Dans ce manuscrit de thèse, mes travaux portent sur la combinaison de méthodes pour la prise de décision séquentielle (plusieurs étapes de décision corrélées) dans des environnements complexes et incertains. Les méthodes mises au point sont essentiellement appliquées à des problèmes de gestion et de production d'électricité tels que l'optimisation de la gestion des stocks d'énergie dans un parc de production pour anticiper au mieux la fluctuation de la consommation des clients.Le manuscrit comporte 7 chapitres regroupés en 4 parties : Partie I, « Introduction générale », Partie II, « État de l'art », Partie III, « Contributions » et Partie IV, « Conclusion générale ».Le premier chapitre (Partie I) introduit le contexte et les motivations de mes travaux, à savoir la résolution de problèmes d' « Unit commitment », c'est à dire l'optimisation des stratégies de gestion de stocks d'énergie dans les parcs de production d'énergie. Les particularités et les difficultés sous-jacentes à ces problèmes sont décrites ainsi que le cadre de travail et les notations utilisées dans la suite du manuscrit.Le second chapitre (Partie II) dresse un état de l'art des méthodes les plus classiques utilisées pour la résolution de problèmes de prise de décision séquentielle dans des environnements incertains. Ce chapitre introduit des concepts nécessaires à la bonne compréhension des chapitres suivants (notamment le chapitre 4). Les méthodes de programmation dynamique classiques et les méthodes de recherche de politique directe y sont présentées.Le 3e chapitre (Partie II) prolonge le précédent en dressant un état de l'art des principales méthodes d’optimisation spécifiquement adaptées à la gestion des parcs de production d'énergie et à leurs subtilités. Ce chapitre présente entre autre les méthodes MPC (Model Predictive Control), SDP (Stochastic Dynamic Programming) et SDDP (Stochastic Dual Dynamic Programming) avec pour chacune leurs particularités, leurs avantages et leurs limites. Ce chapitre complète le précédent en introduisant d'autres concepts nécessaires à la bonne compréhension de la suite du manuscrit.Le 4e chapitre (Partie III) contient la principale contribution de ma thèse : un nouvel algorithme appelé « Direct Value Search » (DVS) créé pour résoudre des problèmes de prise de décision séquentielle de grande échelle en milieu incertain avec une application directe aux problèmes d' « Unit commitment ». Ce chapitre décrit en quoi ce nouvel algorithme dépasse les méthodes classiques présentées dans le 3e chapitre. Cet algorithme innove notamment par sa capacité à traiter des grands espaces d'actions contraints dans un cadre non-linéaire, avec un grand nombre de variables d'état et sans hypothèse particulière quant aux aléas du système optimisé (c'est à dire applicable sur des problèmes où les aléas ne sont pas nécessairement Markovien).Le 5e chapitre (Partie III) est consacré à un concept clé de DVS : l'optimisation bruitée. Ce chapitre expose une nouvelle borne théorique sur la vitesse de convergence des algorithmes d'optimisation appliqués à des problèmes bruités vérifiant certaines hypothèses données. Des méthodes de réduction de variance sont également étudiées et appliquées à DVS pour accélérer sensiblement sa vitesse de convergence.Le 6e chapitre (Partie III) décrit un résultat mathématique sur la vitesse de convergence linéaire d’un algorithme évolutionnaire appliqué à une famille de fonctions non quasi-convexes. Dans ce chapitres, il est prouvé que sous certaines hypothèses peu restrictives sur la famille de fonctions considérée, l'algorithme présenté atteint une vitesse de convergence linéaire.Le 7e chapitre (Partie IV) conclut ce manuscrit en résumant mes contributions et en dressant quelques pistes de recherche intéressantes à explorer. / This thesis is dedicated to sequential decision making (also known as multistage optimization) in uncertain complex environments. Studied algorithms are essentially applied to electricity production ("Unit Commitment" problems) and energy stock management (hydropower), in front of stochastic demand and water inflows. The manuscript is divided in 7 chapters and 4 parts: Part I, "General Introduction", Part II, "Background Review", Part III, "Contributions" and Part IV, "General Conclusion". This first chapter (Part I) introduces the context and motivation of our work, namely energy stock management. "Unit Commitment" (UC) problems are a classical example of "Sequential Decision Making" problem (SDM) applied to energy stock management. They are the central application of our work and in this chapter we explain main challenges arising with them (e.g. stochasticity, constraints, curse of dimensionality, ...). Classical frameworks for SDM problems are also introduced and common mistakes arising with them are be discussed. We also emphasize the consequences of these - too often neglected - mistakes and the importance of not underestimating their effects. Along this chapter, fundamental definitions commonly used with SDM problems are described. An overview of our main contributions concludes this first chapter. The second chapter (Part II) is a background review of the most classical algorithms used to solve SDM problems. Since the applications we try to solve are stochastic, we there focus on resolution methods for stochastic problems. We begin our study with classical Dynamic Programming methods to solve "Markov Decision Processes" (a special kind of SDM problems with Markovian random processes). We then introduce "Direct Policy Search", a widely used method in the Reinforcement Learning community. A distinction is be made between "Value Based" and "Policy Based" exploration methods. The third chapter (Part II) extends the previous one by covering the most classical algorithms used to solve UC's subtleties. It contains a state of the art of algorithms commonly used for energy stock management, mainly "Model Predictive Control", "Stochastic Dynamic Programming" and "Stochastic Dual Dynamic Programming". We briefly overview distinctive features and limitations of these methods. The fourth chapter (Part III) presents our main contribution: a new algorithm named "Direct Value Search" (DVS), designed to solve large scale unit commitment problems. We describe how it outperforms classical methods presented in the third chapter. We show that DVS is an "anytime" algorithm (users immediately get approximate results) which can handle large state spaces and large action spaces with non convexity constraints, and without assumption on the random process. Moreover, we explain how DVS can reduce modelling errors and can tackle challenges described in the first chapter, working on the "real" detailed problem without "cast" into a simplified model. Noisy optimisation is a key component of DVS algorithm; the fifth chapter (Part III) is dedicated to it. In this chapter, some theoretical convergence rate are studied and new convergence bounds are proved - under some assumptions and for given families of objective functions. Some variance reduction techniques aimed at improving the convergence rate of graybox noisy optimization problems are studied too in the last part of this chapter. Chapter sixth (Part III) is devoted to non-quasi-convex optimization. We prove that a variant of evolution strategy can reach a log-linear convergence rate with non-quasi-convex objective functions. Finally, the seventh chapter (Part IV) concludes and suggests some directions for future work.

In this thesis, a new approach for generating reaction time predictions for Bayesian cognitive models of sequential decision-making is proposed. The method is based on a Markov chain Monte Carlo algorithm that, by utilizing prior distributions and likelihood functions of possible action sequences, generates predictions about the time needed to choose one of these sequences. The plausibility of the reaction time predictions produced by this algorithm was investigated for simple exemplary distributions as well as for prior distributions and likelihood functions of a Bayesian model of habit learning. Simulations showed that the reaction time distributions generated by the Markov chain Monte Carlo sampler exhibit key characteristics of reaction time distributions typically observed in decision-making tasks. The introduced method can be easily applied to various Bayesian models for decision-making tasks with any number of choice alternatives. It thus provides the means to derive reaction time predictions for models where this has not been possible before. / In dieser Arbeit wird ein neuer Ansatz zum Generieren von Reaktionszeitvorhersagen für bayesianische Modelle sequenzieller Entscheidungsprozesse vorgestellt. Der Ansatz basiert auf einem Markov-Chain-Monte-Carlo-Algorithmus, der anhand von gegebenen A-priori-Verteilungen und Likelihood-Funktionen von möglichen Handlungssequenzen Vorhersagen über die Dauer einer Entscheidung für eine dieser Handlungssequenzen erstellt. Die Plausibilität der mit diesem Algorithmus generierten Reaktionszeitvorhersagen wurde für einfache Beispielverteilungen sowie für A-priori-Verteilungen und Likelihood-Funktionen eines bayesianischen Modells zur Beschreibung von Gewohnheitslernen untersucht. Simulationen zeigten, dass die vom Markov-Chain-Monte-Carlo-Sampler erzeugten Reaktionszeitverteilungen charakteristische Eigenschaften von typischen Reaktionszeitverteilungen im Kontext sequenzieller Entscheidungsprozesse aufweisen. Das Verfahren lässt sich problemlos auf verschiedene bayesianische Modelle für Entscheidungsparadigmen mit beliebig vielen Handlungsalternativen anwenden und eröffnet damit die Möglichkeit, Reaktionszeitvorhersagen für Modelle abzuleiten, für die dies bislang nicht möglich war.

本研究旨在探討兒童美語補教業消費者決策行為。由於過去相關文獻以量化 調查為主，缺少質化觀點；又因強調首次購買，忽略了時間脈絡及決策環境的影響；另外假定消費決策者為家長，也使得消費使用者-兒童的決策影響力未見著墨；因此本研究試圖以質性方法，探索兒美補教消費本質，並加入時間脈絡及親子互動觀察，試圖建構更完整之消費決策樣貌。 研究設計方面，先就過去文獻之研究發現進行整理，提出值得進一步探索之 問題，而後擬定訪談大綱，分別針對家長及兒童進行訪談，其中家長訪談為求深入以補充過去文獻不足，以一對一方式進行；兒童訪談部分，由於沒有文獻曾經探討兒童在補教消費的觀點，因此以焦點團體方式蒐集多元意見，進行探索研究。 本研究結果可歸納為七大點：(一)兒美補教首次消費接近特殊品之消費性 質，家長涉入深（二）家長對於兒美補教首次消費知覺風險高，決策勝任感不足（三）兒美補教消費為具嘗試性重購（Trial Repeat Purchase）之序貫決策（Sequential Decision）（四）反向代間影響於嘗試性重購時期漸增（五）同儕影響在兒美補教業存在且顯著，以社會支持、社會認同與社會遵從等三種方式影響（六）孩子與家長在選擇補習班決策準則與購買後滿意度準則均不同（七）兒美補習決策存在影響甚大的『詢問脈絡口碑』(Seeker-Initiated WOM)。 / This study intends to inquire consumer’s decision making process of children’s English learning program in a qualitative way. While there are many quantitative studies about this process, few are conducted qualitatively. revious studies are all focused only on the first purchase, neglecting their subsequent repeat purchase processes; and parents are assumed to be the decision makers, ignoring the role of the children in the process. This study thus purports to explore and provide a clearer depiction of the parent-child interaction in the repeat purchase context. After reviewing 27 prior studies, this researcher suggested research questions in this study. Data were collected through several interviews with parents and children respectively. Parent interviews were conducted through one-on-one basis, while children interviews through focus groups. Interviews were taped, transcribed, and analyzed qualitatively. The results of this study can be concluded into seven parts as follows. First of all, the essence of the first purchase decision in children’s English learning program is close to the essence of the purchase decision of special goods, and the consumers’ (parents) involvement is deep. Second, the risk conception of the first purchase is high and consumers are lack of confidence in first purchase in the decision making. Third, the process is a sequential decision with trial repeat purchase process. Fourth, the influence of reverse intergeneration is increasing in the process of trial repeat purchase process. Fifth, the peer-influence exists remarkably in three ways : social support, social identification and social conformity. Sixth, both decision and satisfying criteria are distinct from parents and children. Finally, the seeker-initiated WOM is a tremendous factor in the decision of children’s English learning program.

In ant societies, and, more in general, in insect societies, the activities of the individuals, as well as of the society as a whole, are not regulated by any explicit form of centralized control. On the other hand, adaptive and robust behaviors transcending the behavioral repertoire of the single individual can be easily observed at society level. These complex global behaviors are the result of self-organizing dynamics driven by local interactions and communications among a number of relatively simple individuals.<p><p>The simultaneous presence of these and other fascinating and unique characteristics have made ant societies an attractive and inspiring model for building new algorithms and new multi-agent systems. In the last decade, ant societies have been taken as a reference for an ever growing body of scientific work, mostly in the fields of robotics, operations research, and telecommunications.<p><p>Among the different works inspired by ant colonies, the Ant Colony Optimization metaheuristic (ACO) is probably the most successful and popular one. The ACO metaheuristic is a multi-agent framework for combinatorial optimization whose main components are: a set of ant-like agents, the use of memory and of stochastic decisions, and strategies of collective and distributed learning.<p><p>It finds its roots in the experimental observation of a specific foraging behavior of some ant colonies that, under appropriate conditions, are able to select the shortest path among few possible paths connecting their nest to a food site. The pheromone, a volatile chemical substance laid on the ground by the ants while walking and affecting in turn their moving decisions according to its local intensity, is the mediator of this behavior.<p><p>All the elements playing an essential role in the ant colony foraging behavior were understood, thoroughly reverse-engineered and put to work to solve problems of combinatorial optimization by Marco Dorigo and his co-workers at the beginning of the 1990's.<p><p>From that moment on it has been a flourishing of new combinatorial optimization algorithms designed after the first algorithms of Dorigo's et al. and of related scientific events.<p><p>In 1999 the ACO metaheuristic was defined by Dorigo, Di Caro and Gambardella with the purpose of providing a common framework for describing and analyzing all these algorithms inspired by the same ant colony behavior and by the same common process of reverse-engineering of this behavior. Therefore, the ACO metaheuristic was defined a posteriori, as the result of a synthesis effort effectuated on the study of the characteristics of all these ant-inspired algorithms and on the abstraction of their common traits.<p><p>The ACO's synthesis was also motivated by the usually good performance shown by the algorithms (e.g. for several important combinatorial problems like the quadratic assignment, vehicle routing and job shop scheduling, ACO implementations have outperformed state-of-the-art algorithms).<p><p>The definition and study of the ACO metaheuristic is one of the two fundamental goals of the thesis. The other one, strictly related to this former one, consists in the design, implementation, and testing of ACO instances for problems of adaptive routing in telecommunication networks.<p><p>This thesis is an in-depth journey through the ACO metaheuristic, during which we have (re)defined ACO and tried to get a clear understanding of its potentialities, limits, and relationships with other frameworks and with its biological background. The thesis takes into account all the developments that have followed the original 1999's definition, and provides a formal and comprehensive systematization of the subject, as well as an up-to-date and quite comprehensive review of current applications. We have also identified in dynamic problems in telecommunication networks the most appropriate domain of application for the ACO ideas. According to this understanding, in the most applicative part of the thesis we have focused on problems of adaptive routing in networks and we have developed and tested four new algorithms.<p><p>Adopting an original point of view with respect to the way ACO was firstly defined (but maintaining full conceptual and terminological consistency), ACO is here defined and mainly discussed in the terms of sequential decision processes and Monte Carlo sampling and learning.<p><p>More precisely, ACO is characterized as a policy search strategy aimed at learning the distributed parameters (called pheromone variables in accordance with the biological metaphor) of the stochastic decision policy which is used by so-called ant agents to generate solutions. Each ant represents in practice an independent sequential decision process aimed at constructing a possibly feasible solution for the optimization problem at hand by using only information local to the decision step.<p>Ants are repeatedly and concurrently generated in order to sample the solution set according to the current policy. The outcomes of the generated solutions are used to partially evaluate the current policy, spot the most promising search areas, and update the policy parameters in order to possibly focus the search in those promising areas while keeping a satisfactory level of overall exploration.<p><p>This way of looking at ACO has facilitated to disclose the strict relationships between ACO and other well-known frameworks, like dynamic programming, Markov and non-Markov decision processes, and reinforcement learning. In turn, this has favored reasoning on the general properties of ACO in terms of amount of complete state information which is used by the ACO's ants to take optimized decisions and to encode in pheromone variables memory of both the decisions that belonged to the sampled solutions and their quality.<p><p>The ACO's biological context of inspiration is fully acknowledged in the thesis. We report with extensive discussions on the shortest path behaviors of ant colonies and on the identification and analysis of the few nonlinear dynamics that are at the very core of self-organized behaviors in both the ants and other societal organizations. We discuss these dynamics in the general framework of stigmergic modeling, based on asynchronous environment-mediated communication protocols, and (pheromone) variables priming coordinated responses of a number of ``cheap' and concurrent agents.<p><p>The second half of the thesis is devoted to the study of the application of ACO to problems of online routing in telecommunication networks. This class of problems has been identified in the thesis as the most appropriate for the application of the multi-agent, distributed, and adaptive nature of the ACO architecture.<p><p>Four novel ACO algorithms for problems of adaptive routing in telecommunication networks are throughly described. The four algorithms cover a wide spectrum of possible types of network: two of them deliver best-effort traffic in wired IP networks, one is intended for quality-of-service (QoS) traffic in ATM networks, and the fourth is for best-effort traffic in mobile ad hoc networks.<p><p>The two algorithms for wired IP networks have been extensively tested by simulation studies and compared to state-of-the-art algorithms for a wide set of reference scenarios. The algorithm for mobile ad hoc networks is still under development, but quite extensive results and comparisons with a popular state-of-the-art algorithm are reported. No results are reported for the algorithm for QoS, which has not been fully tested. The observed experimental performance is excellent, especially for the case of wired IP networks: our algorithms always perform comparably or much better than the state-of-the-art competitors.<p><p>In the thesis we try to understand the rationale behind the brilliant performance obtained and the good level of popularity reached by our algorithms. More in general, we discuss the reasons of the general efficacy of the ACO approach for network routing problems compared to the characteristics of more classical approaches. Moving further, we also informally define Ant Colony Routing (ACR), a multi-agent framework explicitly integrating learning components into the ACO's design in order to define a general and in a sense futuristic architecture for autonomic network control.<p><p>Most of the material of the thesis comes from a re-elaboration of material co-authored and published in a number of books, journal papers, conference proceedings, and technical reports. The detailed list of references is provided in the Introduction.<p><p><p> / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

Cette thèse s'inscrit dans le domaine de la prise de décision séquentielle en environnement inconnu, et plus particulièrement dans le cadre des bandits manchots (multi-armed bandits, MAB), défini par Robbins et Lai dans les années 50. Depuis les années 2000, ce cadre a fait l'objet de nombreuses recherches théoriques et algorithmiques centrées sur le compromis entre l'exploration et l'exploitation : L'exploitation consiste à répéter le plus souvent possible les choix qui se sont avérés les meilleurs jusqu'à présent. L'exploration consiste à essayer des choix qui ont rarement été essayés, pour vérifier qu'on a bien identifié les meilleurs choix. Les applications des approches MAB vont du choix des traitements médicaux à la recommandation dans le contexte du commerce électronique, en passant par la recherche de politiques optimales de l'énergie. Les contributions présentées dans ce manuscrit s'intéressent au compromis exploration vs exploitation sous deux angles spécifiques. Le premier concerne la prise en compte du risque. Toute exploration dans un contexte inconnu peut en effet aboutir à des conséquences indésirables ; par exemple l'exploration des comportements d'un robot peut aboutir à des dommages pour le robot ou pour son environnement. Dans ce contexte, l'objectif est d'obtenir un compromis entre exploration, exploitation, et prise de risque (EER). Plusieurs algorithmes originaux sont proposés dans le cadre du compromis EER. Sous des hypothèses fortes, l'algorithme MIN offre des garanties de regret logarithmique, à l'état de l'art ; il offre également une grande robustesse, contrastant avec la forte sensibilité aux valeurs des hyper-paramètres de e.g. (Auer et al. 2002). L'algorithme MARAB s'intéresse à un critère inspiré de la littérature économique(Conditional Value at Risk), et montre d'excellentes performances empiriques comparées à (Sani et al. 2012), mais sans garanties théoriques. Enfin, l'algorithme MARABOUT modifie l'estimation du critère CVaR pour obtenir des garanties théoriques, tout en obtenant un bon comportement empirique. Le second axe de recherche concerne le bandit contextuel, où l'on dispose d'informations additionnelles relatives au contexte de la décision ; par exemple, les variables d'état du patient dans un contexte médical ou de l'utilisateur dans un contexte de recommandation. L'étude se focalise sur le choix entre bras qu'on a tirés précédemment un nombre de fois différent. Le choix repose en général sur la notion d'optimisme, comparant les bornes supérieures des intervalles de confiance associés aux bras considérés. Une autre approche appelée BESA, reposant sur le sous-échantillonnage des valeurs tirées pour les bras les plus visités, et permettant ainsi de se ramener au cas où tous les bras ont été tirés un même nombre de fois, a été proposée par (Baransi et al. 2014). / This thesis focuses on sequential decision making in unknown environment, and more particularly on the Multi-Armed Bandit (MAB) setting, defined by Lai and Robbins in the 50s. During the last decade, many theoretical and algorithmic studies have been aimed at cthe exploration vs exploitation tradeoff at the core of MABs, where Exploitation is biased toward the best options visited so far while Exploration is biased toward options rarely visited, to enforce the discovery of the the true best choices. MAB applications range from medicine (the elicitation of the best prescriptions) to e-commerce (recommendations, advertisements) and optimal policies (e.g., in the energy domain). The contributions presented in this dissertation tackle the exploration vs exploitation dilemma under two angles. The first contribution is centered on risk avoidance. Exploration in unknown environments often has adverse effects: for instance exploratory trajectories of a robot can entail physical damages for the robot or its environment. We thus define the exploration vs exploitation vs safety (EES) tradeoff, and propose three new algorithms addressing the EES dilemma. Firstly and under strong assumptions, the MIN algorithm provides a robust behavior with guarantees of logarithmic regret, matching the state of the art with a high robustness w.r.t. hyper-parameter setting (as opposed to, e.g. UCB (Auer 2002)). Secondly, the MARAB algorithm aims at optimizing the cumulative 'Conditional Value at Risk' (CVar) rewards, originated from the economics domain, with excellent empirical performances compared to (Sani et al. 2012), though without any theoretical guarantees. Finally, the MARABOUT algorithm modifies the CVar estimation and yields both theoretical guarantees and a good empirical behavior. The second contribution concerns the contextual bandit setting, where additional informations are provided to support the decision making, such as the user details in the ontent recommendation domain, or the patient history in the medical domain. The study focuses on how to make a choice between two arms with different numbers of samples. Traditionally, a confidence region is derived for each arm based on the associated samples, and the 'Optimism in front of the unknown' principle implements the choice of the arm with maximal upper confidence bound. An alternative, pioneered by (Baransi et al. 2014), and called BESA, proceeds instead by subsampling without replacement the larger sample set. In this framework, we designed a contextual bandit algorithm based on sub-sampling without replacement, relaxing the (unrealistic) assumption that all arm reward distributions rely on the same parameter. The CL-BESA algorithm yields both theoretical guarantees of logarithmic regret and good empirical behavior.

Dans cette thèse, nous proposons une méthode pour construire un outil d'aide à la décision pour le diagnostic de maladie rare. Nous cherchons à minimiser le nombre de tests médicaux nécessaires pour atteindre un état où l'incertitude concernant la maladie du patient est inférieure à un seuil prédéterminé. Ce faisant, nous tenons compte de la nécessité dans de nombreuses applications médicales, d'éviter autant que possible, tout diagnostic erroné. Pour résoudre cette tâche d'optimisation, nous étudions plusieurs algorithmes d'apprentissage par renforcement et les rendons opérationnels pour notre problème de très grande dimension. Pour cela nous décomposons le problème initial sous la forme de plusieurs sous-problèmes et montrons qu'il est possible de tirer partie des intersections entre ces sous-tâches pour accélérer l'apprentissage. Les stratégies apprises se révèlent bien plus performantes que des stratégies gloutonnes classiques. Nous présentons également une façon de combiner les connaissances d'experts, exprimées sous forme de probabilités conditionnelles, avec des données cliniques. Il s'agit d'un aspect crucial car la rareté des données pour les maladies rares empêche toute approche basée uniquement sur des données cliniques. Nous montrons, tant théoriquement qu'empiriquement, que l'estimateur que nous proposons est toujours plus performant que le meilleur des deux modèles (expert ou données) à une constante près. Enfin nous montrons qu'il est possible d'intégrer efficacement des raisonnements tenant compte du niveau de granularité des symptômes renseignés tout en restant dans le cadre probabiliste développé tout au long de ce travail. / In this thesis, we propose a method to build a decision support tool for the diagnosis of rare diseases. We aim to minimize the number of medical tests necessary to achieve a state where the uncertainty regarding the patient's disease is less than a predetermined threshold. In doing so, we take into account the need in many medical applications, to avoid as much as possible, any misdiagnosis. To solve this optimization task, we investigate several reinforcement learning algorithm and make them operable in our high-dimensional. To do this, we break down the initial problem into several sub-problems and show that it is possible to take advantage of the intersections between these sub-tasks to accelerate the learning phase. The strategies learned are much more effective than classic greedy strategies. We also present a way to combine expert knowledge, expressed as conditional probabilities, with clinical data. This is crucial because the scarcity of data in the field of rare diseases prevents any approach based solely on clinical data. We show, both empirically and theoretically, that our proposed estimator is always more efficient than the best of the two models (expert or data) within a constant. Finally, we show that it is possible to effectively integrate reasoning taking into account the level of granularity of the symptoms reported while remaining within the probabilistic framework developed throughout this work.

<p>Sequential decision-making is a fundamental class of problem that motivates algorithm designs of online machine learning and reinforcement learning. Arguably, the resulting online algorithms have supported modern online service industries for their data-driven real-time automated decision making. The applications span across different industries, including dynamic pricing (Marketing), recommendation (Advertising), and dosage finding (Clinical Trial). In this dissertation, we contribute fundamental statistical design advances for sequential decision-making algorithms, leaping progress in theory and application of online learning and sequential decision making under uncertainty including online sparse learning, finite-armed bandits, and high-dimensional online decision making. Our work locates at the intersection of decision-making algorithm designs, online statistical machine learning, and operations research, contributing new algorithms, theory, and insights to diverse fields including optimization, statistics, and machine learning.</p> <p><br></p> <p>In part I, we contribute a theoretical framework of continuous risk monitoring for regularized online statistical learning. Such theoretical framework is desirable for modern online service industries on monitoring deployed model's performance of online machine learning task. In the first project (Chapter 1), we develop continuous risk monitoring for the online Lasso procedure and provide an always-valid algorithm for high-dimensional dynamic pricing problems. In the second project (Chapter 2), we develop continuous risk monitoring for online matrix regression and provide new algorithms for rank-constrained online matrix completion problems. Such theoretical advances are due to our elegant interplay between non-asymptotic martingale concentration theory and regularized online statistical machine learning.</p> <p><br></p> <p>In part II, we contribute a bootstrap-based methodology for finite-armed bandit problems, termed Residual Bootstrap exploration. Such a method opens a possibility to design model-agnostic bandit algorithms without problem-adaptive optimism-engineering and instance-specific prior-tuning. In the first project (Chapter 3), we develop residual bootstrap exploration for multi-armed bandit algorithms and shows its easy generalizability to bandit problems with complex or ambiguous reward structure. In the second project (Chapter 4), we develop a theoretical framework for residual bootstrap exploration in linear bandit with fixed action set. Such methodology advances are due to our development of non-asymptotic theory for the bootstrap procedure.</p> <p><br></p> <p>In part III, we contribute application-driven insights on the exploration-exploitation dilemma for high-dimensional online decision-making problems. Such insights help practitioners to implement effective high-dimensional statistics methods to solve online decisionmaking problems. In the first project (Chapter 5), we develop a bandit sampling scheme for online batch high-dimensional decision making, a practical scenario in interactive marketing, and sequential clinical trials. In the second project (Chapter 6), we develop a bandit sampling scheme for federated online high-dimensional decision-making to maintain data decentralization and perform collaborated decisions. These new insights are due to our new bandit sampling design to address application-driven exploration-exploitation trade-offs effectively. </p>