The transient analysis of large-scale systems is often difficult even when the systems belong to the simplest M/M/n type of queues. To address analytical difficulties, previous studies have been conducted under various asymptotic regimes by suitably accelerating parameters, thereby establishing some useful mathematical frameworks and giving insights into important characteristics and intuitions. However, some studies show significant limitations when used to approximate real service systems: (i) they are more relevant to steady-state analysis; (ii) they emphasize proofs of convergence results rather than numerical methods to obtain system performance; and (iii) they provide only one set of limit processes regardless of actual system size. Attempting to overcome the drawbacks of previous studies, this dissertation studies the transient analysis of large-scale service systems with time-dependent parameters. The research goal is to develop a methodology that provides accurate approximations based on a technique called uniform acceleration, utilizing the theory of strong approximations. We first investigate and discuss the possible inaccuracy of limit processes obtained from employing the technique. As a solution, we propose adjusted fluid and diffusion limits that are specifically designed to approximate large, finite-sized systems. We find that the adjusted limits significantly improve the quality of approximations and hold asymptotic exactness as well. Several numerical results provide evidence of the effectiveness of the adjusted limits. We study both a call center which is a canonical example of large-scale service systems and an emerging peer-based Internet multimedia service network known as P2P. Based on our findings, we introduce a possible extension to systems which show non-Markovian behavior that is unaddressed by the uniform acceleration technique. We incorporate the denseness of phase-type distributions into the derivation of limit processes. The proposed method offers great potential to accurately approximate performance measures of non-Markovian systems with less computational burden.

published_or_final_version / abstract / toc / Mechanical Engineering / Doctoral / Doctor of Philosophy

As the dimension and the complexity of large interconnected systems grow, so does the necessity for decentralized control. One of the interesting challenges in the field of decentralized control is the arbitrary pole placement using output feedback. The feasibility of this problem depends solely on the identification of the decentralized fixed modes. As a matter of fact, if the system is free of fixed modes, then by increasing the controller’s order, any arbitrary closed loop poles can always be assigned. Due to this fact, reducing the controller’s order constitutes another interesting challenge when dealing with decentralization. This research describes the decentralized pole placement of linear systems. It is assumed that the internal structure of the system is unknown. The only access to the system is from a number of control stations. The decentralized controller consists of output feedback controllers each built at a control station. The research can be divided into two parts. In the first part, conditions for fixed modes existence as well as realization and stability of the overall system under decentralization are established using polynomial matrix algebra. The second part deals with the solution of decentralized pole placement problem, in particular, finding a decentralized controller which assigns some set of desired poles. The solution strategy is to reduce the controller’s order as much as possible using mathematical programming techniques. The idea behind this method is to start with a low order controller and then attempt to shift the poles of the closed loop system to the desired poles. / Ph. D. / incomplete_metadata

Les fournisseurs de services de cloud computing, les réseaux sociaux et les entreprises de gestion des données ont assisté à une augmentation considérable du volume de données qu'ils reçoivent chaque jour. Toutes ces données créent des nouvelles opportunités pour étendre la connaissance humaine dans des domaines comme la santé, l'urbanisme et le comportement humain et permettent d'améliorer les services offerts comme la recherche, la recommandation, et bien d'autres. Ce n'est pas par accident que plusieurs universitaires mais aussi les médias publics se référent à notre époque comme l'époque “Big Data”. Mais ces énormes opportunités ne peuvent être exploitées que grâce à de meilleurs systèmes de gestion de données. D'une part, ces derniers doivent accueillir en toute sécurité ce volume énorme de données et, d'autre part, être capable de les restituer rapidement afin que les applications puissent bénéficier de leur traite- ment. Ce document se concentre sur ces deux défis relatifs aux “Big Data”. Dans notre étude, nous nous concentrons sur le stockage de sauvegarde (i) comme un moyen de protéger les données contre un certain nombre de facteurs qui peuvent les rendre indisponibles et (ii) sur le placement des données sur des systèmes de stockage répartis géographiquement, afin que les temps de latence perçue par l'utilisateur soient minimisés tout en utilisant les ressources de stockage et du réseau efficacement. Tout au long de notre étude, les données sont placées au centre de nos choix de conception dont nous essayons de tirer parti des propriétés de contenu à la fois pour le placement et le stockage efficace. / Cloud service providers, social networks and data-management companies are witnessing a tremendous increase in the amount of data they receive every day. All this data creates new opportunities to expand human knowledge in fields like healthcare and human behavior and improve offered services like search, recommendation, and many others. It is not by accident that many academics but also public media refer to our era as the “Big Data” era. But these huge opportunities come with the requirement for better data management systems that, on one hand, can safely accommodate this huge and constantly increasing volume of data and, on the other, serve them in a timely and useful manner so that applications can benefit from processing them. This document focuses on the above two challenges that come with “Big Data”. In more detail, we study (i) backup storage systems as a means to safeguard data against a number of factors that may render them unavailable and (ii) data placement strategies on geographically distributed storage systems, with the goal to reduce the user perceived latencies and the network and storage resources are efficiently utilized. Throughout our study, data are placed in the centre of our design choices as we try to leverage content properties for both placement and efficient storage.

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1981. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. / Bibliography: leaves 400-410. / by Jose Ignacio Perez Arriaga. / Ph.D.

In recent years, Affective Computing has arisen as a prolific interdisciplinary field for engineering systems that integrate human affections. While human-computer relationships have long revolved around cognitive interactions, it is becoming increasingly important to account for human affect, or feelings or emotions, to avert user experience frustration, provide disability services, predict virality of social media content, etc. In this thesis, we specifically focus on Affective Computing as it applies to large-scale visual multimedia, and in particular, still images, animated image sequences and video streams, above and beyond the traditional approaches of face expression and gesture recognition. By taking a principled psychology-grounded approach, we seek to paint a more holistic and colorful view of computational affect in the context of visual multimedia. For example, should emotions like 'surprise' and `fear' be assumed to be orthogonal output dimensions? Or does a 'positive' image in one culture's view elicit the same feelings of positivity in another culture? We study affect frameworks and ontologies to define, organize and develop machine learning models with such questions in mind to automatically detect affective visual concepts. In the push for what we call "Big Affective Computing," we focus on two dimensions of scale for affect -- scaling up and scaling out -- which we propose are both imperative if we are to scale the Affective Computing problem successfully. Intuitively, simply increasing the number of data points corresponds to "scaling up". However, less intuitive, is when problems like Affective Computing "scale out," or diversify. We show that this latter dimension of introducing data variety, alongside the former of introducing data volume, can yield particular insights since human affections naturally depart from traditional Machine Learning and Computer Vision problems where there is an objectively truthful target. While no one might debate a picture of a 'dog' should be tagged as a 'dog,' but not all may agree that it looks 'ugly'. We present extensive discussions on why scaling out is critical and how it can be accomplished while in the context of large-volume visual data. At a high-level, the main contributions of this thesis include: Multiplicity of Affect Oracles: Prior to the work in this thesis, little consideration has been paid to the affective label generating mechanism when learning functional mappings between inputs and labels. Throughout this thesis but first in Chapter 2, starting in Section 2.1.2, we make a case for a conceptual partitioning of the affect oracle governing the label generation process in Affective Computing problems resulting a multiplicity of oracles, whereas prior works assumed there was a single universal oracle. In Chapter 3, the differences between intended versus expressed versus induced versus perceived emotion are discussed, where we argue that perceived emotion is particularly well-suited for scaling up because it reduces the label variance due to its more objective nature compared to other affect states. And in Chapter 4 and 5, a division of the affect oracle along cultural lines with manifestations along both language and geography is explored. We accomplish all this without sacrificing the 'scale up' dimension, and tackle significantly larger volume problems than prior comparable visual affective computing research. Content-driven Visual Affect Detection: Traditionally, in most Affective Computing work, prediction tasks use psycho-physiological signals from subjects viewing the stimuli of interest, e.g., a video advertisement, as the system inputs. In essence, this means that the machine learns to label a proxy signal rather than the stimuli itself. In this thesis, with the rise of strong Computer Vision and Multimedia techniques, we focus on the learning to label the stimuli directly without a human subject provided biometric proxy signal (except in the unique circumstances of Chapter 7). This shift toward learning from the stimuli directly is important because it allows us to scale up with much greater ease given that biometric measurement acquisition is both low-throughput and somewhat invasive while stimuli are often readily available. In addition, moving toward learning directly from the stimuli will allow researchers to precisely determine which low-level features in the stimuli are actually coupled with affect states, e.g., which set of frames caused viewer discomfort rather a broad sense that a video was discomforting. In Part I of this thesis, we illustrate an emotion prediction task with a psychology-grounded affect representation. In particular, in Chapter 3, we develop a prediction task over semantic emotional classes, e.g., 'sad,' 'happy' and 'angry,' using animated image sequences given annotations from over 2.5 million users. Subsequently, in Part II, we develop visual sentiment and adjective-based semantics models from million-scale digital imagery mined from a social multimedia platform. Mid-level Representations for Visual Affect: While discrete semantic emotions and sentiment are classical representations of affect with decades of psychology grounding, the interdisciplinary nature of Affective Computing, now only about two decades old, allows for new avenues of representation. Mid-level representations have been proposed in numerous Computer Vision and Multimedia problems as an intermediary, and often more computable, step toward bridging the semantic gap between low-level system inputs and high-level label semantic abstractions. In Part II, inspired by this work, we adapt it for vision-based Affective Computing and adopt a semantic construct called adjective-noun pairs. Specifically, in Chapter 4, we explore the use of such adjective-noun pairs in the context of a social multimedia platform and develop a multilingual visual sentiment ontology with over 15,000 affective mid-level visual concepts across 12 languages associated with over 7.3 million images and representations from over 235 countries, resulting in the largest affective digital image corpus in both depth and breadth to date. In Chapter 5, we develop computational methods to predict such adjective-noun pairs and also explore their usefulness in traditional sentiment analysis but with a previously unexplored cross-lingual perspective. And in Chapter 6, we propose a new learning setting called 'cross-residual learning' building off recent successes in deep neural networks, and specifically, in residual learning; we show that cross-residual learning can be used effectively to jointly learn across even multiple related tasks in object detection (noun), more traditional affect modeling (adjectives), and affective mid-level representations (adjective-noun pairs), giving us a framework for better grounding the adjective-noun pair bridge in both vision and affect simultaneously.

A Large Scale Robot Colony (LSRC) is a complex artifact comprising of a significant population of both mobile and static robots. LSRC research is in its literary infancy and it is therefore necessary to rely upon external fields for the appropriate framework, Multi Agent Systems (MAS) and Large Scale Systems (LSS). At the intersection of MAS, LSS and LSRC exist near identical issues, problems and solutions. If attention is paid to coherence then solution portability is possible. The issue of Self-Reliability is poorly addressed by the MAS research field. Disparity between the real world and simulation is another area of concern. Despite these deficiencies, MAS and LSS are perceived as the most appropriate frameworks. MAS research focuses on three prime areas, cognitive science, management and interaction. LSRC is focused on Self-Sustainability, Self-Management and Self-Organization. While LSS research was not primarily intended for populations of mobile robots, it does address key issues of LSRC, such as effective sustainability and management. Implementation of LSRC that is based upon the optimal solution for any one or two of the three aspects will be inferior to a coherent solution based upon all three. LSRC’s are complex organizations with significant populations of both static and mobile robots. The increase in population size and the requirement to address the issue of Self-Reliance give rise to new issues. It is no longer sufficient to speak only in terms of robot intelligence, architecture, interaction or team behaviour, even though these are still valid topics. Issues such as population sustainability and management have greater significance within LSRC. As the size of a robot populations increases, minor uneconomical decisions and actions inhibit the performance of the population. Interaction must be made economical within the context of the LSRC. Sustainability of the population becomes significant as it enables stable performance and extended operational lifespan. Management becomes significant as a mechanism to direct the population so as to achieve near optimal performance. The Self-Sustainability, Self-Management and Self-Organization of LSRC are vastly more complex than in team robotics. Performance of the overall population becomes more significant than individual or team achievement. This thesis is a presentation of the Cooperative Autonomous Robot Colony (CARC) architecture. The CARC architecture is novel in that it offers a coherent baseline solution to the issue of mobile robot Self-Reliance. This research uses decomposition as a mechanism to reduce problem complexity. Self-Reliance is decomposed into Self-Sustainability, Self-Management, and Self-Organization. A solution to the issue of Self-Reliance will comprise of conflicting sub-solutions. A product of this research is a set of guidelines that manages the conflict of sub-solutions and maintains a coherent solution. In addressing the issue of Self-Reliance, it became apparent that Economies of Scale, played an important role. The effects of Economies of Scale directed the research towards LSRC’s. LSRC’s demonstrated improved efficiency and greater capability to achieve the requirements of Self-Reliance. LSRC’s implemented with the CARC architecture would extend human capability, enabling large scale operations to be performed in an economical manner, within real world and real time environments, including those of a remote and hostile nature. The theory and architecture are supported using published literature, experiments, observations and mathematical projections. Contributions of this work are focused upon the three pillars of Self-Reliance addressed by CARC: Self-Sustainability, Self-Management and Self-Organization. The chapter on Self-Sustainability explains and justifies the relevance of this issue, what it is, why it is important and how it can be achieved. Self-Sustainability enables robots to continue to operate beyond disabling events by addressing failure and routine maintenance. Mathematical projections are used to compare populations of non-sustained and sustained robots. Computer modeling experiments are used to demonstrate the feasibility of Self-Sustainability, including extended operational life, the maintenance of optimal work flow and graceful physical degradation (GPD). A detailed explanation is presented of Sustainability Functions, Colony Sites, Static Robot Roles, Static Robot Failure Options, and Polymorphism. The chapter on Self-Management explores LSS research as a mechanism to exert influence over a LSRC. An experimental reactive management strategy is demonstrated. This strategy while limited does indicate promising potential directions for future research including the Man in the Loop (MITL) strategy highly desired by NASA JPL for off world command and control of a significant robot colony (Huntsberger, et. al., 2000). Experiments on Communication evaluate both Broadcast Conveyance (BC) and Message Passing Conveyance (MPC). These experiments demonstrate the potential of Message Passing as a low cost system for LSRC communication. Analysis of Metrics indicates that a Performance Based Feedback Method (PBFM) and a Task Achievement Method (TAM) are both necessary and sufficient to monitor a LSRC. The chapter on Self-Organization describes a number of experiments, algorithms and protocols on Reasoning Robotics, a minor variant of Reactive Robotics. Reasoning Robotics utilizes an Event Driven Architecture (EDA) rather than a Stimulus Driven Architecture (SDA) common to Reactive Robotics. Enhanced robot performance is demonstrated by a combination of EDA and environmental modification enabling stigmergy. These experiments cover Intersection Navigation with contingency for Multilane Intersections, a Radio Packet Controller (RPC) algorithm, Active and Passive Beacons including a communication protocol, mobile robot navigation using Migration Decision Functions (MDF’s), including MDF positional errors. The central issue addressed by this thesis is the production of Self-Reliance guidelines for LSRC’s. Self-Reliance is perceived as a critical issue in advancing the useful and productive applications for LSRC’s. LSRC’s are complex with many issues in related fields of MAS and LSS. Decomposition of Self-Reliance into Self-Sustainability, Self-Management and Self-Organization were used to aid in problem understanding. It was found that Self-Sustainability extends the operational life of individual robots and the LSRC. Self-Management enables the exertion of human influence over the LSRC, such that the ratio of humans to robots is reduced but not eliminated. Self-Organization achieves and enhances performance through a routine and reliable LSRC environment. The product of this research was the novel CARC architecture, which consists of a set of Self-Reliance guidelines and algorithms. The Self-Reliance guidelines manage conflict between optimal solutions and provide a framework for LSRC design. This research was supported by literature, experiments, observations and mathematical projections.

by Michael Athans and Sanjoy K. Mitter. / Research supported by Air Force Office of Scientific Research (AFSC) Research Grant AF-AFOSR 72-2273. Report for 1975/76 distributed through Industrial Liaison Program.

by Michael Athans. / Caption title. "This report summarizes the developments to date (November 1 1978) ... ." / Travel support provided in part by National Science Foundation Grant NSF/ENG77-07777

M. Coderch ... [et al.]. / Bibliography: leaf 6. / "September, 1981." / Supported in part by the DOE under Grant ET-76-C-01-2295