Em Aprendizagem Colaborativa com Suporte Computacional (ACSC), o planejamento instrucional consiste em obter uma sequência de interações instrucionais que definem o conteúdo instrucional como a representação do que deve ser ensinado e da forma em que os participantes devem interagir, denominada informação de planejamento instrucional. O desenvolvimento, adaptação e personalização de unidades concisas de estudo compostas por recursos instrucionais e informação de planejamento instrucional, denominadas unidades de aprendizagem, envolve um processo de planejamento instrucional complexo que consome muito tempo e apresenta um conjunto de tarefas repetitivas a serem efetuadas pelos projetistas instrucionais. Neste trabalho, o planejamento instrucional em ACSC é modelado como um problema de planejamento hierárquico para dar suporte ao desenvolvimento, adaptação e personalização das unidades de aprendizagem de forma automática. A modelagem consiste na representação do domínio a ser ensinado, das caraterísticas dos estudantes e das estratégias de planejamento instrucional na linguagem do sistema JSHOP2ip, um sistema de planejamento hierárquico desenvolvido para dar solução aos problemas específicos de planejamento instrucional. Para avaliar a modelagem proposta, efetuamos o desenvolvimento de um gerador de cursos colaborativos como um serviço Web usando a modelagem proposta e o sistema JSHOP2ip, no qual foram avaliados o desempenho, a modelagem das estratégias e a saída do planejador. Além disso, para demonstrar a viabilidade do modelo proposto em situações reais, efetuamos o desenvolvimento de uma ferramenta de autoria de unidades de aprendizagem que emprega o gerador de cursos colaborativos. / In Computer Supported Collaborative Learning (CSCL), the goal of instructional design is to obtain a instructional interaction sequence that define the instructional content as a representation of what should be taught and the way in which participants must interact, called instructional planning information. The development, adaptation and personalization of basic units of study comprised of the instructional resources and instructional planning information, called units of learning, that involves a complex instructional planning process, time consuming and repetitive. In this work, the instructional design in CSCL is modeled as hierarchical planning problem to support the development, adaptation and personalization for units of learning. The modeling is the representation of the domain to be taught, the characteristics of students and instructional strategies in JSHOP2ip, an independent hierarchical planning system designed to solve problems of instructional design. To evaluate the proposed model, we developed a collaborative course generator as a Web service using the proposed model and JSHOP2ip system, upon which we evaluated the performance, modeling strategies and the output scheduler. Furthermore, to demonstrate the feasibility of the proposed model in real situations, we developed an authoring tool for units of learning employing the collaborative course generator

Ambientes Pervasivos são ambientes povoados por diversos dispositivos (sensores, atuadores, etc.) e aplicações de software (locais ou distribuídas) incorporadas nestes ambientes físicos de forma transparente para o usuário. Ambientes deste tipo devem ser capazes de interagir e satisfazer as requisições do usuário de forma autônoma e transparente. Nesse sentido, um dos maiores desafios de pesquisa em ambientes pervasivos é a de estabelecer mecanismos automáticos para compor, de forma dinâmica, funcionalidades que satisfaçam as requisições dos usuários. Nesse sentido, nós partimos da hipótese que mecanismos automáticos de interação entre ambientes e usuários podem ser abordados como um problema de composição automática de serviços em ambientes pervasivos. Portanto, nossa proposta é desenvolver um modelo referência, a partir do qual podem ser implementados sistemas que permitam ao ambiente pervasivo interagir com o usuário de forma natural, automática e dinâmica. Desta forma, o sistema de composição de serviços gerencia e automatiza o processo de resolução de requisições feitas pelo usuário (de forma implícita e/ou explicita) através das funcionalidades (serviços) disponíveis no ambiente ou através de novas funcionalidades criadas pelo processo de composição. Para tal fim, este trabalho apresenta um Modelo de Referência que permita projetar, implementar e avaliar sistemas de composição de serviços que gerenciem e automatizem o processo de interação em diversos tipos de ambientes pervasivos. Nossa proposta traz vantagens como: o baixo acoplamento e a interoperabilidade, isto porque é possível selecionar, integrar e reutilizar de forma eficiente e efetiva serviços heterogêneos provenientes de diversos tipos de dispositivos e/ou aplicações. Além disso, o modelo ontológico WSMO (Web Services Modelling Ontology) nos permite descrever semanticamente as capacidades dos serviços como também as informações contextuais presentes no ambiente, o que torna nosso sistema mais perto de um ambiente pervasivo real como o idealizado por Mark Weiser. / Pervasive environments are populated by several devices (sensors, actuators, etc.) and software applications (local or distributed) incorporated these physical environments transparently to the user. Environments of this type should be able to interact and process user requests autonomously and transparently. In this sense, one of the biggest research challenges in pervasive environments is to establish automatic mechanisms to compose dynamically, features that meet the user requirements. In this sense, we set the hypothesis that automatic mechanisms of interaction between users and environments can be addressed as a problem of automatic composition of services in pervasive environments. Therefore, our proposal is to develop a reference model, from which systems can be implemented to enable the pervasive environment interact with the user in a natural, automatic and dynamic. Thus, the system service composition management and automates the process of resolving requests made by the user (implicitly and / or explicitly) through the functionality (services) available in the environment or through new features created by the process of composition. To this end, this paper presents a Reference Model that allows to design, implement and evaluate systems of composition of services that manage and automate the interaction process in different types of pervasive environments. Our proposal brings benefits such as loose coupling and interoperability, because it is possible to select, integrate and reuse in an efficient and effective heterogeneous services from different types of devices and / or applications. Furthermore, the ontological model WSMO (Web Services Modeling Ontology) allows us to semantically describe the capabilities of the services as well as contextual information in the environment, which makes our system closer to a real pervasive environment as conceived by Mark Weiser.

O grande avanço das técnicas de Planejamento em Inteligência Artificial fez com que a Engenharia de Requisitos e a Engenharia do Conhecimento ganhassem extrema importância entre as disciplinas relacionadas a projeto de engenharia (Engineering Design). A especificação, modelagem e análise dos domínios de planejamento automático se tornam etapas fundamentais para melhor entender e classificar os domínios de planejamento, servindo também de guia na busca de soluções. Neste trabalho, é apresentada uma proposta de um ambiente integrado de modelagem e análise de domínios de planejamento, que leva em consideração o ciclo de vida de projeto, representado por uma ferramenta gráfica de modelagem que utiliza diferentes representações: a UML para modelar e analisar as características estáticas dos domínios; XML para armazenar, integrar, e exportar informação para outras linguagens (ex.: PDDL); as Redes de Petri para fazer a análise dinâmica; e a PDDL para testes com planejadores. / The great development in Artificial Intelligence Planning has emphasized the role of Requirements Engineering and Knowledge Engineering among the disciplines that contributes to Engineering Design. The modeling and specification of automated planning domains turn out to be fundamental tasks in order to understand and classify planning domains and guide the application of problem solving techniques. In this work, it is presented the proposed integrated environment for modeling and analyzing automated planning domains, which considered the life cycle of a project, represented by a tool that uses several language representations: UML to model and perform static analyses of planning environments; XML to hold, integrate, share and export information to other language representations (e.g. PDDL); Petri Nets, where dynamic analyses are made; and PDDL for testing models with planners.

Em Aprendizagem Colaborativa com Suporte Computacional (ACSC), o planejamento instrucional consiste em obter uma sequência de interações instrucionais que definem o conteúdo instrucional como a representação do que deve ser ensinado e da forma em que os participantes devem interagir, denominada informação de planejamento instrucional. O desenvolvimento, adaptação e personalização de unidades concisas de estudo compostas por recursos instrucionais e informação de planejamento instrucional, denominadas unidades de aprendizagem, envolve um processo de planejamento instrucional complexo que consome muito tempo e apresenta um conjunto de tarefas repetitivas a serem efetuadas pelos projetistas instrucionais. Neste trabalho, o planejamento instrucional em ACSC é modelado como um problema de planejamento hierárquico para dar suporte ao desenvolvimento, adaptação e personalização das unidades de aprendizagem de forma automática. A modelagem consiste na representação do domínio a ser ensinado, das caraterísticas dos estudantes e das estratégias de planejamento instrucional na linguagem do sistema JSHOP2ip, um sistema de planejamento hierárquico desenvolvido para dar solução aos problemas específicos de planejamento instrucional. Para avaliar a modelagem proposta, efetuamos o desenvolvimento de um gerador de cursos colaborativos como um serviço Web usando a modelagem proposta e o sistema JSHOP2ip, no qual foram avaliados o desempenho, a modelagem das estratégias e a saída do planejador. Além disso, para demonstrar a viabilidade do modelo proposto em situações reais, efetuamos o desenvolvimento de uma ferramenta de autoria de unidades de aprendizagem que emprega o gerador de cursos colaborativos. / In Computer Supported Collaborative Learning (CSCL), the goal of instructional design is to obtain a instructional interaction sequence that define the instructional content as a representation of what should be taught and the way in which participants must interact, called instructional planning information. The development, adaptation and personalization of basic units of study comprised of the instructional resources and instructional planning information, called units of learning, that involves a complex instructional planning process, time consuming and repetitive. In this work, the instructional design in CSCL is modeled as hierarchical planning problem to support the development, adaptation and personalization for units of learning. The modeling is the representation of the domain to be taught, the characteristics of students and instructional strategies in JSHOP2ip, an independent hierarchical planning system designed to solve problems of instructional design. To evaluate the proposed model, we developed a collaborative course generator as a Web service using the proposed model and JSHOP2ip system, upon which we evaluated the performance, modeling strategies and the output scheduler. Furthermore, to demonstrate the feasibility of the proposed model in real situations, we developed an authoring tool for units of learning employing the collaborative course generator

Cette thèse explore l'application de la planification HTN afin d'animer une section d'infanterie dans un simulateur informatique temps réel. Afin de produire des plans en ligne pour près de 40 soldats, on montre qu'il est possible d'optimiser le planificateur pour un domaine HTN en compilant les éléments de planifications en structures statiques et en procédures C++. On montre ensuite que la structure du problème se prête à une combinaison de la planification HTN avec la planification par abstraction, obtenue en modélisant des effets abstraits aux tâches composées. Sous certaines conditions, la recherche de solutions est alors accélérée en détectant les réseaux de tâches pour lesquels aucune solution n'est exécutable. Enfin, on montre que la structure du problème permet aussi de formuler des fonctions d'évaluation exploitables dans un algorithme de recherche heuristique non admissible, capable de retourner rapidement des solutions presque optimales. / This thesis explores the application of HTN planning to the animation of an infantry platoon in a real-time simulation software. In order to achieve online planning for nearly 40 soldiers, we show that it is possible to optimize the planner for one HTN domain with a compilation of planning elements into C++ static structures and procedures. Then, we demonstrate that the problem structure lends itself to a combination of HTN planning with abstraction planning, achieved with the modelisation of abstract effects for compound tasks. In some conditions, we can detect those task networks that never lead to any executable solution, and therefore improve the search. Eventually, we show that the problem structure enables to formulate evaluation functions that can be input into a non admissible heuristic search algorithm, and that near optimal solutions can be obtained within a short run-time.

No description available.

In radiation therapy treatment planning, recent works have used machine learning based on historically delivered plans to automate the process of producing clinically acceptable plans. Compared to traditional approaches such as repeated weighted-sum optimization or multicriteria optimization (MCO), automated planning methods have, in general, the benefits of low computational times and minimal user interaction, but on the other hand lack the flexibility associated with general-purpose frameworks such as MCO. Machine learning approaches can be especially sensitive to deviations in their dose prediction due to certain properties of the optimization functions usually used for dose mimicking and, moreover, suffer from the fact that there exists no general causality between prediction accuracy and optimized plan quality.In this thesis, we present a means of unifying ideas from machine learning planning methods with the well-established MCO framework. More precisely, given prior knowledge in the form of either a previously optimized plan or a set of historically delivered clinical plans, we are able to automatically generate Pareto optimal plans spanning a dose region corresponding to plans which are achievable as well as clinically acceptable. For the former case, this is achieved by introducing dose--volume constraints; for the latter case, this is achieved by fitting a weighted-data Gaussian mixture model on pre-defined dose statistics using the expectation--maximization algorithm, modifying it with exponential tilting and using specially developed optimization functions to take into account prediction uncertainties.Numerical results for conceptual demonstration are obtained for a prostate cancer case with treatment delivered by a volumetric-modulated arc therapy technique, where it is shown that the methods developed in the thesis are successful in automatically generating Pareto optimal plans of satisfactory quality and diversity, while excluding clinically irrelevant dose regions. For the case of using historical plans as prior knowledge, the computational times are significantly shorter than those typical of conventional MCO. / Inom strålterapiplanering har den senaste forskningen använt maskininlärning baserat på historiskt levererade planer för att automatisera den process i vilken kliniskt acceptabla planer produceras. Jämfört med traditionella angreppssätt, såsom upprepad optimering av en viktad målfunktion eller flermålsoptimering (MCO), har automatiska planeringsmetoder generellt sett fördelarna av lägre beräkningstider och minimal användarinteraktion, men saknar däremot flexibiliteten hos allmänna ramverk som exempelvis MCO. Maskininlärningsmetoder kan vara speciellt känsliga för avvikelser i dosprediktionssteget på grund av särskilda egenskaper hos de optimeringsfunktioner som vanligtvis används för att återskapa dosfördelningar, och lider dessutom av problemet att det inte finns något allmängiltigt orsakssamband mellan prediktionsnoggrannhet och kvalitet hos optimerad plan. I detta arbete presenterar vi ett sätt att förena idéer från maskininlärningsbaserade planeringsmetoder med det väletablerade MCO-ramverket. Mer precist kan vi, givet förkunskaper i form av antingen en tidigare optimerad plan eller en uppsättning av historiskt levererade kliniska planer, automatiskt generera Paretooptimala planer som täcker en dosregion motsvarande uppnåeliga såväl som kliniskt acceptabla planer. I det förra fallet görs detta genom att introducera dos--volym-bivillkor; i det senare fallet görs detta genom att anpassa en gaussisk blandningsmodell med viktade data med förväntning--maximering-algoritmen, modifiera den med exponentiell lutning och sedan använda speciellt utvecklade optimeringsfunktioner för att ta hänsyn till prediktionsosäkerheter.Numeriska resultat för konceptuell demonstration erhålls för ett fall av prostatacancer varvid behandlingen levererades med volymetriskt modulerad bågterapi, där det visas att metoderna utvecklade i detta arbete är framgångsrika i att automatiskt generera Paretooptimala planer med tillfredsställande kvalitet och variation medan kliniskt irrelevanta dosregioner utesluts. I fallet då historiska planer används som förkunskap är beräkningstiderna markant kortare än för konventionell MCO.

Nous nous positionnons dans un contexte d'informatique ambiante dans lequel il arrive que les besoins de l'utilisateur n'aient pas été prévus, notamment en situation exceptionnelle. Dans ce cas, il peut ne pas exister de système préconçu qui réponde exactement à ces besoins. Pour les satisfaire, il faut alors pouvoir composer les systèmes disponibles dans l'environnement, et le système composé doit permettre à l'utilisateur de faire des choix à l'exécution. Ainsi, l'utilisateur a la possibilité d'adapter l'exécution de la composition à son contexte. Cela signifie que la composition intègre des structures de contrôle de l'exécution, destinées à l'utilisateur : la composition est dite flexible. Dans cette thèse, nous proposons de répondre au problème de la composition flexible en contexte d'intelligence ambiante avec un planificateur produisant des plans flexibles. Dans un premier temps, nous proposons une modélisation de la planification flexible. Pour cela, nous définissons les opérateurs de séquence et d'alternative, utilisés pour caractériser les plans flexibles. Nous définissons deux autres opérateurs au moyen de la séquence et de l'alternative : l'entrelacement et l'itération. Nous nous référons à ce cadre théorique pour délimiter la flexibilité traitée par notre planificateur Lambda-Graphplan. L'originalité de Lambda-Graphplan est de produire des itérations en s'appuyant sur une approche par graphe de planification. Nous montrons notamment que Lambda-Graphplan est très performant avec les domaines se prêtant à la construction de structures itératives. / In a context of Ambient Intelligence, some of the user's needs might not be anticipated, e.g. when the user is in an unforeseen situation. In this case, there could exist no system that exactly meets their needs. By composing the available systems, the user could obtain a new system that satisfies their needs. In order to adapt the composition to the context, the composition must allow the user to make choices at runtime. So the composition includes control structures for the user: the composition is flexible. In this thesis, I deal with the problem of the flexible composition by automated planning. I propose a model of flexible planning. The sequence and the choice operators are defined and used to characterize flexible plans. Then, two other operators are derived from the sequence and the choice operators: the interleaving and the iteration operators. I refer to this framework to define the flexibility produced by my planner, Lambda-Graphplan, which is based on the planning graph. The originality of Lambda-Graphplan is to produce iterations. I show that Lambda-Graphplan is very efficient on domains that allow the construction of iterative structures.

Construction of radiotherapy treatments for cancer is a laborious and time consuming task. At the same time, when presented with a treatment plan, an oncologist can quickly judge whether or not it is suitable. This means that the problem of constructing these treatment plans is well suited for automation. This thesis investigates a novel way of automatic treatment planning. The treatment planning system this pipeline is constructed for provides dose mimicking functionality with probability density functions of dose–volume histograms (DVHs) and spatial dose as inputs. Therefore this will be the output of the pipeline. The input is historically treated patient scans, segmentations and spatial doses. The approach involves three modules which are individually replaceable with little to no impact on the remaining two modules. The modules are: an autoencoder as a feature extractor to concretise important features of a patient segmentation, a distance optimisation step to learn a distance in the previously constructed feature space and, finally, a probabilistic spatial dose estimation module using sparse pseudo-input Gaussian processes trained on voxel features. Although performance evaluation in terms of clinical plan quality was beyond the scope of this thesis, numerical results show that the proposed pipeline is successful in capturing salient features of patient geometry as well as predicting reasonable probability distributions for DVH and spatial dose. Its loosely connected nature also gives hope that some parts of the pipeline can be utilised in future work. / Skapandet av strålbehandlingsplaner för cancer är en tidskrävande uppgift. Samtidigt kan en onkolog snabbt fatta beslut om en given plan är acceptabel eller ej. Detta innebär att uppgiften att skapa strålplaner är väl lämpad för automatisering. Denna uppsats undersöker en ny metod för att automatiskt generera strålbehandlingsplaner. Planeringssystemet denna metod utvecklats för innehåller funktionalitet för dosrekonstruktion som accepterar sannolikhetsfördelningar för dos–volymhistogram (DVH) och dos som input. Därför kommer detta att vara utdatan för den konstruerade metoden. Metoden är uppbyggd av tre beståndsdelar som är individuellt utbytbara med liten eller ingen påverkan på de övriga delarna. Delarna är: ett sätt att konstruera en vektor av kännetecken av en patients segmentering, en distansoptimering för att skapa en distans i den tidigare konstruerade känneteckensrymden, och slutligen en skattning av sannolikhetsfördelningar med Gaussiska processer tränade på voxelkännetecken. Trots att utvärdering av prestandan i termer av klinisk plankvalitet var bortom räckvidden för detta projekt uppnåddes positiva resultat. De estimerade sannolikhetsfördelningarna uppvisar goda karaktärer för både DVHer och doser. Den löst sammankopplade strukturen av metoden gör det dessutom möjligt att delar av projektet kan användas i framtida arbeten.