilidad de los modelos matemáticos e identificar las principales limitaciones y futuras líneas de trabajo. / [CA] En aquesta tesi doctoral es desenvolupen models de programació matemàtica per al disseny estratègic-tàctic d'una xarxa de transport intermodal que combina dos tipus de problemes de decisions: la localització d'instal·lacions i el disseny de la xarxa de transport. Aquesta combinació es reconeix en la literatura com problemes combinats LI-DR. El problema combinat s'estudia per a una situació real i s'analitzen el comportament de la solució òptima, a partir de diferents aspectes com: la disponibilitat de l'pressupost d'inversió, capacitat de les instal·lacions intermodals, múltiples períodes de decisions, interdependència en la priorització de les decisions, múltiples fonts de finançament i criteris d'optimització per a les tres dimensions de sostenibilitat. Per a la situació real es referencien les condicions d'accés i connectivitat de la Zona de Desenvolupament Econòmic i Social (ZODES) Magdalena Medio de el departament de Bolívar a Colòmbia, i el potencial de la indústria agroalimentària per a aquesta subregió. Amb el propòsit d'identificar factors clau que perfilin la formulació respecte a la composició i funcionament dels sistemes intermodals i en els problemes combinats LI-DR s'analitza la literatura des d'aquestes dues perspectives. L'anàlisi de la literatura ha permès aportar dues classificacions noves i identificar reptes per a la investigació futura. Per a la formulació dels models es porta a terme la sistemàtica de passos definits per l'aplicació de les tècniques de programació matemàtica. Amb aquests passos s'aconsegueix transformar el problema de l'món real a un problema manejable amb aquestes tècniques. La transformació afavoreix la interpretació matemàtica de el problema combinat LI-DR intermodal, la modelació de les dades i la definició d'una estructura de xarxa d'entrada per indexar les decisions estratègiques i tàctiques. Els models de programació matemàtica es construeixen de manera gradual. En concret, es proposen 2 versions que es representen en 5 variants. Es comença formulant un model de programació lineal sencer-mixt (MPLEM) mono-període per analitzar, des d'un enfocament econòmic, la sensibilitat de les capacitats de sistema intermodal, la capacitat financera dels prenedors de decisions i la variació de la demanda. A continuació, sobre la base d'aquest model es proposa un MPLEM multi-període i dues variants per validar les condicions d'interdependència en la presa de decisions estratègica i la participació de múltiples actors en el finançament dels projectes d'inversió. Finalment, es formula un MPLEM multi-objectiu per optimitzar simultàniament les tres dimensions de sostenibilitat. Per resoldre i validar els models es van implementar dos esquemes de resolució. En els esquemes s'utilitzen els llenguatges de programació R i Python amb el programari d'optimització matemàtica Gurobi Optimizer. Es realitzen experiments numèrics per a diferents escenaris i s'analitza el comportament de les solucions considerant diferents valors als paràmetres. Els resultats obtinguts permeten comprovar la utilitat dels models matemàtics i identificar les principals limitacions i futures línies de treball. / [EN] In this doctoral thesis, mathematical programming models are developed aiming at the strategic-tactical design of an intermodal transport network that combines two types of decision problems: the location of facilities and the transport network design. In the literature, this combination is recognized as combined LI-DR problems. The combined problem is studied for a real situation and the performance of the optimal solution is analyzed in relation to different aspects such as the investment budget availability, the intermodal facilities capacity, multiple decision periods, interdependence in the prioritization of the decisions, multiple sources of funding and optimization criteria for the three dimensions of sustainability. In what concern to the real situation, the access and connectivity conditions of the Magdalena Medio Economic and Social Development Zone (ZODES, for its acronym in Spanish) of the Bolívar department in Colombia, and the potential of the agri-food industry for this subregion, are considered. Aiming the identification of key factors that outlines the formulation regarding the composition and operation of intermodal systems and, in combined LI-DR problems, the literature is reviewed from these two perspectives. The literature analysis has made it possible to provide two novel classifications and to identify challenges for future research. The formulation of models follows the systematic steps already defined for the application of mathematical programming techniques. Following these steps, it is possible to transform the problem from a real-world problem to a manageable one. The transformation promotes the mathematical interpretation of the intermodal LI-DR combined problem, the data modeling, and the definition of an input network structure to index strategic and tactical decisions. Mathematical programming models are built gradually. Specifically, 2 versions are proposed, which are represented by 5 variants. Firstly, it is formulated a single-period mixed-integer linear programming model (MILPM) in order to analyze, from an economic perspective, the sensitivity of the intermodal system capacities, the financial capacity of the decision-makers, and the demand changes. Based on the aforementioned model, a multi-period MILPM and two variants are proposed aiming to validate the conditions of interdependence in strategic decision-making and the participation of multiple actors in the investment projects financing. Finally, a multi-objective MILPM is formulated to simultaneously optimize all three dimensions of sustainability. To solve and validate the models, two resolution schemes were implemented. The schematics use the R and Python programming languages with the mathematical optimization software Gurobi Optimizer. Numerical tests are carried out for different scenarios and the performance of the solutions is analyzed considering different values for the parameters. The results obtained allow us to verify the usefulness of the models proposed and identify the main limitations and future lines of work. / Agamez Arias, ADM. (2021). Modelos de optimización para el diseño estratégico-táctico de una red de transporte intermodal [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/177015 / TESIS