Dissertations / Theses on the topic 'Sequential decision processes'

Ce travail contribue à la théorie de la décision possibiliste et plus précisément à la prise de décision séquentielle dans le cadre de la théorie des possibilités, à la fois au niveau théorique et pratique. Bien qu'attrayante pour sa capacité à résoudre les problèmes de décision qualitatifs, la théorie de la décision possibiliste souffre d'un inconvénient important : les critères d'utilité qualitatives possibilistes comparent les actions avec les opérateurs min et max, ce qui entraîne un effet de noyade. Pour surmonter ce manque de pouvoir décisionnel, plusieurs raffinements ont été proposés dans la littérature. Les raffinements lexicographiques sont particulièrement intéressants puisqu'ils permettent de bénéficier de l'arrière-plan de l'utilité espérée, tout en restant "qualitatifs". Cependant, ces raffinements ne sont définis que pour les problèmes de décision non séquentiels. Dans cette thèse, nous présentons des résultats sur l'extension des raffinements lexicographiques aux problèmes de décision séquentiels, en particulier aux Arbres de Décision et aux Processus Décisionnels de Markov possibilistes. Cela aboutit à des nouveaux algorithmes de planification plus "décisifs" que leurs contreparties possibilistes. Dans un premier temps, nous présentons des relations de préférence lexicographiques optimistes et pessimistes entre les politiques avec et sans utilités intermédiaires, qui raffinent respectivement les utilités possibilistes optimistes et pessimistes. Nous prouvons que les critères proposés satisfont le principe de l'efficacité de Pareto ainsi que la propriété de monotonie stricte. Cette dernière garantit la possibilité d'application d'un algorithme de programmation dynamique pour calculer des politiques optimales. Nous étudions tout d'abord l'optimisation lexicographique des politiques dans les Arbres de Décision possibilistes et les Processus Décisionnels de Markov à horizon fini. Nous fournissons des adaptations de l'algorithme de programmation dynamique qui calculent une politique optimale en temps polynomial. Ces algorithmes sont basés sur la comparaison lexicographique des matrices de trajectoires associées aux sous-politiques. Ce travail algorithmique est complété par une étude expérimentale qui montre la faisabilité et l'intérêt de l'approche proposée. Ensuite, nous prouvons que les critères lexicographiques bénéficient toujours d'une fondation en termes d'utilité espérée, et qu'ils peuvent être capturés par des utilités espérées infinitésimales. La dernière partie de notre travail est consacrée à l'optimisation des politiques dans les Processus Décisionnels de Markov (éventuellement infinis) stationnaires. Nous proposons un algorithme d'itération de la valeur pour le calcul des politiques optimales lexicographiques. De plus, nous étendons ces résultats au cas de l'horizon infini. La taille des matrices augmentant exponentiellement (ce qui est particulièrement problématique dans le cas de l'horizon infini), nous proposons un algorithme d'approximation qui se limite à la partie la plus intéressante de chaque matrice de trajectoires, à savoir les premières lignes et colonnes. Enfin, nous rapportons des résultats expérimentaux qui prouvent l'efficacité des algorithmes basés sur la troncation des matrices
This work contributes to possibilistic decision theory and more specifically to sequential decision-making under possibilistic uncertainty, at both the theoretical and practical levels. Even though appealing for its ability to handle qualitative decision problems, possibilisitic decision theory suffers from an important drawback: qualitative possibilistic utility criteria compare acts through min and max operators, which leads to a drowning effect. To overcome this lack of decision power, several refinements have been proposed in the literature. Lexicographic refinements are particularly appealing since they allow to benefit from the expected utility background, while remaining "qualitative". However, these refinements are defined for the non-sequential decision problems only. In this thesis, we present results on the extension of the lexicographic preference relations to sequential decision problems, in particular, to possibilistic Decision trees and Markov Decision Processes. This leads to new planning algorithms that are more "decisive" than their original possibilistic counterparts. We first present optimistic and pessimistic lexicographic preference relations between policies with and without intermediate utilities that refine the optimistic and pessimistic qualitative utilities respectively. We prove that these new proposed criteria satisfy the principle of Pareto efficiency as well as the property of strict monotonicity. This latter guarantees that dynamic programming algorithm can be used for calculating lexicographic optimal policies. Considering the problem of policy optimization in possibilistic decision trees and finite-horizon Markov decision processes, we provide adaptations of dynamic programming algorithm that calculate lexicographic optimal policy in polynomial time. These algorithms are based on the lexicographic comparison of the matrices of trajectories associated to the sub-policies. This algorithmic work is completed with an experimental study that shows the feasibility and the interest of the proposed approach. Then we prove that the lexicographic criteria still benefit from an Expected Utility grounding, and can be represented by infinitesimal expected utilities. The last part of our work is devoted to policy optimization in (possibly infinite) stationary Markov Decision Processes. We propose a value iteration algorithm for the computation of lexicographic optimal policies. We extend these results to the infinite-horizon case. Since the size of the matrices increases exponentially (which is especially problematic in the infinite-horizon case), we thus propose an approximation algorithm which keeps the most interesting part of each matrix of trajectories, namely the first lines and columns. Finally, we reports experimental results that show the effectiveness of the algorithms based on the cutting of the matrices

Algorithmically designed reward functions can influence groups of learning agents toward measurable desired sequential joint behaviours. Influencing learning agents toward desirable behaviours is non-trivial due to the difficulties of assigning credit for global success to the deserving agents and of inducing coordination. Quantifying joint behaviours lets us identify global success by ranking some behaviours as more desirable than others. We propose a real-valued metric for turn-taking, demonstrating how to measure one sequential joint behaviour. We describe how to identify the presence of turn-taking in simulation results and we calculate the quantity of turn-taking that could be observed between independent random agents. We demonstrate our turn-taking metric by reinterpreting previous work on turn-taking in emergent communication and by analysing a recorded human conversation. Given a metric, we can explore the space of reward functions and identify those reward functions that result in global success in groups of learning agents. We describe 'medium access games' as a model for human and machine communication and we present simulation results for an extensive range of reward functions for pairs of Q-learning agents. We use the Nash equilibria of medium access games to develop predictors for determining which reward functions result in turn-taking. Having demonstrated the predictive power of Nash equilibria for turn-taking in medium access games, we focus on synthesis of reward functions for stochastic games that result in arbitrary desirable Nash equilibria. Our method constructs a reward function such that a particular joint behaviour is the unique Nash equilibrium of a stochastic game, provided that such a reward function exists. This method builds on techniques for designing rewards for Markov decision processes and for normal form games. We explain our reward design methods in detail and formally prove that they are correct.

Nous proposons une nouvelle approche pour l'apprentissage de représentation parcimonieuse, où le but est de limiter le nombre de caractéristiques sélectionnées \textbf{par donnée}, résultant en un modèle que nous appellerons \textit{Modèle de parcimonie locale pour la classification} --- \textit{Datum-Wise Sparse Classification} (DWSC) en anglais. Notre approche autorise le fait que les caractéristiques utilisées lors de la classification peuvent être différentes d'une donnée à une autre: une donnée facile à classifier le sera ainsi en ne considérant que quelques caractéristiques, tandis que plus de caractéristiques seront utilisées pour les données plus complexes. Au contraire des approches traditionnelles de régularisation qui essaient de trouver un équilibre entre performance et parcimonie au niveau de l'ensemble du jeu de données, notre motivation est de trouver cet équilibre au niveau des données individuelles, autorisant une parcimonie moyenne plus élevée, pour une performance équivalente. Ce type de parcimonie est intéressant pour plusieurs raisons~: premièrement, nous partons du principe que les explications les plus simples sont toujours préférables~; deuxièmement, pour la compréhension des données, une représentation parcimonieuse par donnée fournit une information par rapport à la structure sous-jacente de celles-ci~: typiquement, si un jeu de données provient de deux distributions disjointes, DWSC autorise le modèle à choisir automatiquement de ne prendre en compte que les caractéristiques de la distribution génératrice de chaque donnée considérée
This thesis introduces a body of work on sequential models for classification. These models allow for a more flexible and general approach to classification tasks. Many tasks ultimately require the classification of some object, but cannot be handled with a single atomic classification step. This is the case for tasks where information is either not immediately available upfront, or where the act of accessing different aspects of the object being classified may present various costs (due to time, computational power, monetary cost, etc.). The goal of this thesis is to introduce a new method, which we call datum-wise classification, that is able to handle these more complex classifications tasks by modelling them as sequential processes

Decision Support systems have been gaining in importance recently. Yet one of the bottlenecks of designing such systems lies in understanding how the user values different decision outcomes, or more simply what the user preferences are. Preference elicitation promises to remove the guess work of designing decision making agents by providing more formal methods for measuring the `goodness' of outcomes. This thesis aims to address some of the challenges of preference elicitation such as the high dimensionality of the underlying problem. The problem is formulated as a partially observable Markov decision process (POMDP) using a factored representation to take advantage of the structure inherent to preference elicitation problems. Moreover, simple preference knowledge on problem attributes are used to acquire more accurate preferences without increasing the burden on the user. Sparse terminal actions are defined to allow a flexible trade-off between speed and accuracy of the elicited preference function. Empirical simulations are used to validate the proposed methodology. The result is a framework that is flexible enough to be applied to a wide range of domains that addresses some of the challenges facing preference elicitation methods
Les systèmes d'aide à la décision ont gagné en importance récemment. Pourtant, un des problèmes importants liés au design de tels systèmes demeure: comprendre comment l'usager évalue les différents résultats, ou plus simplement, déterminer quelles sont ses préférences. L'extraction des préférences vise à éliminer certains aspects arbitraires du design d'agents de décision en offrant des méthodes plus formelles pour mesurer la qualité des résultats. Cette thèse tente de résoudre certains problèmes ayant trait à l'extraction des préférences, tel que celui de la haute dimensionnalité du problème sous-jacent. Le problème est formulé en tant que processus de décision markovien partiellement observable (POMDP), et utilise une représentation factorisée afin de profiter de la structure inhérente aux problèmes d'extraction des préférences. De plus, des connaissances simples quant aux caractéristiques de ces problèmes sont exploitées afin d'obtenir des préférences plus précises, sans pour autant augmenter la tâche de l'usager. Les actions terminales "sparse" sont définies de manière à permettre un compromis flexible entre vitesse et précision. Le résultat est un système assez flexible pour être appliqué à un grand nombre de domaines qui ont à faire face aux problèmes liés aux méthodes d'extraction des préférences.

My thesis is entitled "Contributions to Simulation-based High-dimensional Sequential Decision Making". The context of the thesis is about games, planning and Markov Decision Processes. An agent interacts with its environment by successively making decisions. The agent starts from an initial state until a final state in which the agent can not make decision anymore. At each timestep, the agent receives an observation of the state of the environment. From this observation and its knowledge, the agent makes a decision which modifies the state of the environment. Then, the agent receives a reward and a new observation. The goal is to maximize the sum of rewards obtained during a simulation from an initial state to a final state. The policy of the agent is the function which, from the history of observations, returns a decision. We work in a context where (i) the number of states is huge, (ii) reward carries little information, (iii) the probability to reach quickly a good final state is weak and (iv) prior knowledge is either nonexistent or hardly exploitable. Both applications described in this thesis present these constraints : the game of Go and a 3D simulator of the european project MASH (Massive Sets of Heuristics). In order to take a satisfying decision in this context, several solutions are brought : 1. Simulating with the compromise exploration/exploitation (MCTS) 2. Reducing the complexity by local solving (GoldenEye) 3. Building a policy which improves itself (RBGP) 4. Learning prior knowledge (CluVo+GMCTS) Monte-Carlo Tree Search (MCTS) is the state of the art for the game of Go. From a model of the environment, MCTS builds incrementally and asymetrically a tree of possible futures by performing Monte-Carlo simulations. The tree starts from the current observation of the agent. The agent switches between the exploration of the model and the exploitation of decisions which statistically give a good cumulative reward. We discuss 2 ways for improving MCTS : the parallelization and the addition of prior knowledge. The parallelization does not solve some weaknesses of MCTS; in particular some local problems remain challenges. We propose an algorithm (GoldenEye) which is composed of 2 parts : detection of a local problem and then its resolution. The algorithm of resolution reuses some concepts of MCTS and it solves difficult problems of a classical database. The addition of prior knowledge by hand is laborious and boring. We propose a method called Racing-based Genetic Programming (RBGP) in order to add automatically prior knowledge. The strong point is that RBGP rigorously validates the addition of a prior knowledge and RBGP can be used for building a policy (instead of only optimizing an algorithm). In some applications such as MASH, simulations are too expensive in time and there is no prior knowledge and no model of the environment; therefore Monte-Carlo Tree Search can not be used. So that MCTS becomes usable in this context, we propose a method for learning prior knowledge (CluVo). Then we use pieces of prior knowledge for improving the rapidity of learning of the agent and for building a model, too. We use from this model an adapted version of Monte-Carlo Tree Search (GMCTS). This method solves difficult problems of MASH and gives good results in an application to a word game.

Este trabalho é dedicado ao desenvolvimento teórico e algorítmico de processos de decisão markovianos com probabilidades imprecisas e representações relacionais. Na literatura, essa configuração tem sido importante dentro da área de planejamento em inteligência artificial, onde o uso de representações relacionais permite obter descrições compactas, e o emprego de probabilidades imprecisas resulta em formas mais gerais de incerteza. São três as principais contribuições deste trabalho. Primeiro, efetua-se uma discussão sobre os fundamentos de tomada de decisão sequencial com probabilidades imprecisas, em que evidencia-se alguns problemas ainda em aberto. Esses resultados afetam diretamente o (porém não restrito ao) modelo de interesse deste trabalho, os processos de decisão markovianos com probabilidades imprecisas. Segundo, propõe-se três algoritmos para processos de decisão markovianos com probabilidades imprecisas baseadas em programação (otimização) matemática. E terceiro, desenvolvem-se ideias propostas por Trevizan, Cozman e de Barros (2008) no uso de variantes do algoritmo Real-Time Dynamic Programming para resolução de problemas de planejamento probabilístico descritos através de versões estendidas da linguagem de descrição de domínios de planejamento (PPDDL).
This work is devoted to the theoretical and algorithmic development of Markov Decision Processes with Imprecise Probabilities and relational representations. In the literature, this configuration is important within artificial intelligence planning, where the use of relational representations allow compact representations and imprecise probabilities result in a more general form of uncertainty. There are three main contributions. First, we present a brief discussion of the foundations of decision making with imprecise probabilities, pointing towards key questions that remain unanswered. These results have direct influence upon the model discussed within this text, that is, Markov Decision Processes with Imprecise Probabilities. Second, we propose three algorithms for Markov Decision Processes with Imprecise Probabilities based on mathematical programming. And third, we develop ideas proposed by Trevizan, Cozman e de Barros (2008) on the use of variants of Real-Time Dynamic Programming to solve problems of probabilistic planning described by an extension of the Probabilistic Planning Domain Definition Language (PPDDL).

Dans cette thèse, nous avons étudié les problèmes de décisions séquentielles, avec comme application la gestion de stocks d'énergie. Traditionnellement, ces problèmes sont résolus par programmation dynamique stochastique. Mais la grande dimension, et la non convexité du problème, amènent à faire des simplifications sur le modèle pour pouvoir faire fonctionner ces méthodes.Nous avons donc étudié une méthode alternative, qui ne requiert pas de simplifications du modèle: Monte Carlo Tree Search (MCTS). Nous avons commencé par étendre le MCTS classique (qui s’applique aux domaines finis et déterministes) aux domaines continus et stochastiques. Pour cela, nous avons utilisé la méthode de Double Progressive Widening (DPW), qui permet de gérer le ratio entre largeur et profondeur de l’arbre, à l’aide de deux méta paramètres. Nous avons aussi proposé une heuristique nommée Blind Value (BV) pour améliorer la recherche de nouvelles actions, en utilisant l’information donnée par les simulations passées. D’autre part, nous avons étendu l’heuristique RAVE aux domaines continus. Enfin, nous avons proposé deux nouvelles méthodes pour faire remonter l’information dans l’arbre, qui ont beaucoup amélioré la vitesse de convergence sur deux cas tests.Une part importante de notre travail a été de proposer une façon de mêler MCTS avec des heuristiques rapides pré-existantes. C’est une idée particulièrement intéressante dans le cas de la gestion d’énergie, car ces problèmes sont pour le moment résolus de manière approchée. Nous avons montré comment utiliser Direct Policy Search (DPS) pour rechercher une politique par défaut efficace, qui est ensuite utilisée à l’intérieur de MCTS. Les résultats expérimentaux sont très encourageants.Nous avons aussi appliqué MCTS à des processus markoviens partiellement observables (POMDP), avec comme exemple le jeu de démineur. Dans ce cas, les algorithmes actuels ne sont pas optimaux, et notre approche l’est, en transformant le POMDP en MDP, par un changement de vecteur d’état.Enfin, nous avons utilisé MCTS dans un cadre de méta-bandit, pour résoudre des problèmes d’investissement. Le choix d’investissement est fait par des algorithmes de bandits à bras multiples, tandis que l’évaluation de chaque bras est faite par MCTS.Une des conclusions importantes de ces travaux est que MCTS en continu a besoin de très peu d’hypothèses (uniquement un modèle génératif du problème), converge vers l’optimum, et peut facilement améliorer des méthodes suboptimales existantes
In this thesis, we study sequential decision making problems, with a focus on the unit commitment problem. Traditionally solved by dynamic programming methods, this problem is still a challenge, due to its high dimension and to the sacrifices made on the accuracy of the model to apply state of the art methods. We investigate on the applicability of Monte Carlo Tree Search methods for this problem, and other problems that are single player, stochastic and continuous sequential decision making problems. We started by extending the traditional finite state MCTS to continuous domains, with a method called Double Progressive Widening (DPW). This method relies on two hyper parameters, and determines the ratio between width and depth in the nodes of the tree. We developed a heuristic called Blind Value (BV) to improve the exploration of new actions, using the information from past simulations. We also extended the RAVE heuristic to continuous domain. Finally, we proposed two new ways of backing up information through the tree, that improved the convergence speed considerably on two test cases.An important part of our work was to propose a way to mix MCTS with existing powerful heuristics, with the application to energy management in mind. We did so by proposing a framework that allows to learn a good default policy by Direct Policy Search (DPS), and to include it in MCTS. The experimental results are very positive.To extend the reach of MCTS, we showed how it could be used to solve Partially Observable Markovian Decision Processes, with an application to game of Mine Sweeper, for which no consistent method had been proposed before.Finally, we used MCTS in a meta-bandit framework to solve energy investment problems: the investment decision was handled by classical bandit algorithms, while the evaluation of each investment was done by MCTS.The most important take away is that continuous MCTS has almost no assumption (besides the need for a generative model), is consistent, and can easily improve existing suboptimal solvers by using a method similar to what we proposed with DPS

Les problèmes de décision séquentielle dans l’incertain requièrent qu’un agent prenne des décisions, les unes après les autres, en fonction de l’état de l’environnement dans lequel il se trouve. Dans la plupart des travaux, l’environnement dans lequel évolue l’agent est supposé stationnaire, c’est-à-dire qu’il n’évolue pas avec le temps. Toute- fois, l’hypothèse de stationnarité peut ne pas être vérifiée quand, par exemple, des évènements exogènes au problème interviennent. Dans cette thèse, nous nous intéressons à la prise de décision séquentielle dans des environnements non-stationnaires. Nous proposons un nouveau modèle appelé HS3MDP permettant de représenter les problèmes non-stationnaires dont les dynamiques évoluent parmi un ensemble fini de contextes. Afin de résoudre efficacement ces problèmes, nous adaptons l’algorithme POMCP aux HS3MDP. Dans le but d’apprendre les dynamiques des problèmes de cette classe, nous présentons RLCD avec SCD, une méthode utilisable sans connaître à priori le nombre de contextes. Nous explorons ensuite le domaine de l’argumentation où peu de travaux se sont intéressés à la décision séquentielle. Nous étudions deux types de problèmes : les débats stochastiques (APS ) et les problèmes de médiation face à des agents non-stationnaires (DMP). Nous présentons dans ce travail un modèle formalisant les APS et permettant de les transformer en MOMDP afin d’optimiser la séquence d’arguments d’un des agents du débat. Nous étendons cette modélisation aux DMP afin de permettre à un médiateur de répartir stratégiquement la parole dans un débat
In sequential decision-making problems under uncertainty, an agent makes decisions, one after another, considering the current state of the environment where she evolves. In most work, the environment the agent evolves in is assumed to be stationary, i.e., its dynamics do not change over time. However, the stationarity hypothesis can be invalid if, for instance, exogenous events can occur. In this document, we are interested in sequential decision-making in non-stationary environments. We propose a new model named HS3MDP, allowing us to represent non-stationary problems whose dynamics evolve among a finite set of contexts. In order to efficiently solve those problems, we adapt the POMCP algorithm to HS3MDPs. We also present RLCD with SCD, a new method to learn the dynamics of the environments, without knowing a priori the number of contexts. We then explore the field of argumentation problems, where few works consider sequential decision-making. We address two types of problems: stochastic debates (APS ) and mediation problems with non-stationary agents (DMP). In this work, we present a model formalizing APS and allowing us to transform them into an MOMDP in order to optimize the sequence of arguments of one agent in the debate. We then extend this model to DMPs to allow a mediator to strategically organize speak-turns in a debate

Aerospace systems are complex systems with interacting disciplines and technologies. As a result, the Decision Makers (DMs) dealing with such problems are involved in balancing the multiple, potentially conflicting attributes/criteria, transforming a large amount of customer supplied guidelines into a solidly defined set of requirement definitions. A variety of existing decision making methods are available to deal with this type of decision problems. The selection of a most appropriate decision making method is of particular importance since inappropriate decision methods are likely causes of misleading engineering design decisions. The research presented in this dissertation proposes a knowledge-based Multi-criteria Interactive Decision-making Advisor and Synthesis process (MIDAS), which can facilitate the selection of the most appropriate decision making method and which provides insight to the user for fulfilling different preferences. Once the most appropriate method is selected for the given problem, the advisor is also able to aid the DM to reach the final decision by following the rigorous problem solving procedure of the selected method. The MIDAS can also provide guidance as to the requirements needed to be fulfilled by a potentially new method for cases where no suitable method is found. In many other domains, such as complex system operation, decisions are often made in an environment with continuously changing situations. In addition, the decisions are usually completed based on uncertain or incomplete information due to the data availability and the environmental variation. This fact exacerbates the complexity of the decision making process because it results in the difficulties in perfectly and deterministically reasoning about the effects of the decisions and thus make it hard in determining the further decisions. In order to make proper decision and increase the system’s effectiveness, an advanced decision strategy is needed to capture the system’s dynamic characteristics and environmental uncertainty. An autonomous decision making advisor is developed to perform the real-time decision making under uncertainty. The development of the advisor system aims to solve a resource allocation problem to redistribute the limited resources to different agents under various scenarios and try to maximize the total rewards obtained from the resource allocation actions.

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.

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.

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.

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.

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.

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.

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.

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).

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

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.

Doctorat en sciences appliquées
info:eu-repo/semantics/nonPublished

Made available in DSpace on 2016-06-02T20:04:52Z (GMT). No. of bitstreams: 1 6306.pdf: 1062380 bytes, checksum: de31a51e2d0a59e52556156a08c37b41 (MD5) Previous issue date: 2014-05-30
Financiadora de Estudos e Projetos
In this work, we make a study of the Bayes sequential decision procedure applied to capture-recapture with fixed sample sizes, to estimate the size of a finite and closed population process. We present the statistical model, review the Bayesian decision theory, presenting the pure decision problem, the statistical decision problem and the sequential decision procedure. We illustrate the theoretical methods discussed using simulated data.
Neste trabalho, fazemos um estudo do procedimento de decisão sequencial de Bayes aplicado ao processo de captura-recaptura com tamanhos amostrais fixados, para estimação do tamanho de uma população finita e fechada. Apresentamos o modelo estatístico, revisamos a teoria de decisão bayesiana, apresentando o problema de decisão puro, o problema de decisão estatística e o procedimento de decisão sequencial. Ilustramos os métodos teóricos discutidos através de dados simulados.

Markov Decision Processes (MDP) is a common framework for modeling sequential decision-making problems, with applications ranging from inventory and supply chains to healthcare applications, autonomous driving and solving repeated games. Despite its modeling power, two fundamental challenges arise when using the MDP framework in real-worldapplications. First, the optimal decision rule may be highly dependent on the MDP parameters (e.g., transition rates across states and rewards for each state-action pair). When the parameters are miss-estimated, the resulting decision rule may be suboptimal when deployed in practice. Additionally, the optimal policies computed by state-of-the-art algorithms may not be interpretable and can be seen as a black-box. Among other reasons, this is problematic as the policy may not be understandable for the people that are supposed to operationalize it. In this thesis, we aim to broaden the applicability of the MDP framework by addressing the challenges of robustness and interpretability. In the first part of the thesis, we focus on robustness. We introduce a novel model for parameter uncertainty in Chapter 2, that is significantly less pessimistic than prior models of uncertainty while enabling the efficient computation of a robust policy. In Chapter 3, we consider a healthcare application, where we focus on proactively transferring patients of a hospital to the Intensive Care Unit, to ameliorate the overall survival rates and patients’ flow. In the second part of this thesis, we focus on interpretable algorithms, with an emphasis on the application to find novel triage protocols for ventilator allocations for COVID-19 patients. In Chapter 4, we introduce a simulation model to estimate the performance of the official New York State (NYS) triage protocol at various levels of shortages of ventilators, using a real data set of patients intubated during Spring 2020 because of COVID-19 complications. In Chapter 5, we introduce our algorithmic framework for computing interpretable (tree) policies and apply our methods to learn novel triage protocols.

Decision-theoretic systems, such as Markov Decision Processes (MDPs), are used for sequential decision-making under uncertainty. MDPs provide a generic framework that can be applied in various domains to compute optimal policies. This thesis presents techniques that offer explanations of optimal policies for MDPs and then refine decision theoretic models (Bayesian networks and MDPs) based on feedback from experts. Explaining policies for sequential decision-making problems is difficult due to the presence of stochastic effects, multiple possibly competing objectives and long-range effects of actions. However, explanations are needed to assist experts in validating that the policy is correct and to help users in developing trust in the choices recommended by the policy. A set of domain-independent templates to justify a policy recommendation is presented along with a process to identify the minimum possible number of templates that need to be populated to completely justify the policy. The rejection of an explanation by a domain expert indicates a deficiency in the model which led to the generation of the rejected policy. Techniques to refine the model parameters such that the optimal policy calculated using the refined parameters would conform with the expert feedback are presented in this thesis. The expert feedback is translated into constraints on the model parameters that are used during refinement. These constraints are non-convex for both Bayesian networks and MDPs. For Bayesian networks, the refinement approach is based on Gibbs sampling and stochastic hill climbing, and it learns a model that obeys expert constraints. For MDPs, the parameter space is partitioned such that alternating linear optimization can be applied to learn model parameters that lead to a policy in accordance with expert feedback. In practice, the state space of MDPs can often be very large, which can be an issue for real-world problems. Factored MDPs are often used to deal with this issue. In Factored MDPs, state variables represent the state space and dynamic Bayesian networks model the transition functions. This helps to avoid the exponential growth in the state space associated with large and complex problems. The approaches for explanation and refinement presented in this thesis are also extended for the factored case to demonstrate their use in real-world applications. The domains of course advising to undergraduate students, assisted hand-washing for people with dementia and diagnostics for manufacturing are used to present empirical evaluations.