Ce travail a pour objectif le développement d’une méthodologie de conception de procédés éco-efficaces, avec prise en compte simultanée de considérations écologiques et économiques lors de la phase de conception préliminaire de procédés chimiques. L’aspect environnemental est quantifié à travers l’utilisation d’un ensemble d’indicateurs selon les lignes directrices de concepts de développement durable. Le cadre conceptuel est basé sur une modélisation du procédé et de l’unité de production d’utilités, car l’impact environnemental d’un procédé est lié non seulement aux effluents directs du procédé, mais également à la consommation énergétique, l’effet du recyclage, de la conversion des matières premières … Dans ce but, le logiciel d’aide à la décision ARIANETM dédié à la gestion des utilités des unités de production (vapeur, électricité, eau …) et des effluents (CO2, SO2, NOx, etc..) a été couplé au modèle d’un procédé pour déterminer les besoins en énergies primaires et quantifier les émissions polluantes. Ces modèles ont ensuite été intégrés au sein d’une boucle d’optimisation multiobjectif, basée sur une variante d’un algorithme génétique multiobjectif de type NSGA-II (Non Sorted Genetic Algorithm). Le compromis entre les objectifs économiques et écologiques est illustré à travers la génération de fronts de Pareto. La sélection des meilleures solutions parmi cet ensemble est effectuée par utilisation de techniques d’analyse multicritère. L’exemple test bien connu du procédé de production de benzène par hydrodésalkylation du toluène (HDA) est revisité ici dans un mode multiobjectif pour illustrer l’utilité de l’approche pour trouver des solutions de conception économiques et écologiques. / This study aims at the development of a design methodology for eco-efficient processes, meaning that ecological and economic considerations are taken into account simultaneously at the preliminary design phase of chemical processes. The environmental aspect is quantified by using of a set of indicators following the guidelines of sustainability concepts. The design framework is based on a modelling approach considering both process and utility production units, since the environmental impact of a chemical process not only contains the material involved in the process but also the energy consumption, the effect of flow recycle, material conversion and so on... For this purpose, a decision support tool dedicated to the management of plant utilities (steam, electricity, water...) and pollutants (CO2, SO2, NOx, etc..), (ARIANETM package) was coupled to process modelling and used here both to compute the primary energy requirements of the process and to quantify its pollutant emissions. Both models were thus integrated in an outer multiobjective optimization loop, based on a variant of the so-called NSGA-II (Non Sorted Genetic Algorithm) multiobjective genetic algorithm. The trade-off between economic and environmental objectives is illustrated through the generation of Pareto fronts. The selection of the best design alternatives is performed through the use of multicriteria analysis. The well-known benchmark process for hydrodealkylation (HDA) of toluene to produce benzene, revisited here in a multi-objective mode, is used to illustrate the usefulness of the approach in finding environmentally friendly and cost-effective designs.

Essa pesquisa tem como foco a tomada de decisão para a escolha de modais de transporte urbano coletivo e analisa os desempenhos operacionais e funcionais dos sistemas Bus Rapid Transit (BRT), Veículo Leve sobre Trilhos (VLT) e Monotrilho. Um modelo de apoio multicritério à decisão (AMD) em cenários complexos é utilizado na análise e tem como dados de entrada as preferências subjetivas de agentes de decisão (stakeholders) e os desempenhos objetivos e subjetivos dos três modais com relação a um sistema de cinco critérios e 22 subcritérios. Tal sistema cobre os eixos de eficiência energética, custo, viabilidade econômica e financeira, qualidade da viagem e impactos ambientais. Stakeholders selecionados com perfis de operador, vizinho, usuário, fornecedor de equipamentos e serviços e consultor em transportes urbanos atribuem suas preferências, pontuando-as em escala numérica de 1 a 9. Visando a aplicação do modelo na região da cidade de São Paulo, foram coletadas 138 opiniões de preferências através de formulários de avaliação e entrevistas conduzidas nessa região. Cada um dos três modais, após ser configurado para atender aos requisitos operacionais e funcionais do serviço de transporte definido por uma unidade funcional comum, tem os seus desempenhos com relação ao mesmo conjunto de critérios e subcritérios pontuados em uma escala numérica. Um índice global (IG), que define a prioridade de cada alternativa, é obtido pela agregação das preferências e desempenhos com uma função aditiva. O modelo foi aplicado em um estudo de caso que simulou a operação dos três modais no trecho entre as estações Vila Prudente e São Mateus da Linha 15 do Metrô de São Paulo, cujos requisitos definiram a unidade funcional. Dentro dos limites e premissas estabelecidas para o estudo de caso, os índices globais calculados indicaram a preferência pelo modal Monotrilho, seguido, na ordem, pelo VLT e BRT. / This research focuses on decision-making support methods to choice public urban transport modes and analyzes the operational and functional performances of the Bus Rapid Transit (BRT), Light Rail Transit (LRT) and Monorail systems. A multicriteria decision aided model (MDAM) in complex scenarios is used in the analysis and has as input data the subjective preferences of stakeholders and the objective and subjective performances of the three modalities in relation to a set of five criteria and 22 subcriteria. This system covers the axes of energy efficiency, cost, economic and financial feasibility, travel quality and environmental impacts. Stakeholders selected with profiles of operator, neighbor, user, equipment and service provider and consultant in urban transport assign their preferences, scoring them on a numerical scale from 1 to 9. In order to apply the model in the region of the city of São Paulo, 138 opinions were collected through evaluation forms and interviews conducted in that region. Each of the three modes, once configured to meet the operational and functional requirements of the transport service defined by a common functional unit, has its performances against the same set of criteria and subcriteria scored on a numerical scale. A global index (GI), which defines the priority of each alternative, is obtained by aggregating the preferences and performances according to an additive function. The MDAM model was applied in a case study that has simulated the operation of the three transport modalities in the stretch between the stations Vila Prudente and São Mateus of the Line 15 of the São Paulo Metro, whose general requirements did define the functional unit of the case. Within the limits and assumptions defined for the case study, the calculated global indices indicated the first preference for the Monorail modal, followed, in sequence, by the LRT and BRT.

La reconversion des friches urbaines est unepriorité pour préserver les sols. Ce sont souvent des sitespollués ; ce qui engendre de nombreux obstacles pour lesreconvertir. Les bénéfices découlant de la reconversionsont sous-estimés tandis que les coûts sont surestimés parles opérateurs techniques et fonciers. L’impact de lastigmatisation du site amène les riverains à ne pass’approprier les nouveaux usages. Le choix de ces usagespeut être à l'origine de conflits entre les parties prenantes.Cette thèse s'intéresse à la reconversion des frichespolluées en tenant compte à la fois de lamultidimensionnalité des parties prenantes, de leurspréférences, et de leurs perceptions. Nous étudions lesobstacles et les leviers pouvant être mobilisés pourpromouvoir leur reconversion à partir d’une enquête auprès de 76 opérateurs techniques et fonciers dans unepremière partie.Dans la seconde, nous analysonsl’importance des stigmates associés aux friches polluéesauprès des riverains et futurs usagers. Une premièreenquête, menée au niveau national (803 observations)analyse leurs perceptions, représentations et préférencesvis-à-vis de ces sites. Ces dernières sont précisées à l’aided’une expérience à choix discrets (338 observations)administrée sur cinq communes impactées par ce type desite. Dans une troisième partie, nous appliquons unedémarche d’aide multicritère à la décision participativedans le cas d’une friche urbaine polluée. Elle identifie lesprojets les plus consensuels en regroupant l’ensemble desparties prenantes. Ils corresponde / Brownfield redevelopment is a key priority topreserve soils. Brownfields are often contaminated yet.Therefore, it creates numerous obstacles to reuse them.Firstly, brownfields redevelopment’s benefits areunderestimated whereas costs are overestimated bydevelopers. Secondly, brownfields are plagued with thestigma effect. This effect persists even after remediationprocess (e.g., individuals may not use facilities on aformer contaminated brownfield). Thirdly, new uses cancause conflicts between the stakeholders. As such, thisthesis deals with contaminated brownfieldredevelopment taking into account both themultidimensionality of stakeholders, their preferencesand their perceptions. developers. In a first part, we study barriers to brownfieldredevelopment and how to tackle them using a surveyamong 76 French developers.In a second part, weanalyse the importance of the stigma associated withcontaminated brownfields. A first survey (803observations), conducted at national level, investigatesindividuals’ perceptions, representations and preferencesregarding brownfield redevelopment. A second survey(338 observations), conducted among five municipalitiesimpacted by such sites, specifies their preferences usinga discrete choice experiment. In the third part, we applya participatory multicriteria decision aid. It determinesthe most consensual projects in the case of acontaminated site. They correspond to individuals’preferences that were previously analysed.

La present tesi proposa una metodologia de disseny conceptual d'estacions depuradores d'aigües residuals (EDAR) mitjançant la combinació del procés de decisió jeràrquic i l'anàlisi de decisions multicriteri. El document s'inicia amb una breu introducció als principals camps abordats pel treball: el disseny dels processos químics en general, el disseny de les estacions depuradores d'aigües residuals en particular, i l'anàlisi de decisions multicriteri aplicada a la gestió ambiental. Seguidament, es fixen els objectius del treball i es descriuen tant la metodologia com el material de suport informàtic utilitzats. Per validar i contrastar la metodologia de disseny presentada, es desenvolupa un cas d'estudi on es porta a terme el disseny conceptual d'una EDAR que presenta els mateixos requeriments que l'EDAR que opera actualment al municipi de Granollers. Inicialment es presenta la informació de partida i tot seguit es defineixen els objectius de disseny, així com el conjunt de criteris que s'utilitzaran per avaluar en quina mesura es compleixen aquests objectius. Els objectius de disseny són de diferents tipus: ambientals, tècnics, socials i econòmics, i el conjunt de criteris utilitzats, concretament 33, també es classifica segons aquestes quatre categories. Cadascun dels criteris presenta un determinat pes d'importància relativa en la presa de decisions. Finalment, es desenvolupa tot el procés de decisió fins a obtenir el disseny complet de l'EDAR. El procés de decisió s'ha dividit en dues parts diferenciades però que alhora s'entrellacen: la línia d'aigua i la línia de fang. El procés de decisió presenta un total de divuit qüestions amb un màxim de quatre alternatives per pregunta (dotze qüestions corresponen a la línia d'aigua, i sis a la línia de fangs). Per solucionar cadascuna d'aquestes qüestions, s'avaluen les alternatives proposades respecte a un conjunt de criteris triats de la llista inicial. Aplicant el procés de decisió multicriteri anomenat SMART (simple multiattribute rating technique), es combinen els resultats de les alternatives respecte a cada criteri, tenint en compte la importància de cada criteri per obtenir un sol valor per alternativa. Per quantificar els criteris referents a l'operació del procés i les de tipus econòmic s'han utilitzat els programes GPS-X i CapdetWorks respectivament. Pel que fa als criteris no quantificats mitjançant aquests programes, s'han resolt de manera qualitativa i mitjançant manuals de disseny i també tenint en compte l'opinió d'experts en aquest camp. L'alternativa que obté un pes més elevat és la recomanada per al procés de decisió. El cas d'estudi finalitza un cop s'obté el disseny complet de l'EDAR. Per integrar tots aquests elements que hem esmentat i donar suport al desenvolupament del procés de decisió s'ha utilitzat el programa DRAMA (Design Rationale Management). A continuació, es fa una anàlisi comparativa entre l'EDAR que hi ha actualment al municipi de Granollers i l'EDAR resultat del cas d'estudi. Es descriu el diagrama de flux que conforma l'EDAR de Granollers i el diagrama de flux de l'EDAR resultat de l'estudi, se'n fa una anàlisi comparativa justificant cadascuna de les decisions preses en el cas d'estudi i, finalment, es fa una discussió de resultats on es reflecteixen els avantatges associats d'aplicar la metodologia de disseny conceptual proposada. Finalment, es presenten les conclusions de la tesi. Els principals resultats de la tesi es van publicar el 2002 a la revista internacional Industrial and Engineering Chemistry Research (N. Vidal, R. Bañares-Alcántara, I. Rodríguez-Roda i M. Poch: "Design of wastewater treatment plants using a conceptual design methodology", Industrial and Engineering Chemistry Research, 41 (20), pàg. 4993-5005) i la continuació de la línia de recerca al Laboratori d'Enginyeria Química i Ambiental de la UdG ha comportat la presentació del treball de recerca de Xavi Flores "Procés de decisió jeràrquic combinat amb anàlisi multicriteri per al suport al disseny conceptual de sistemes de fangs actius d'una estació depuradora d'aigües residuals" i la presentació dels resultats parcials al congrés internacional de la 9th IWA Conference on Design, Operation and Economics of Large Wastewater Treatment, que va tenir lloc el setembre passat a Praga ("Combining hierarchical decision process with multi-criteria analysis for conceptual design of WWTP", X. Flores, N. Vidal, A. Bonmatí, J. B. Copp i I. Rodríguez-Roda). / This thesis proposes a methodology for the conceptual design of wastewater treatment plants (WWTP) via a combination of a hierarchical decision process and multicriteria decision analysis. We begin with a brief introduction into the main fields of study involved in this work: the design of chemical processes in general; the design of wastewater treatment plants in particular and multicriteria decision analysis as applied to environmental management. We go on to outline the objectives of the study and then describe our methodology as well as the computer-based support tools used. In order to evaluate and contrast our design methodology, we present a study case in which a conceptual design is developed for a WWTP which has the same requirements as the actual WWTP currently in operation in the town of Granollers. First, we present the initial information and then define the design objectives, as well as the set of criteria that will be used in order to evaluate the degree to which the objectives have been met.The design objectives are of different types: environmental, technical, social and economical; the set of criteria, of which there are 33 in all, are also classified into these four categories. Each of the criteria has a specific weighting in terms of their importance in taking decisions. Next, we go through the whole decision process leading to the completion of the design of the WWTP. The decision process has been divided into two parts which are differentiated but at the same time, interconnected: the water line and the sludge line. The decision process involves a total of eighteen questions (twelve for the water line and six for the sludge line) with a maximum of four alternatives per question. In order to answer each question, the proposed alternatives are evaluated in relation to a set of criteria chosen from the initial list. Applying the multicriteria decision process known as SMART (simple multiattribute rating technique), the results for the alternatives with respect to each criteria are combined, bearing in mind the importance of each criteria, in order to get a single value for each alternative. To get the results for those alternatives relating to the operation of the process and those to do with economical factors, we used the GPS-X and CapdetWorks programmes, respectively. The criteria not quantified by means of these programmes were resolved by qualitative means and through the use of design manuals, in addition to taking expert opinion into account. The alternative that obtains the highest weighting is the one which is recommended by the decision process. The study case is completed once a complete design of the WWTP is obtained. In order to integrate all the elements we have mentioned, and to assist the development of the decision process, we employed the programme known as DRAMA (Design Rationale Management). We then go on to give a comparative analysis between the real WWTP in Granollers and the WWTP resulting from our study case. We describe and compare the flow diagrams in both cases, providing justification for each of the decisions taken in the study case and then discuss the results, reflecting on the advantages to be gained from using the methodology of conceptual design we propose. Finally, we present the conclusions of our thesis.The main results of this thesis were published, in 2002, in the international magazine, Industrial and Engineering Chemistry Research (N. Vidal, R. Bañares-Alcántara, I. Rodríguez-Roda and M. Poch: "Design of wastewater treatment plants using a conceptual design methodology", Industrial and Engineering Chemistry Research, 41 (20), pages 4993-5005). Continuing work in this line of research at the Laboratori d'Enginyeria Química i Ambiental at the University of Girona has led to the research work by Xavi Flores, Procés de decisió jeràrquic combinat amb anàlisi multicriteri per al suport al disseny conceptual de sistemes de fangs actius d'una estació depuradora d'aigües residuals (A hierarchical decision process combined with multicriteria analysis to assist conceptual design of active sludge systems in a wastewater treatment plant) and the presentation of the preliminary results at the 9th IWA Conference on Design, Operation and Economics of Large Wastewater Treatment, which took place in September, 2003, in Prague ("Combining hierarchical decision process with multi-criteria analysis for conceptual design of WWTP", X. Flores, Núria Vidal, August Bonmatí, J. B. Copp and Ignasi Rodríguez-Roda).

Η ποιότητα, άρα και ο τρόπος παρακολούθησης της, δηλαδή ο ποιοτικός έλεγχος, σχετίζεται άμεσα με την κερδοφορία μιας επιχείρησης• είτε αυτό επιτευχθεί μέσω αύξησης των πωλήσεων είτε μέσω μείωσης του κόστους παραγωγής ή παροχής υπηρεσιών. Ολοένα και περισσότερο γίνεται αποδεκτό το γεγονός ότι η υψηλή ποιότητα του προϊόντος και των υπηρεσιών και η σύνδεσή τους με την ικανοποίηση του πελάτη, είναι το κλειδί για την επιβίωση οποιασδήποτε επιχείρησης. Επομένως, κρίνεται επιτακτική ανάγκη η μελέτη και έρευνα γύρω από την σχέση της ποιότητας και των επιχειρήσεων και ειδικά σε περιόδους οικονομικής κρίσης και έντονων πιέσεων των καταναλωτών και της αγοράς. Σκοπός της παρούσας εργασίας είναι να παρουσιάσει τη μεθοδολογία σχεδιασμού ενός συστήματος ποιοτικού ελέγχου με τη χρήση μεθόδων πολυκριτηριακής ανάλυσης και να γίνει σύγκριση αυτών των μεθόδων. / The purpose of this thesis is to describe the methodology of a Quality Control System planning using methods of multictiteria analysis and to compare these methods.

Les travaux effectués dans le cadre de cette thèse porte sur le contrôle intelligent, isoarchique et multicritère de flux de personnes dans une zone fermée. Nous proposons un système de contrôle basé sur la biométrie multimodale et le RFID qui sont deux techniques complémentaires pour une sécurisation robuste et flexible du flux de personnes. La biométrie multimodale est utilisée pour une reconnaissance plus fiable des individus, et le RFID pour la sécurisation et le stockage des informations identitaires des personnes à surveiller. Ce système est complètement décentralisé et la décision concernant une demande d'accès est prise de manière autonome au niveau de chaque porte de chaque zone sous contrôle. Les entités internes participantes au processus de prise de décision répondent à des concepts exprimés via le paradigme holonique. L'ouverture automatique d'une porte est conditionnée à la conjonction de plusieurs critères. Une méthode d'aide multicritère à la décision est ainsi déployée au sein de chaque porte d'accès pour fusionner les réponses des identifications biométriques et pour traiter en temps réel les demandes d'autorisation d'accès. Tout d'abord, un état de l'art a été réalisé sur la biométrie, la multimodalité biométrique, la technologie RFID et les systèmes de contrôle d'accès physique. Ensuite, un système de contrôle intelligent, isoarchique et multicritère a été proposé, intégrant l'utilisation simultanée de la multimodalité biométrique et du RFID. Enfin, un démonstrateur du système a été implémenté dans le cadre du contrôle de flux de détenus dans une prison. / The proposed work deals with the intelligent control, isoarchic and multicriteria of people flow in a restricted area. Our proposal is a control system based on a multimodal biometrics and RFID which are considered as two secured complementary techniques for robust and flexible people flow control. Multimodal biometrics is used for more reliable individual recognitions and the RFID for securing and storing supervised individuals identity information. This system is completely decentralized and the decision related to a control access request is made autonomously at each gate of each controlled area. The internal entities which participate to the decision making process respond to the holonic paradigm concepts and principles. The automatic gate opening is conditioned with several criteria conjunction (biometrics identifications, RFID identification, access permissions, authorized paths, status of the zone at time t, etc.). A multicriteria decision aid method is thus deployed in each access gate to merge biometrics identifications responses and to automatically treat the real-time access authorization requests. First, a state of art related to the biometric recognition, the contribution of multimodal biometric, the RFID technology and the physical access control based on biometric, was done. Then, an intelligent, isoarchic and multicriteria control of people flow system was proposed, including the use of multimodal biometric and RFID. At the end, a system simulation test bed was implemented to control prisoners flow in a jail. It supports the integration of various biometrics and RFID technologies.

La qualité de service est devenue aujourd'hui une notion incontournable dans le développement des applications logicielles, en particulier dans le cadre des architectures orientées services. Les travaux de cette thèse se focalisent sur la supervision de la qualité de service des applications orientées services, définies sous forme d'orchestrations de services. L'approche de supervision proposée dans ce contexte est générique. Elle repose sur des patrons de flux de contrôle des orchestrations de services pouvant être implémentés en intégralité ou en partie par tout langage d'orchestration de services. D'autre part, elle ne pose aucune restriction, ni sur les attributs qualité à surveiller par le système de supervision, ni sur leurs représentations. Cette approche de supervision se distingue des approches existantes par l'exploitation d'un modèle de préférences orienté utilisateur, permettant de représenter fidèlement la satisfaction de ce dernier. Le degré de satisfaction, issu du modèle de préférences, constitue une information de haut niveau représentant la qualité globale de l'orchestration étudiée. Sur la base de ce degré de satisfaction, de nouvelles stratégies de surveillance sont proposées afin de satisfaire les attentes de l'utilisateur. L'élaboration du modèle de préférences exploite la méthode d'aide à la décision multi-critères MACBETH étendue avec l'opérateur d'agrégation de l'intégrale de Choquet 2-additive. Une illustration de l'approche de supervision a été réalisée sur une orchestration de services, représentant un processus industriel dans le domaine du pilotage d'atelier de production. Les travaux de cette thèse ont été réalisés dans le cadre d'un projet R&D regroupant sept éditeurs de progiciels dans le domaine du MES (Manufacturing Execution System). / Quality of Service (QoS) is an important issue today in the developement of software applications, especially in the context of Service-Oriented Architectures (SOA). The work of this thesis focuses on QoS supervision of service-oriented applications, defined as service orchestrations. The proposed supervision approach is generic. It is based on workflow control-flow patterns, which can be entirety or partillay implemented by any service orchestration language. On the other hand, it does not make any restrictions, neither on the monitored QoS attributes , nor on their representations. This supervision approach is based on a user-oriented preferences model, that represents faithfully the user satisfaction. The satisfaction degree derived from the preferences model is a high-level information representing the overall quality of the orchestration. New monitoring strategies are proposed on the basis of this satisfaction degree in order to satisfy user expectations. The elaboration of the preferences model uses the MACBETH method extended to the 2-additive Choquet integral operator as a multi criteria decision aiding method. An illustration of the approach is carried out on a service orchestration representing a Manufacturing Execution System (MES) process. This work was realized in a R&D project involving seven software vendors in the field of MES.STAR

Cette thèse traite du processus d’évaluation et de sélection de projets sur la base de critères multiples. Outre la capacité du modèle à permettre une identification efficace des meilleurs projets et leur intégration à un portefeuille, l’équité et la transparence sont des considérations importantes dans la conception de modèles d’appui à ce processus. Nous proposons un cadre de travail général pour l’évaluation de projets, Il reprend les codes de l’analyse SWOT, dont de nombreuses organisations orientées projets sont familières. Nos contributions apportent des éléments de réponse à la question de « l’après SWOT », à laquelle ces organisations peuvent éprouver des difficultés à répondre. Dans ce cadre de travail, nous introduisons et discutons un modèle de préférences permettant de mesurer l’importance des critères sur deux dimensions, représentant de manière indépendante leurs capacités de conviction et d’opposition. Suivant l’évaluation et en préalable à la sélection, le filtrage consiste à écarter les projets trop inadéquats. Nous proposons un mécanisme basé sur la dominance pour effectuer cette opération. Nous proposons, enfin, deux méthodes de sélection de projets, chacune étant basée sur une procédure d’agrégation multicritère originale. La première méthode, SPADE (pour Structure de Préférence pour l’Aide à la Décision) est une approche de surclassement, destinée à des contextes où les préférences exprimées concernent essentiellement les projets individuels, et dans lesquels les décisions concernant un projet peuvent être argumentées en référence à des projets tiers. Nous garantissons la validité théorique de SPADE, en amont, ce qui permet un temps de mise en œuvre réduit et une utilisation en temps réel. En pratique, nous illustrons l’application de SPADE, en la comparant à deux autres approches d’aide multicritères à la décision, MAUT et ELECTRE, en mettant en exergue ses spécificités. La seconde méthode, RADAR (Règles d’Aide à la Décision et à l’ARgumentation) est une approche à base de règles logiques. Elle est destinée à des contextes plus contraints dans lesquels les préférences exprimées concernent à la fois les projets individuels, mais aussi le portefeuille de projets (degré de diversification, budget total, etc.). De plus, l'argumentation des décisions est ici basée exclusivement sur la qualité intrinsèque des projets en référence à une norme fixe. RADAR permet également la construction automatique de tels arguments. Nous proposons un programme linéaire en variable mixtes permettant de valider théoriquement cette approche. Cependant, sa résolution est nécessaire à chaque mise en œuvre de RADAR, ce qui limite l’application de cette approche au temps différé. Nous illustrons une telle application sur un jeu de données représentant des évaluations de projets financés par le Fond des Nations Unies pour la Démocratie (UNDEF). / Project portfolio management (PPM) involves the use of methods and tools, allowing an organization to plan, evaluate, analyze and screen the execution of a set of projects or project proposals, sharing common resources or aiming at the attainment of common objectives. Multicriteria decision aid models are useful tools to support this process, given their ability to accurately model preferences, and rationally agregate points of view. However, existing models present some lacks that limit their use outside of academic circles : (i) They neglect the non-symetrical nature of the importance of some criteria that are relevant in PPM. (ii) The black box effect makes it hard to use them for the argumentation of decisions and to gain their acceptance by users (iii) They are implicitly fitted for private/for-profit projects, which limits their use in public organizations. In this thesis, our contribution consists in proposing two multicriteria methods for supporting the activities of evaluating, selecting and arguing decision, for project portfolio management. We propose: (i) An analysis of the specific features of public and private projects and their consequences for decision support (ii) A framework that allows an independent modeling of the abilities of a criterion to oppose and convince (iii) Two transparent multicriteria agregation procedures, fitted for different decision contexts. We ensure the theoretical validity of our approaches and illustrate their applicability on real data, with satisfying results.

[pt] O objetivo da tese é desenvolver uma metodologia multicritério de apoio à decisão, orientada para a negociação em grupo e apoiada em metáforas técnicas e comportamentais. Em tal metodologia, os modelos multicritério são as ferramentas que os agentes de negociação podem utilizar para análise, comparação, seleção e ordenação das alternativas não-dominadas, segundo as suas preferências e recursos. A busca do consenso (pela discussão e entendimento claro, na participação da essência de pontos de vista individuais mais importantes) é um dos princípios de trabalho do grupo no quadro de referência do processo de decisão. O apoio de metáforas técnicas e comportamentais dos modelos de decisão existentes é fundamental. Essas metáforas, quando combinadas com as tecnologias computacional e de comunicação, ajudam os decisores a utilizar mais eficaz e eficientemente os métodos multicritério de apoio à decisão. A metodologia de apoio à decisão em grupo é ilustrada pela simulação de dois casos (caso 1 - a Compra de um Apartamento e caso 2 - a Fábrica de Montagem de Caminhões - a Manufatura Virtual). Eles mostram a importância e a utilidade desse padrão metodológico no ambiente de decisão em grupo: as metas são melhor explicitadas e evoluem para patamares de preferência superiores e homogêneos no grupo (os objetivos ficam mais claros, explícitos e dimensionados nas discussão do grupo); os critérios individuais acompanham essa dinâmica, transformam-se em critérios decisores do grupo, (independentes, exaustivos e operacionais) e, na etapa final do processo, no(s) critério(s) fundamental(is) decisor (es), como a soma das propriedades fundamentais dos critérios; as soluções são comparadas entre si em sucessivas etapas, conforme o grupo explicita e consensa os critérios individuais, os critérios de grupo e o(s) critério (s) fundamental(is) decisor(es); esse(s) último(s) apontará (ão) a solução não-dominada e final. Em cada uma dessas etapas, as soluções dominadas são abandonadas e só permanecem no processo as não-dominadas. Os resultados nos dois casos mostram uma maior eficiência e satisfação no grupo, pela solução final consensada. Ela é superior qualitativa e quantitativamente à solução indicada pelos modelos tradicionais, até mesmo se comparada à(s) alternativa(s) julgada(s) pelo critério fundamental(is) decisor(es) financeiro, além de mais criativa e envolvida com as variáveis ambientais do processo de decisão. Os riscos assumidos pelo grupo são discutidos, avaliados e minimizados na análise e avaliação das alternativas finais (por meio das ações e reações) e mostram elas serem alcançáveis se implementadas. Elas podem ser maiores, mas são melhor analisadas e ponderadas na racionalidade do consenso do grupo e a solução final aplaudida. / [en] The main purpose of the thesis is to develop a multicriteria decision methodology to support the group negotiation, based on the classic multicritera approaches technical and behaviour metaphors. Such a methodology combines the multicriteria approaches as tool, where the decision agents can use them to analyse, to compare, to select and to rank the non domain solutions, by their preferences and resources. Negotiation is a interactive decision process between two or more agents to research their preferences and needs agreement. The search of the consensus on the group is one important source to the framework and it is primordial the basis of the multicritera approaches technical and behaviour metaphors. These metaphors mix the computer and communication technologies and help, capacity and efficiency, the agents to use the multicriteria approaches. The vigour and value of the multicriteria group system is confirmed by two simulate cases (case 1 - Purchase a Flat; case 2 - The new VW Truck Assemble). They show the approach utility on group decision: the goals flow to high group preferences and homogenous levels; the criteria s increase their value and dynamics, they exchange themselves to decision and last group criteria s (independent, robust and better quantifies ones) to compound a total factor or an identical sum to the criteria properties; the non-domain solutions are compared each other in progression stages, first by the individual criteria properties, after by the group criteria s and finally by the total consensus decision factor, it will mark the winner non domain solution. On each step, the dominate solutions are disregarded and only the non-dominate solutions remain on the process. The winner consensus solution for each case has an higher add value and a group gratification. It has a better quality and quantity dimension (financial and qualifying - industrial production environmental) when it is compared with the classic approach ones; or even with the last non- domain solutions, before the essential decision criteria(s) runs, it is more inventive and the decision process environmental sources can involve it. During the group discussion the risks are contested, valued and minimised when the final solutions (activities and reactions) are compared: they can be greater but they will be better analysed and measured by the rationality group consensus and the final solution is succeeded. / [es] El objetivo de esta tesis es desarrollar una metodología multicriterio de apoyo a la decisión, orientada a la negociación en grupo y apoyada en metáforas técnicas y comportamentales. En tal metodología, los modelos multicriterio constituyen la herramienta principal que los agentes de negociación pueden utilizar para el análisis, comparación, selección y ordenamiento de las alternativas no-dominadas, según sus preferencias y recursos. La búsqueda del consenso (a través de la discusión y entendimiento claro) es uno de los principios de trabajo del grupo en el cuadro de referencia del proceso de decisión. Es fundamental el apoyo de metáforas técnicas y comportamentales de los modelos de decisión existentes. Estas metáforas, cuando se combinan con las tecnologías computacional y de comunicación, ayudan a los decisores a utilizar más eficaz y eficientemente los métodos multicritério de apoyo a la decisión. La metodología de apoyo a la decisión en grupo se ilustra a través de la simulación de dos casos (caso 1 - la Compra de un Apartamento y caso 2 - la Fábrica de Montaje de Camiones - la Manufactura Virtual). Estas simulaciones muestran la importancia y utilidad del padrón metodológico en el ambiente de decisión en grupo: las metas son mejor explicitadas y se observa una evolución en la preferencia y homogeneidad del grupo (los objetivos fican más claros, explícitos y dimensionados en las discusión del grupo). Los criterios individuales acompañan esa dinámica, se transforman en criterios decisores del grupo, (independientes, exhaustivos y operacionales). En la etapa final del proceso, en el(s) criterio(s) fundamental(es) decisor(es), visto como la suma de las propiedades fundamentales de los criterios, las soluciones se compararan entre sí en sucesivas etapas, conforme el grupo explicita y consensa los criterios individuales, los criterios de grupo y el(los) criterio(s) fundamental(es) decisor(es); este(estos) último(s) indicará(an) la solución no-dominada y final. En cada una desas etapas, se abandonan las soluciones dominadas y sólo permanecen en el proceso las no-dominadas. Los resultados en los dos casos muestran una mayor eficiencia y satisfacción en el grupo, por la solución final consensada. Ésta es superior cualitativa y cuantitativamente a la solución indicada por los modelos tradicionales, incluso cuando se compara con la(s) alternativa(s) juzgadas(s) por el criterio fundamental(es) decisor(es) financiero, además de ser más creativa y relacionada con las variables ambientales del proceso de decisión. Los riesgos asumidos por el grupo se discuten, evalúan y minimizan en el análisis y evaluación de las alternativas finales (por medio de las acciones y reacciones). Éstas pueden no ser mayores, mas son mejor analizadas y ponderadas en la racionalidad del consenso del grupo y la solución final, aplaudida.

Essa pesquisa tem como foco a tomada de decisão para a escolha de modais de transporte urbano coletivo e analisa os desempenhos operacionais e funcionais dos sistemas Bus Rapid Transit (BRT), Veículo Leve sobre Trilhos (VLT) e Monotrilho. Um modelo de apoio multicritério à decisão (AMD) em cenários complexos é utilizado na análise e tem como dados de entrada as preferências subjetivas de agentes de decisão (stakeholders) e os desempenhos objetivos e subjetivos dos três modais com relação a um sistema de cinco critérios e 22 subcritérios. Tal sistema cobre os eixos de eficiência energética, custo, viabilidade econômica e financeira, qualidade da viagem e impactos ambientais. Stakeholders selecionados com perfis de operador, vizinho, usuário, fornecedor de equipamentos e serviços e consultor em transportes urbanos atribuem suas preferências, pontuando-as em escala numérica de 1 a 9. Visando a aplicação do modelo na região da cidade de São Paulo, foram coletadas 138 opiniões de preferências através de formulários de avaliação e entrevistas conduzidas nessa região. Cada um dos três modais, após ser configurado para atender aos requisitos operacionais e funcionais do serviço de transporte definido por uma unidade funcional comum, tem os seus desempenhos com relação ao mesmo conjunto de critérios e subcritérios pontuados em uma escala numérica. Um índice global (IG), que define a prioridade de cada alternativa, é obtido pela agregação das preferências e desempenhos com uma função aditiva. O modelo foi aplicado em um estudo de caso que simulou a operação dos três modais no trecho entre as estações Vila Prudente e São Mateus da Linha 15 do Metrô de São Paulo, cujos requisitos definiram a unidade funcional. Dentro dos limites e premissas estabelecidas para o estudo de caso, os índices globais calculados indicaram a preferência pelo modal Monotrilho, seguido, na ordem, pelo VLT e BRT. / This research focuses on decision-making support methods to choice public urban transport modes and analyzes the operational and functional performances of the Bus Rapid Transit (BRT), Light Rail Transit (LRT) and Monorail systems. A multicriteria decision aided model (MDAM) in complex scenarios is used in the analysis and has as input data the subjective preferences of stakeholders and the objective and subjective performances of the three modalities in relation to a set of five criteria and 22 subcriteria. This system covers the axes of energy efficiency, cost, economic and financial feasibility, travel quality and environmental impacts. Stakeholders selected with profiles of operator, neighbor, user, equipment and service provider and consultant in urban transport assign their preferences, scoring them on a numerical scale from 1 to 9. In order to apply the model in the region of the city of São Paulo, 138 opinions were collected through evaluation forms and interviews conducted in that region. Each of the three modes, once configured to meet the operational and functional requirements of the transport service defined by a common functional unit, has its performances against the same set of criteria and subcriteria scored on a numerical scale. A global index (GI), which defines the priority of each alternative, is obtained by aggregating the preferences and performances according to an additive function. The MDAM model was applied in a case study that has simulated the operation of the three transport modalities in the stretch between the stations Vila Prudente and São Mateus of the Line 15 of the São Paulo Metro, whose general requirements did define the functional unit of the case. Within the limits and assumptions defined for the case study, the calculated global indices indicated the first preference for the Monorail modal, followed, in sequence, by the LRT and BRT.