Bygg- och fastighetssektorn står för omkring 22 procent av Sveriges totala årliga växthusgasutsläpp, vilket understryker dess betydande miljöpåverkan. Utsläppen härrör från produktion av byggmaterial, hantering av bygg- och rivningsavfall samt utvinning av fossila bränslen för transporter och maskiner (Boverket, 2024). Trots att branschen strävar efter att minska sitt koldioxidavtryck genom ny teknik och digitalisering, utgör dess traditionella och konservativa natur en utmaning för den digitala transformationen (Fastighetsägarna, 2024). Denna studie utforskar rollen som digitala lösningar, särskilt AI-baserad energioptimering, kan spela för att öka energieffektiviteten inom fastighetssektorn och minska dess klimatpåverkan. Behovet av hållbara energilösningar har intensifierats av de historiskt höga energipriserna i Europa sedan sommaren 2021, drivna av geopolitiska spänningar och stigande fossilbränslepriser (Naturskyddsföreningen, 2022). FN:s Agenda 2030-mål 7 (Överkomlig och ren energi) och 11 (Hållbara städer och samhällen) ger ytterligare incitament för fastighetsägare att fokusera på energieffektivitet, i linje med globala hållbarhetsmål (Regeringskansliet, 2022). Dessutom erbjuder EU:s gröna taxonomi ett ramverk för att identifiera och rapportera miljömässigt hållbara aktiviteter, relevant för fastighetsbolag som använder AI-baserad energioptimering för att bidra till begränsningen av klimatförändringarna (SEB, 2021; Birlev, 2023). AI-system kan övervaka och analysera energiförbrukningen för att identifiera ineffektiviteter och optimera driften, vilket minskar energianvändningen och koldioxidutsläppen (Byggvesta, 2023; Piigab, 2023). Implementeringen av AI-teknik möter dock utmaningar, inklusive behovet av omfattande data samt höga kostnader för hårdvara och personalutbildning (Bratt & Philipson., 2023). Denna studie ger insikter om hur AI kan främja hållbarhet inom fastighetssektorn, och belyser nödvändigheten av att övervinna tekniska, ekonomiska och beteendemässiga hinder för en framgångsrik implementering. / The building and real estate sector contributes approximately 22% of Sweden's total annual greenhouse gas emissions, underscoring its significant environmental impact. Emissions arise from building material production, construction and demolition waste management, and fossil fuel extraction for transportation and machinery (Boverket, 2024). While the industry strives to reduce its carbon footprint through new technologies and digitalization, its traditional and conservative nature poses challenges to digital transformation (Fastighetsägarna, 2024). This study explores the role of digital solutions, particularly AI-based energy optimization, in enhancing energy efficiency within the real estate sector and reducing its carbon footprint. The urgency for sustainable energy solutions has intensified with historically high energy prices across Europe since summer 2021, driven by geopolitical tensions and fossil fuel price surges (Naturskyddsföreningen, 2022). The UN's Agenda 2030 goals 7 (Affordable and Clean Energy) and 11 (Sustainable Cities and Communities) provide additional incentives for property owners to focus on energy efficiency, aligning with global sustainability targets (Regeringskansliet, 2022). Additionally, the EU's green taxonomy offers a framework for identifying and reporting environmentally sustainable activities, relevant for real estate companies using AI-based energy optimization to contribute to climate change mitigation (SEB, 2021; Birlev, 2023). AI systems can monitor and analyze energy consumption to identify inefficiencies and optimize operations, reducing energy usage and carbon emissions (Byggvesta, 2023; Piigab, 2023). However, implementing AI technology faces challenges, including the need for extensive data and the high costs of hardware and personnel training (Bratt & Philipson, 2023). This study provides insights into how AI can promote sustainability in the real estate sector, highlighting the necessity of overcoming technical, economic, and behavioral barriers for successful implementation.

<p dir="ltr">The energy consumption in the building sector is responsible for over 36% of the total energy consumption across the globe. Of all the energy-consumer devices within a building, heating, ventilation, and air conditioning (HVAC) systems account for over 50% of the total energy consumed. This makes HVAC systems a source of preventable and unexplored energy waste that can be tackled by incorporating intelligent operations. Since its inception, model predictive control (MPC) has been one of the prospective solutions for HVAC management systems to reduce both costs and energy usage. Additionally, MPC is becoming increasingly practical as the processing capacity of building automation systems increases and a large quantity of monitored building data becomes available. MPC also provides the potential to improve the energy efficiency of HVAC systems via its capacity to consider limitations, to predict disruptions, and to factor in multiple competing goals such as interior thermal comfort and building energy consumption. In this regard, the opening chapter delves into the evolving landscape of the HVAC industry. It explores how rapid advancements in technology, growing concerns about climate change, and the ever-present need for energy efficiency are driving innovation. The chapter highlights the shift from static to dynamic HVAC systems, where buildings become sensor-rich networks enabling advanced control strategies like Model Predictive Control (MPC) and Fault Detection and Diagnosis (FDD). we first provide a comprehensive review of the literature concerning the application of MPC in HVAC systems. Detailed discussions of modeling approaches and optimization algorithms are included. Numerous design aspects such as prediction horizon, time step, and cost function, that impact MPC performance are discussed in detail. The technical characteristics, advantages, and disadvantages of various types of modeling software are discussed. Next, a thorough, real-world case study for the design and implementation of a generalized data-collection and control architecture for HVAC systems in an educational building is proposed. The proposed MPC method adds a supervisory control layer on top of the current BMS by delivering temperature setpoints to the legacy controller. This means that the technique may be used to a variety of current HVAC systems in different commercial buildings. In addition, the utilization of remote web services to host the cloud-based architecture significantly minimizes the amount of technical expertise generally necessary to create such systems. In addition, we provide significant lessons learned from the installation process and we list indicative prices, therefore minimizing uncertainty for other researchers and promoting the use of comparable solutions. Chapter two focuses on Fault Detection and Diagnosis (FDD), a critical component of maintaining optimal HVAC performance and minimizing energy waste. HVAC systems are susceptible to malfunctions over time, leading to increased energy consumption and higher maintenance costs. FDD techniques play a vital role in identifying and diagnosing these faults early on, allowing for timely repairs and preventing further deterioration. This chapter introduces a novel bi-level machine learning framework for diagnosing faults in air handling units. This framework addresses key challenges associated with FDD. A bi-level machine learning framework is developed for diagnosing faults in air handling units (AHUs) and rooftop units (RTUs) based on principal component analysis (PCA), time series anomaly detection, and random forest (RF). By proposing this framework, we address three persistent challenges in this field: (I) minimizing false positives; (II) accounting for data imbalance; and (III) normal condition monitoring of equipment. It is shown that PCA can reduce the dataset dimension with one principal component accounting for 95% of data variance. Also, the random forest could classify the faults with 89% precision for single-zone AHU, 85% precision for RTU, and 79% for multi-zone AHU. Chapter three tackles the practical implementation of Model Predictive Control (MPC) in a real-world commercial building setting. It details the development, implementation, and cost analysis of a universally applicable cloud-based MPC framework for HVAC control systems. This chapter offers valuable insights into the feasibility and effectiveness of MPC in achieving energy efficiency goals while maintaining occupant comfort. The chapter delves into the hardware and software components used for data acquisition and MPC implementation. It emphasizes the use of cloud-based microservices to ensure seamless integration with existing building management systems, promoting wider adoption of this advanced control strategy. Three innovative control strategies are presented and evaluated in this chapter. The chapter presents compelling evidence for the effectiveness of these strategies, showcasing significant energy savings of up to 19.21%. Chapter four focuses on Occupancy-based Demand Controlled Ventilation (DCV) as a means to optimize indoor air quality (IAQ) while minimizing energy consumption. This chapter highlights the growing importance of IAQ in the wake of the COVID-19 pandemic and its impact on occupant health and well-being. Current ventilation standards often rely on static occupancy assumptions, which can lead to over-ventilation during unoccupied pe riods and wasted energy. This chapter proposes a dynamic occupant behavior model using machine learning algorithms to predict CO2 concentrations within buildings. The chapter investigates the performance of various machine learning algorithms, ultimately identify ing a Multilayer Perceptron (MLP) as the most effective in predicting CO2 levels under dynamic occupancy conditions. This model allows for real-time modulation of ventilation rates, ensuring adequate IAQ while minimizing energy consumption. The concluding chapter presents experimental findings on the effectiveness of adaptive Variable Frequency Drive (VFD) control strategies in optimizing HVAC energy consump tion. Variable Frequency Drives allow for adjusting the speed of electric motors, including those powering HVAC fans. This chapter explores the potential of using real-time occu pancy predictions to optimize VFD operation. The proposed control strategy demonstrates impressive energy savings, achieving a 51.4% reduction in HVAC fan energy consumption while adhering to ASHRAE IAQ standards. This chapter paves the way for occupant-centric ventilation strategies that prioritize both human health and energy efficiency. These results underscore the potential of predictive control systems to transform building operations to ward greater sustainability and efficiency. The chapter acknowledges the need for further validation through extended monitoring and analysis. In summary, this thesis contributes significantly to the advancement of smart building technologies by proposing practical frameworks for implementing advanced control strategies in HVAC systems. The findings presented here offer valuable insights for building designers, engineers, facility managers, and policymakers interested in creating sustainable, energy efficient, and occupant-centric buildings. The developed frameworks have the potential to be applied across a wide range of building types and climatic conditions, promoting broader adoption of smart building technologies and contributing to a more sustainable built environment.</p>

Energieffektivisering börjar bli ett eftersträvande mål runtom i världen. Detta grundar sig i att energiförbrukningen för byggnader uppgår till ca 40 % globalt, en siffra som man vill få ner. Men att uppnå en energieffektiv byggnad är inte lätt. Detta kan göras på många och olika sätt. Ett av dem är att energieffektivisera fönstren, som är en byggnads svagaste punkt pga dess höga U-värde. Val av fönster är inte lätt, och det finns ett flertal alternativ att välja bland. I kalla klimat som Sverige söker man fönster med lågt U-värde och högt g-värde, samt en hög avskärmningsfaktor. I varmare länder vill man också ha ett lågt U-värde hos fönster men fokusen ligger främst på en låg avskärmningsfaktor. Syftet med uppsatsen var att hitta de mest energieffektiva fönstren, oavsett kostnad, för byggnader som befinner sig i länder med kallare klimat. Det var också av vikt att få veta lönsamheten för fönstren i fråga, därför har även kostnadsfrågor belysts. Metodvalen var informationssökning i olika databaser och litteratur samt att olika företag inom fönsterbranschen kontaktades, vilket ledde till att relevant och önskvärd information erhölls. Därefter fortskred arbetet genom kalkyleringar för energibalansen och lönsamheten. Vacuumfönster, aerogelfönster samt kromogena fönster hör till framtida fönster som kan tillföra positiva inverkan på energibalansen för byggnader. Men dessa fönster är i nuläget inte helt färdigutvecklade, fast har potential att bli världsledande. Vacuumfönster och kromogena fönster är i nuläget bättre lämpade för varmare klimat. Lyckas man komma längre med deras nutida utveckling är det inte omöjligt att anpassa de för kallare klimat. Aerogelfönster ger mest energibesparing vid byte av fönster, men pga vissa optiska egenskaper och komplicerad tillverkning av produkten är den i nuläget inte optimal vid ett fönsterbyte. De framtida fönstren är ej heller ekonomiskt försvarbara, det finns i dagsläget kommersiella energieffektiva fönster som är billigare att införskaffa och ger ett ansenligt bra resultat för en byggnads energibalans. / Energy optimization is starting to be a pursued worldwide main goal. This is based on the global energy consumption that occurs in buildings, which is about 40 percent. There is no doubt that this value needs to be lowered. But to achieve an energy efficient building is not easy. Although, this can be done in many and different ways. One of them is to optimize the windows, which is a buildings weakest point due to its high U-value.The choice of windows can be a harsh decision, there’s plenty of windows to choose among. In heating dominated climates, as the one in Sweden, it is necessary to choose windows with low U-values and high g-values, also a high solar heat gain coefficient/shading coefficient is required. A window with a low U-value is also important in cooling dominated climates but the main focus is instead on a low shading coefficient, which is not the case in this thesis. The purpose is to find the most energy efficient window that lowers the need for active heating in buildings, and also reveal and discuss the cost issues for the chosen windows.By searching in scientific databases and contacting companies active in the window industry the desired information was obtained. Calculations including the energy balance and present value were made, which gave an indication of the profitability for the different windows. Vacuum, aerogel and chromogenic window are examples of future windows which can have a positive impact on the energy balance in buildings. Yet these windows are currently not fully developed, but have potential to be highly valuable types of windows. Vacuum and chromogenic windows are better suited for cooling dominated climates. However if the development succeed where a big progress will be made it will not be impossible to suit them for heating dominated climates too. Aerogel windows have the best impact on the energy savings when replacing windows, but due to some optical attributes and a complicated manufacturing of the product aerogel windows are currently not an optimal choice for window replacement. The future windows isn’t either economically viable. For now, there are other commercially energy efficient windows that are cheaper to purchase. They also show an acceptable good result on the energy balance for a building.

Det finns ekonomiska, säkerhetsmässiga och miljömässiga vinster att göra genom att tillämpa ecodriving. Tidigare forskning pekar på att transportslagen bilar, tåg och flyg har gjort stora besparingar i ekonomiskt och miljömässigt hänseende, men hur ser det ut inom sjöfarten? Med denna frågeställning som bakgrund är syftet med studien att studera intresset för ecodriving bland svenska rederier och svenska myndigheter med en fartygsflotta. Dataunderlaget för studien utgörs av material från kvalitativa intervjuer med personer i exekutiv position. Resultatet av studien visar på att de flesta verksamheterna står i startgroparna eller redan arbetar utifrån en eller flera aktivt valda metoder för ecodriving. Resultatet visar också att det finns en blandning av förutsättningar och uppfattningar om vad ecodriving är och vad det kan bli inom sjöfarten. När frågor om automatisering i samband med ecodriving behandlas är resultatet tvetydigt. / There are economical, safetylike and environmental benefits of applying eco-driving. Previous research has shown that cars, trains and aviation have made significant savings economically and environmentaly speaking, but how does that transcend into the maritime business? With this question as a background, the aim with this thesis is to examine the interest of ecodriving among Swedish shipowners and authorities. The data in this thesis is derived from qualitative interviews with employees in executive land-based positions. The result shows that several of the shipowners and authorities are in the starting pits or are already conducting one or more eco-driving methods in their operations. The result also shows that there is a variety of prerequisites and perceptions of what eco-driving is and what is could become in the future among the respondents. When questions about automatization in relation to eco-driving are brought up, the result is ambiguous.

Ce travail s'inscrit dans le cadre du projet FUI RIDER (Research for IT Driven Energy efficiency) qui vise à développer un système de gestion de l'énergie faiblement dépendant du bâtiment à contrôler et propose une nouvelle approche pour réduire les coûts énergétiques. Cette approche exploite la notion de confort thermique afin de calculer de nouvelles consignes à fournir au système de contrôle du conditionnement du bâtiment. L'approche s'appuie sur l'idée que le confort thermique est une notion multidimensionnelle subjective. La littérature propose des modèles statistiques pour appréhender le confort thermique. Malheureusement, ces modèles sont fortement non linéaires et non interprétables ce qui rend difficile leur utilisation pour la conduite ou l'optimisation. Nous proposons un nouveau modèle de confort basé sur la théorie de l'utilité multi attributs et les intégrales de Choquet. L'intérêt d'un tel modèle est qu'il est interprétable en termes de préférences pour la conduite, linéaire par simplexe ce qui facilite la résolution des problèmes d'optimisation, et plus concis qu'un système de contrôle à base de règles. Dans la seconde partie de ce travail, le THermal Process Enhancement (THPE) s'intéresse à l'obtention efficiente des consignes calculées avec le modèle du confort thermique. Le THPE se base sur un raisonnement approché établi à partir d'un modèle qualitatif enrichi EQM (Extended Qualitative Model). L'EQM est le résultat de l'étude mathématique et qualitative des équations différentielles régissant les processus thermiques. Il est enrichi en continu par un système de gestion de l'expérience basé sur un apprentissage avec pénalités qui fournit les informations quantitatives nécessaires pour inférer des recommandations de conduite quantifiées à partir des tendances modélisées dans l'EQM. L'EQM et les raisonnements associés requièrent peu de paramètres et sont opérationnels même si la base d'apprentissage est initialement vide au lancement de RIDER. Le système de gestion de l'expérience permet simplement de quantifier les recommandations et de converger plus vite vers une commande optimale. Le raisonnement à base de modèles qui supporte notre approche est faiblement dépendant du processus thermique, pertinent dès le lancement de RIDER et se prête facilement au changement d'échelle de l'analyse thermique d'un bâtiment. Les performances de notre THPE, sa stabilité et son adaptation par rapport aux variations de l'environnement sont illustrées sur différents problèmes de contrôle et d'optimisation. Les commandes optimales sont généralement obtenues en quelques itérations et permettent d'avoir un contrôle adaptatif et individuel des pièces d'un bâtiment. / The present work is part of the FUI RIDER project (Research for IT Driven Energy efficiency). It aims to develop an energy management system that has to be weakly dependent on building's specificities in order to be easily deployed in different kinds of buildings. This work proposes a new approach based on the thermal comfort concept in order to reduce energy costs. This approach takes advantage of the thermal comfort concept in order to compute new optimized setpoints for the building energy control system. It relies on the idea that thermal comfort is a subjective multidimensional concept that can be used to reduce energy consumption. The literature provides statistical thermal comfort models but their complexity and non-linearity make them not useful for the control and optimization purposes. Our new thermal comfort model is based on the multi attributes utility theory and Choquet integrals. The advantages of our model are: its interpretability in term of preference relationships, its linearity in simplex regions which simplifies optimization problems' solving, and its compact form which is more tractable than a rule based control formalism. In the second part of this work, the THermal Process Enhancement (THPE) proposes a control system approach to efficiently reach the optimized setpoints provided by the comfort model. The THPE proposes an efficient and simple thermal control approach based on imprecise knowledge of buildings' special features. Its weak data-dependency ensures the scalability and simplicity of our approach. For this, an extended thermal qualitative model (EQM) is proposed. It is based on a qualitative description of influences that actions' parameters may have on buildings' thermal performances. This description results from the mathematical and qualitative analysis of dynamical thermal behaviors. Our thermal qualitative model is then enriched by online collecting and assessing previous thermal control performances. The online learning provides the necessary quantitative information to infer quantified control recommendations from the qualitative tendencies displayed by the EQM. Thus, an approximate reasoning based on the EQM and an online learning coupled with a penalty function provides smart thermal control functionalities. The EQM based approximate reasoning guarantees our control system weak dependency with regard to the building special features as well as its multi-scale applicability and its relevancy even for RIDER's first start when the learning database lacks of information. The performances of our THPE are assessed on various types of control and optimization issues. An optimal control is generally achieved in a few iterations which allows providing an adaptive and individual control of building's rooms.

La thèse porte sur la conception de protocoles de routage pour les réseaux de capteurs. Les problèmes de recherche du routage de données dans un réseau multi-sauts sont d’une part l’optimisation de l’énergie et d’autre part le routage sous contraintes de la qualité de service (QoS) multicritères (e.g., énergie, fiabilité, délai, …). Cette thèse apporte deux contributions par rapport à l’état de l’art : une optimisation d'un protocole de routage probabiliste pour l'équilibre de l'usage d'énergie et un protocole de routage capable de prendre en compte simultanément des métriques de QoS multiples. En effet, pour équilibrer la consommation de l’énergie du routage lorsque des chemins multiples existent, les protocoles de routage probabiliste existants affectent une probabilité de choix à chaque chemin, soit de façon empirique, soit proportionnelle au niveau de l’énergie disponible du chemin. Nous ne savions pas quelles sont les probabilités optimales qui permettent d’avoir la durée de vie maximale du réseau. Cette thèse a permis d’établir ces probabilités optimales à l’aide de la modélisation sous forme d’un problème d’optimisation linéaire. Quant au problème du routage multicritères, bien que des métriques multiples soient définies par RPL (un standard d’IETF), les protocoles existants choisissent la route soit sur une métrique, soit sur une fonction de coût combinant plusieurs (qui introduit par conséquent un biais de pondération), mais jamais plusieurs simultanément. Dans cette thèse, nous avons d’abord évalué numériquement les performances de l’approche « operator calculus algebra » introduit par R. Schott et S. Staples qui définit un algorithme efficace permettant de trouver tous les chemins satisfaisant les contraintes multiples dans un graphe , puis dérivé une version distribuée sur laquelle nous avons conçu un protocole de routage multi-métriques. Ces deux contributions ont été implémentées dans l’environnement Contiki et émulées/simulées sous Cooja (un logiciel permettant de simuler des protocoles des réseaux de capteurs) / In this thesis, we focus on routing protocols for Wireless Sensor Networks (WSNs). The main research problems in the domain of routing data packets in a multi-hop network are the optimisation of the energy and the routing under multi-criteria QoS constraints (e.g., energy, reliability, delay, …). To address these problems, this dissertation proposes two contributions. Firstly, an optimal probabilistic routing protocol which balances the usage of energy and secondly, a routing protocol which is able to simultaneously take into account multiple QoS metrics. In fact, for balancing the energy consumption between the multiple existing links, the existing probabilistic routing protocols assign a probability to each link, either in an empirical way or depending on proportional energy level of the path. We did not know what are the optimal probabilities which lead to the optimal network lifetime. Our first contribution proposes optimal probabilities by modeling and solving a linear programming problem. As for the multi-constrainted QoS routing problem, multiple metrics are defined by RPL (a standard of IETF) but the existing protocols chose paths either according to only one metric or using a single aggregated function with multiple metrics, but never all the metrics simultaneously. In this dissertation, we first evaluate the performance of the operator calculus algebra introduced by R. Schott and S. Staples which defines an efficient algorithm allowing to find all the paths which satisfy the multiple constraints in a graph, and secondly we proposed a distributed version of this algorithm based on which a routing protocol has been designed. Both contributions are implemented in Contiki environment and simulated/emulated under Cooja (a software designed for simulating protocols of WSNs)

Submitted by PPG Ci?ncia da Computa??o (ppgcc@pucrs.br) on 2018-05-22T12:22:46Z No. of bitstreams: 1 ANDR?_LU?S_DEL_MESTRE_MARTINS_TES.pdf: 10284806 bytes, checksum: 089cdc5e5c91b6ab23816b94fdbe3d1d (MD5) / Approved for entry into archive by Sheila Dias (sheila.dias@pucrs.br) on 2018-06-04T11:21:09Z (GMT) No. of bitstreams: 1 ANDR?_LU?S_DEL_MESTRE_MARTINS_TES.pdf: 10284806 bytes, checksum: 089cdc5e5c91b6ab23816b94fdbe3d1d (MD5) / Made available in DSpace on 2018-06-04T11:37:12Z (GMT). No. of bitstreams: 1 ANDR?_LU?S_DEL_MESTRE_MARTINS_TES.pdf: 10284806 bytes, checksum: 089cdc5e5c91b6ab23816b94fdbe3d1d (MD5) Previous issue date: 2018-03-19 / Sistemas many-core integram m?ltiplos cores em um chip, fornecendo alto desempenho para v?rios segmentos de mercado. Novas tecnologias introduzem restri??es de pot?ncia conhecidos como utilization-wall ou dark-silicon, onde a dissipa??o de pot?ncia no chip impede que todos os PEs sejam utilizados simultaneamente em m?ximo desempenho. A carga de trabalho (workload) em sistemas many-core inclui aplica??es tempo real (RT), com restri??es de vaz?o e temporiza??o. Al?m disso, workloads t?picos geram vales e picos de utiliza??o de recursos ao longo do tempo. Este cen?rio, sistemas complexos de alto desempenho sujeitos a restri??es de pot?ncia e utiliza??o, exigem um gerenciamento de recursos (RM) multi-objetivos capaz de adaptar dinamicamente os objetivos do sistema, respeitando as restri??es impostas. Os trabalhos relacionados que tratam aplica??es RT aplicam uma an?lise em tempo de projeto com o workload esperado, para atender ?s restri??es de vaz?o e temporiza??o. Para abordar esta limita??o do estado-da-arte, ecis?es em tempo de projeto, esta Tese prop?e um gerenciamento hier?rquico de energia (REM), sendo o primeiro trabalho que considera a execu??o de aplica??es RT e ger?ncia de recursos sujeitos a restri??es de pot?ncia, sem uma an?lise pr?via do conjunto de aplica??es. REM emprega diferentes heur?sticas de mapeamento e de DVFS para reduzir o consumo de energia. Al?m de n?o incluir as aplica??es RT, os trabalhos relacionados n?o consideram um workload din?mico, propondo RMs com um ?nico objetivo a otimizar. Para tratar esta segunda limita??o do estado-da-arte, RMs com objetivo ?nico a otimizar, esta Tese apresenta um gerenciamento de recursos multi-objetivos adaptativo e hier?rquico (MORM) para sistemas many-core com restri??es de pot?ncia, considerando workloads din?micos com picos e vales de utiliza??o. MORM pode mudar dinamicamente os objetivos, priorizando energia ou desempenho, de acordo com o comportamento do workload. Ambos RMs (REM e MORM) s?o abordagens multi-objetivos. Esta Tese emprega o paradigma Observar-Decidir-Atuar (ODA) como m?todo de projeto para implementar REM e MORM. A Observa??o consiste em caracterizar os cores e integrar monitores de hardware para fornecer informa??es precisas e r?pidas relacionadas ? energia. A Atua??o configura os atuadores do sistema em tempo de execu??o para permitir que os RMs atendam ?s decis?es multi-objetivos. A Decis?o corresponde ? implementa??o do REM e do MORM, os quais compartilham os m?todos de Observa??o e Atua??o. REM e MORM destacam-se dos trabalhos relacionados devido ?s suas caracter?sticas de escalabilidade, abrang?ncia e estimativa de pot?ncia e energia precisas. As avalia??es utilizando REM em manycores com at? 144 cores reduzem o consumo de energia entre 15% e 28%, mantendo as viola??es de temporiza??o abaixo de 2,5%. Resultados mostram que MORM pode atender dinamicamente a objetivos distintos. Comparado MORM com um RM estado-da-arte, MORM otimiza o desempenho em vales de workload em 11,56% e em picos workload em at? 49%. / Many-core systems integrate several cores in a single die to provide high-performance computing in multiple market segments. The newest technology nodes introduce restricted power caps so that results in the utilization-wall (also known as dark silicon), i.e., the on-chip power dissipation prevents the use of all resources at full performance simultaneously. The workload of many-core systems includes real-time (RT) applications, which bring the application throughput as another constraint to meet. Also, dynamic workloads generate valleys and peaks of resources utilization over the time. This scenario, complex high-performance systems subject to power and performance constraints, creates the need for multi-objective resource management (RM) able to dynamically adapt the system goals while respecting the constraints. Concerning RT applications, related works apply a design-time analysis of the expected workload to ensure throughput constraints. To cover this limitation, design-time decisions, this Thesis proposes a hierarchical Runtime Energy Management (REM) for RT applications as the first work to link the execution of RT applications and RM under a power cap without design-time analysis of the application set. REM employs different mapping and DVFS (Dynamic Voltage Frequency Scaling) heuristics for RT and non-RT tasks to save energy. Besides not considering RT applications, related works do not consider the workload variation and propose single-objective RMs. To tackle this second limitation, single-objective RMs, this Thesis presents a hierarchical adaptive multi-objective resource management (MORM) for many-core systems under a power cap. MORM addresses dynamic workloads with peaks and valleys of resources utilization. MORM can dynamically shift the goals to prioritize energy or performance according to the workload behavior. Both RMs (REM and MORM), are multi-objective approaches. This Thesis employs the Observe-Decide-Act (ODA) paradigm as the design methodology to implement REM and MORM. The Observing consists on characterizing the cores and on integrating hardware monitors to provide accurate and fast power-related information for an efficient RM. The Actuation configures the system actuators at runtime to enable the RMs to follow the multi-objective decisions. The Decision corresponds to REM and MORM, which share the Observing and Actuation infrastructure. REM and MORM stand out from related works regarding scalability, comprehensiveness, and accurate power and energy estimation. Concerning REM, evaluations on many-core systems up to 144 cores show energy savings from 15% to 28% while keeping timing violations below 2.5%. Regarding MORM, results show it can drive applications to dynamically follow distinct objectives. Compared to a stateof- the-art RM targeting performance, MORM speeds up the workload valley by 11.56% and the workload peak by up to 49%.

Depuis trois décennies, la tendance du marché répond à la demande actuelle de miniaturisation et d'augmentation de performances des appareils multimédias. Or, toute réduction des dimensions d'un facteur donné impose une diminution des tensions (pour des raisons de fiabilité). Afin d'y répondre, la réduction de taille des circuits intégrés CMOS atteint des échelles d'intégration submicroniques entrainant une baisse importante de la fiabilité des composants et en particulier des transistors. La création de porteurs chauds, ainsi que la dissipation thermique à l'intérieur des circuits submicroniques, sont les deux phénomènes physiques principaux à l'origine de la baisse de fiabilité. La solution technique permettant de garder un bon degré de fiabilité, tout en réduisant la taille des composants, consiste à réduire la tension d'alimentation des circuits. Parallèlement aux contraintes de performances, les normes environnementales demandent une consommation la plus réduite possible. La difficulté consiste alors en la réalisation de circuits associant une alimentation basse puissance (tension et courant) d'où la notion de circuits " Low Power ". Ces circuits sont pour certains déjà utilisés dans le domaine du multimédia, du médical, avec des contraintes d'intégration différentes (possibilité de composants externes, stabilité, etc.). L'augmentation des performances en vitesse des circuits digitaux nécessite par ailleurs l'utilisation de technologies générant des fuites de plus en plus importantes qui sont incompatibles avec une réduction de la consommation dans des modes de veille sans la mise en place de nouvelles techniques / For three decades, the market trend answers the current demand of miniaturization and performance increase of the multimedia devices. Yet, any reduction of the dimensions of a given factor imposes a decrease of the tensions (for reasons of reliability). To answer this question, the downsizing of CMOS integrated circuits reaches submicron scales of integration resulting in a significant decrease in the reliability of components and in particular transistors. The hot carriers creations, as well as heat dissipation within the submicron circuits, are the two main physical phenomena behind the reliability decline. The technical solution to maintain a good degree of reliability, while reducing component size, is to reduce the supply voltage of circuits. In parallel to performance constraints, environmental standards require consumption as small as possible. The challenge is then to build circuits combining low power supply (voltage and current) where the concept of circuits "Low Power". These circuits are used for some already in the field of multimedia, medical, integration with various constraints (possibility of external components, stability, etc..). The speed increase performance of digital circuits also requires the use of technologies that generate leaks increasingly important that are inconsistent with consumption reduction in standby modes without the introduction of new techniques.

Au cours des dernières années, l'informatique “green” est devenue un sujet important dans le calcul intensif. Cependant, les plates-formes informatiques continuent de consommer de plus en plus d'énergie en raison de l'augmentation du nombre de noeuds qui les composent. Afin de minimiser les coûts d'exploitation de ces plates-formes de nombreuses techniques ont été étudiées, parmi celles-ci, il y a le changement de la fréquence dynamique des processeurs (DVFS en anglais). Il permet de réduire la consommation d'énergie d'un CPU, en abaissant sa fréquence. Cependant, cela augmente le temps d'exécution de l'application. Par conséquent, il faut trouver un seuil qui donne le meilleur compromis entre la consommation d'énergie et la performance d'une application. Cette thèse présente des algorithmes développés pour optimiser la consommation d'énergie et les performances des applications parallèles avec des itérations synchrones et asynchrones sur des clusters ou des grilles. Les modèles de consommation d'énergie et de performance proposés pour chaque type d'application parallèle permettent de prédire le temps d'exécution et la consommation d'énergie d'une application pour toutes les fréquences disponibles.La contribution de cette thèse peut être divisé en trois parties. Tout d'abord, il s'agit d'optimiser le compromis entre la consommation d'énergie et les performances des applications parallèles avec des itérations synchrones sur des clusters homogènes. Deuxièmement, nous avons adapté les modèles de performance énergétique aux plates-formes hétérogènes dans lesquelles chaque noeud peut avoir des spécifications différentes telles que la puissance de calcul, la consommation d'énergie, différentes fréquences de fonctionnement ou encore des latences et des bandes passantes réseaux différentes. L'algorithme d'optimisation de la fréquence CPU a également été modifié en fonction de l'hétérogénéité de la plate-forme. Troisièmement, les modèles et l'algorithme d'optimisation de la fréquence CPU ont été complètement repensés pour prendre en considération les spécificités des algorithmes itératifs asynchrones.Tous ces modèles et algorithmes ont été appliqués sur des applications parallèles utilisant la bibliothèque MPI et ont été exécutés avec le simulateur Simgrid ou sur la plate-forme Grid'5000. Les expériences ont montré que les algorithmes proposés sont plus efficaces que les méthodes existantes. Ils n’introduisent qu’un faible surcoût et ne nécessitent pas de profilage au préalable car ils sont exécutés au cours du déroulement de l’application. / In recent years, green computing has become an important topic in the supercomputing research domain. However, the computing platforms are still consuming more and more energy due to the increase in the number of nodes composing them. To minimize the operating costs of these platforms many techniques have been used. Dynamic voltage and frequency scaling (DVFS) is one of them. It can be used to reduce the power consumption of the CPU while computing, by lowering its frequency. However, lowering the frequency of a CPU may increase the execution time of the application running on that processor. Therefore, the frequency that gives the best trade-off between the energy consumption and the performance of an application must be selected.This thesis, presents the algorithms developed to optimize the energy consumption and theperformance of synchronous and asynchronous message passing applications with iterations runningover clusters or grids. The energy consumption and performance models for each type of parallelapplication predicts its execution time and energy consumption for any selected frequency accordingto the characteristics of both the application and the architecture executing this application.The contribution of this thesis can be divided into three parts: Firstly, optimizing the trade-offbetween the energy consumption and the performance of the message passing applications withsynchronous iterations running over homogeneous clusters. Secondly, adapting the energy andperformance models to heterogeneous platforms where each node can have different specificationssuch as computing power, energy consumption, available frequency gears or network’s latency andbandwidth. The frequency scaling algorithm was also modified to suit the heterogeneity of theplatform. Thirdly, the models and the frequency scaling algorithm were completely rethought to takeinto considerations the asynchronism in the communication and computation. All these models andalgorithms were applied to message passing applications with iterations and evaluated over eitherSimGrid simulator or Grid’5000 platform. The experiments showed that the proposed algorithms areefficient and outperform existing methods such as the energy and delay product. They also introducea small runtime overhead and work online without any training or profiling.

Thesis (M.Ing. (Mechanical Engineering))--North-West University, Potchefstroom Campus, 2008.