Dissertations / Theses on the topic 'Network Equilibrium Modeling'

Network design decisions, especially those pertaining to urban infrastructure, are made by a central authority or network leader, and taking into consideration the network users or followers. These network decision problems are formulated as non-linear bi-level programming problems. In this work, a continuous network design problem (CNDP) and discrete network design problem (DNDP) bi-level optimization programs are proposed and solved in the context of transportation planning. The solution strategy involved reformulation and linearization as a single-level program by introducing the optimality conditions of the lower level problem into the upper level problem. For the CNDP, an alternative linearization algorithm (modified least squares partitioning, MLSPA) is proposed. MLSPA takes into consideration the current arc capacity and potential expansion to find a reduced set of planes to generalize the flow-capacity surface behavior. The concepts of flow capacity surface was introduced as a way to model of congested network and capture the effect of capacity on travel time/cost. It was found that the quality of the linear approximation depends on the goodness of fit the bottleneck arcs. The proposed approach was tested with well-known benchmark problems in transportation which yielded promising results in terms of efficiency, without sacrificing solution quality.

A gaseificação é um processo de conversão termoquímica que compreende a oxidação parcial de um combustível para convertê-lo em uma mistura gasosa (“syngas”). Geralmente a modelagem desses processos utiliza uma descrição cinética detalhada ou os aproximam ao equilíbrio químico. Ambas as abordagens têm vantagens e desvantagens, bem como limitações. O objetivo deste trabalho foi o desenvolvimento de uma nova modelagem fenomenológica de processos de gaseificação através de um modelo “híbrido” aqui chamado de modelo híbrido adaptativo por zonas (HAZ). Este modelo assumiu que o gaseificador é representado por zonas de dois tipos: uma dominada pela cinética química, representada por um modelo cinético, e outra onde a cinética química é rápida e as espécies químicas estão em equilíbrio químico, representada por um modelo de equilíbrio. Um critério de transição entre as zonas foi proposto através de um número de Damköhler (Da) que relaciona tempos de residência e de reação química. Desta forma, o modelo adapta-se conforme os processos dominantes em cada zona. Em um primeiro momento, um modelo de equilíbrio multifásico (EM) foi desenvolvido e aplicado para um estudo da cogaseificação de carvão mineral e biomassas disponíveis no Brasil. A seguir, o modelo HAZ foi construído através da técnica de rede equivalente de reatores químicos (ERN) a partir do modelo EM e de um modelo cinético, também desenvolvido neste trabalho. Uma metodologia de aplicação do modelo HAZ foi proposta, aplicada e validada para duas configurações de gaseificadores: dois casos de gaseificadores de biomassa em leito fluidizado borbulhante e um caso de gaseificador de carvão mineral em leito de arraste. Para os dois primeiros casos foi estimada que a transição ocorra para Da ≥ 10+5 e para o último caso; chegou-se a Da ≥ 10+3. A aplicação do modelo HAZ se mostrou satisfatória sendo que foi possível a redução do tempo computacional em pelo menos 40% com relação a uma abordagem puramente cinética. Cabe ressaltar ainda que o modelo HAZ possibilitou um maior entendimento físico e químico ao identificar os processos dominantes locais.
Gasification is a thermochemical conversion process consisting of partial oxidation of a fuel to convert it to a gas mixture (“syngas”). Generally, the gasification process modeling uses a kinetic detailed description, or approach it to a chemical equilibrium state. Both approaches have advantages and disadvantages, as well as limitations. The objective of this work was to develop a new phenomenological modeling of gasification processes through a “hybrid” model here called hybrid adaptive zone model (HAZ). This proposed modeling assumed the gasifier is represented by two types of zones: one dominated by chemical kinetics, represented by a kinetic model, and another where chemical kinetics is fast so chemical species are assumed in chemical equilibrium states, represented by an equilibrium model. A transition criterion between zones was defined by a Damköhler number (Da) which relates residence time and chemical reaction time. Therefore, the HAZ model can adapted according to the dominant processes in each zone. Firstly, a multi-phase equilibrium model (ME) was developed and applied to study the coal-biomass co-gasification of Brazilian sources. Hereafter, the HAZ model was built using the technique of equivalent reactor network (ERN) with the ME model and a kinetic model developed in this work. A methodology of use of the HAZ model was proposed, applied and validated for two configurations of gasifiers: two cases of biomass bubbling fluidized-bed gasifiers and one case of coal entrained-flow gasifier. In the first two cases the transition was estimated to occur on Da ≥ 10+5 and in the last case; we estimated on Da ≥ 10+3. The application of the HAZ model proved to be satisfactory since it could reduce the computation time by at least 40% compared to a pure kinetic approach. It should already be emphasized that the HAZ model allowed a better physical and chemical understanding of gasification by identifying the dominant local processes.

The question of how to design a congestion pricing scheme is difficult to answer and involves a number of complex decisions. This thesis is devoted to the quantitative parts of designing a congestion pricing scheme with link tolls in an urban car traffic network. The problem involves finding the number of tolled links, the link toll locations and their corresponding toll level. The road users are modeled in a static framework, with elastic travel demand.

Assuming the toll locations to be fixed, we recognize a level setting problem as to find toll levels which maximize the social surplus. A heuristic procedure based on sensitivity analysis is developed to solve this optimization problem. In the numerical examples the heuristic is shown to converge towards the optimum for cases when all links are tollable, and when only some links are tollable.

We formulate a combined toll location and level setting problem as to find both toll locations and toll levels which maximize the net social surplus, which is the social surplus minus the cost of collecting the tolls. The collection cost is assumed to be given for each possible toll location, and to be independent of toll level and traffic flow. We develop a new heuristic method which is based on repeated solutions of an approximation to the combined toll location and level setting problem. Also, a known heuristic method for locating a fixed number of toll facilities is extended, to find the optimal number of facilities to locate. Both heuristics are evaluated on two small networks, where our approximation procedure shows the best results.

Our approximation procedure is also employed on the Sioux Falls network. The result is compared with different judgmental closed cordon structures, and the solution suggested by our method clearly improves the net social surplus more than any of the judgmental cordons.

In this study a dynamic assignment model is developed which estimates travellers' route and departure time choices and the resulting time varying traffic patterns during the morning peak. The distinctive feature of the model is that it does not restrict the geometry of the network to specific forms. The proposed framework of analysis consists of a travel time model, a demand model and a demand adjustment mechanism. Two travel time models are proposed. The first is based on elementary relationships from traffic flow theory and provides the framework for a macroscopic simulation model which calculates the time varying flow patterns and link travel times given the time dependent departure rate distributions; the second is based on queueing theory and models roads as bottlenecks through which traffic flow is either uncongested or fixed at a capacity independent of traffic density. The demand model is based on the utility maximisation decision rule and defines the time dependent departure rates associated with each reasonable route connecting, the O-D pairs of the network, given the total utility associated with each combination of departure time and route. Travellers' choices are assumed to result from the trade-off between travel time and schedule delay and each individual is assumed to first choose a departure time t, and then select a reasonable route, conditional on the choice of t. The demand model has therefore the form of a nested logit. The demand adjustment mechanism is derived from a Markovian model, and describes the day-to-day evolution of the departure rate distributions. Travellers are assumed to modify their trip choice decisions based on the information they acquire from recent trips. The demand adjustment mechanism is used in order to find the equilibrium state of the system, defined as the state at which travellers believe that they cannot increase their utility of travel by unilaterally changing route or departure time. The model outputs exhibit the characteristics of real world traffic patterns observed during the peak, i. e., time varying flow patterns and travel times which result from time varying departure rates from the origins. It is shown that increasing the work start time flexibility results in a spread of the departure rate distributions over a longer period and therefore reduces the level of congestion in the network. Furthermore, it was shown that increasing the total demand using the road network results in higher levels of congestion and that travellers tend to depart earlier in an attempt to compensate for the increase in travel times. Moreover, experiments using the queueing theory based travel time model have shown that increasing the capacity of a bottleneck may cause congestion to develop downstream, which in turn may result in an increase of the average travel time for certain O-D pairs. The dynamic assignment model is also applied to estimate the effects that different road pricing policies may have on trip choices and the level of congestion; the model is used to demonstrate the development of the shifting peak phenomenon. Furthermore, the effect of information availability on the traffic patterns is investigated through a number of experiments using the developed dynamic assignment model and assuming that guided drivers form a class of users characterised by lower variability of preferences with respect to route choice.

El mundo de las telecomunicaciones está cambiando de un escenario donde únicamente las personas estaban conectadas a un modelo donde prácticamente todos los dispositivos y sensores se encuentran conectados, también conocido como Internet de las cosas (IoT), donde miles de millones de dispositivos se conectarán a Internet a través de conexiones móviles y redes fijas. En este contexto, hay muchos retos que superar, desde el desarrollo de nuevos estándares de comunicación al estudio de la viabilidad económica de los posibles escenarios futuros. En esta tesis nos hemos centrado en el estudio de la viabilidad económica de diferentes escenarios mediante el uso de conceptos de microeconomía, teoría de juegos, optimización no lineal, economía de redes y redes inalámbricas. La tesis analiza la transición desde redes centradas en el servicio de tráfico HTC a redes centradas en tráfico MTC desde un punto de vista económico. El primer escenario ha sido diseñado para centrarse en las primeras etapas de la transición, en la que ambos tipos de tráfico son servidos bajo la misma infraestructura de red. En el segundo escenario analizamos la siguiente etapa, en la que el servicio a los usuarios MTC se realiza mediante una infraestructura dedicada. Finalmente, el tercer escenario analiza la provisión de servicios basados en MTC a usuarios finales, mediante la infraestructura analizada en el escenario anterior. Gracias al análisis de todos los escenarios, hemos observado que la transición de redes centradas en usuarios HTC a redes MTC es posible y que la provisión de servicios en tales escenarios es viable. Además, hemos observado que el comportamiento de los usuarios es esencial para determinar la viabilidad de los diferentes modelos de negocio, y por tanto, es necesario estudiar el comportamiento y las preferencias de los usuarios en profundidad en estudios futuros. Específicamente, los factores más relevantes son la sensibilidad de los usuarios al retardo en los datos recopilados por los sensores y la cantidad de los mismos. También hemos observado que la diferenciación del tráfico en categorías mejora el uso de las redes y permite crear nuevos servicios empleando datos que, de otro modo, no se aprovecharían, lo cual nos permite mejorar la monetización de la infraestructura. También hemos demostrado que la provisión de capacidad es un mecanismo válido, alternativo a la fijación de precios, para la optimización de los beneficios de los proveedores de servicio. Finalmente, se ha demostrado que es posible crear roles específicos para ofrecer servicios IoT en el mercado de las telecomunicaciones, específicamente, los IoT-SPs, que proporcionan servicios basados en sensores inalámbricos utilizando infraestructuras de acceso de terceros y sus propias redes de sensores. En resumen, en esta tesis hemos intentado demostrar la viabilidad económica de modelos de negocio basados en redes futuras IoT, así como la aparición de nuevas oportunidades y roles de negocio, lo cual nos permite justificar económicamente el desarrollo y la implementación de las tecnologías necesarias para ofrecer servicios de acceso inalámbrico masivo a dispositivos MTC.
The communications world is moving from a standalone devices scenario to a all-connected scenario known as Internet of Things (IoT), where billions of devices will be connected to the Internet through mobile and fixed networks. In this context, there are several challenges to face, from the development of new standards to the study of the economical viability of the different future scenarios. In this dissertation we have focused on the study of the economic viability of different scenarios using concepts of microeconomics, game theory, non-linear optimization, network economics and wireless networks. The dissertation analyzes the transition from a Human Type Communications (HTC) to a Machine Type Communications (MTC) centered network from an economic point of view. The first scenario is designed to focus on the first stages of the transition, where HTC and MTC traffic are served on a common network infrastructure. The second scenario analyzes the provision of connectivity service to MTC users using a dedicated network infrastructure, while the third stage is centered in the analysis of the provision of services based on the MTC data over the infrastructure studied in the previous scenario. Thanks to the analysis of all the scenarios we have observed that the transition from HTC users-centered networks to MTC networks is possible and that the provision of services in such scenarios is viable. In addition, we have observed that the behavior of the users is essential in order to determine the viability of a business model, and therefore, it is needed to study their behavior and preferences in depth in future studios. Specifically, the most relevant factors are the sensitivity of the users to the delay and to the amount of data gathered by the sensors. We also have observed that the differentiation of the traffic in categories improves the usage of the networks and allows to create new services thanks to the data that otherwise would not be used, improving the monetization of the infrastructure and the data. In addition, we have shown that the capacity provision is a valid mechanism for providers' profit optimization, as an alternative to the pricing mechanisms. Finally, it has been demonstrated that it is possible to create dedicated roles to offer IoT services in the telecommunications market, specifically, the IoT-SPs, which provide wireless-sensor-based services to the final users using a third party infrastructure. Summarizing, this dissertation tries to demonstrate the economic viability of the future IoT networks business models as well as the emergence of new business opportunities and roles in order to justify economically the development and implementation of the new technologies required to offer massive wireless access to machine devices.
El món de les telecomunicacions està canviant d'un escenari on únicament les persones estaven connectades a un model on pràcticament tots els dispositius i sensors es troben connectats, també conegut com a Internet de les Coses (IoT) , on milers de milions de dispositius es connectaran a Internet a través de connexions mòbils i xarxes fixes. En aquest context, hi ha molts reptes que superar, des del desenrotllament de nous estàndards de comunicació a l'estudi de la viabilitat econòmica dels possibles escenaris futurs. En aquesta tesi ens hem centrat en l'estudi de la viabilitat econòmica de diferents escenaris per mitjà de l'ús de conceptes de microeconomia, teoria de jocs, optimització no lineal, economia de xarxes i xarxes inalàmbriques. La tesi analitza la transició des de xarxes centrades en el servici de tràfic HTC a xarxes centrades en tràfic MTC des d'un punt de vista econòmic. El primer escenari ha sigut dissenyat per a centrar-se en les primeres etapes de la transició, en la que ambdós tipus de tràfic són servits davall la mateixa infraestructura de xarxa. En el segon escenari analitzem la següent etapa, en la que el servici als usuaris MTC es realitza per mitjà d'una infraestructura dedicada. Finalment, el tercer escenari analitza la provisió de servicis basats en MTC a usuaris finals, per mitjà de la infraestructura analitzada en l'escenari anterior. Als paràgrafs següents es descriu amb més detall cada escenari. Gràcies a l'anàlisi de tots els escenaris, hem observat que la transició de xarxes centrades en usuaris HTC a xarxes MTC és possible i que la provisió de servicis en tals escenaris és viable. A més a més, hem observat que el comportament dels usuaris és essencial per a determinar la viabilitat dels diferents models de negoci, i per tant, és necessari estudiar el comportament i les preferències dels usuaris en profunditat en estudis futurs. Específicament, els factors més rellevants són la sensibilitat dels usuaris al retard en les dades recopilats pels sensors i la quantitat dels mateixos. També hem observat que la diferenciació del tràfic en categories millora l'ús de les xarxes i permet crear nous servicis emprant dades que, d'una altra manera, no s'aprofitarien, la qual cosa ens permet millorar la monetització de la infraestructura. També hem demostrat que la provisió de capacitat és un mecanisme vàlid, alternatiu a la fixació de preus, per a l'optimització dels beneficis dels proveïdors de servici. Finalment, s'ha demostrat que és possible crear rols específics per a oferir servicis IoT en el mercat de les telecomunicacions, específicament, els IoT-SPs, que proporcionen servicis basats en sensors inalàmbrics utilitzant infraestructures d'accés de tercers i les seues pròpies xarxes de sensors. En resum, en aquesta tesi hem intentat demostrar la viabilitat econòmica de models de negoci basats en xarxes futures IoT, així com l'aparició de noves oportunitats i rols de negoci, la qual cosa ens permet justificar econòmicament el desenrotllament i la implementació de les tecnologies necessàries per a oferir servicis d'accés inalàmbric massiu a dispositius MTC.
Sanchis Cano, Á. (2018). Economic analysis of wireless sensor-based services in the framework of the Internet of Things. A game-theoretical approach [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/102642
TESIS

Cette thèse porte sur les modèles d'équilibre dynamique sur un réseau de transport et de leurs applications à l'affectation de trafic. Elle vise à proposer une formulation à la fois générale, mathématiquement rigoureuse et microéconomiquement cohérente de l'équilibre dynamique. Une attention toute particulière est accordée à la représentation de la demande de transport et plus spécifiquement à la modélisation des hétérogénéités dans les préférences des usagers du réseau, ainsi que de leurs stratégies de choix d'horaires dans leurs déplacements. Tout d'abord nous exprimons l'équilibre dynamique comme un jeu de Nash avec un continuum de joueurs. Cette formulation nous permet d'obtenir un résultat d'existence. Celui-ci s'applique notamment au plus simple des modèles d'équilibre dynamique, où les usagers sont identiques et ne choisissent pas leurs horaires de départ. Ensuite, nous présentons deux modèles d'équilibre pour lesquels une solution analytique peut être établie. Le premier est une généralisation du modèle de goulot de Vickrey. Alors que Vickrey considère une distribution des horaires préférés d'arrivée en forme de S, nous traitons de distributions quelconques. Le deuxième modèle proposé est un réseau à péage avec deux routes et des usagers dont la valeur du temps est distribuée. Ce modèle nous permet d'investiguer les efficacités relatives de différentes stratégies de tarification et de voir comment celles-ci sont affectés par le niveau d'hétérogénéité dans la valeur du temps. Pour finir, un modèle calculable est présenté et des méthodes de résolution sont proposées. Le modèle est testé sur le réseau routier national. Par ailleurs, il est exploité pour tester une tarification modulée en fonction du temps dont l'objectif est d'atténuer la congestion lors des grands départs de vacances
This thesis is focused on dynamic user equilibrium (DUE) models and theirapplications to traffic assignment. It aims at providing a mathematically rigorous and general formulation for the DUE. Particular attention is paid to the representation of transport demand and more specifically to trip scheduling and users with heterogeneous preferences.The DUE is first expressed as a Nash game with a continuum of players. It strongly relies on up-to-date results from mathematical economics. This formulation allows to prove an existence result for DUE. This results notably applies to one of the simplest dynamic user equilibrium model, where users are homogeneous and departure time choice is not allowed.Then, two simple DUE models for which the solutions can be derived analytically are presented. The first one is a generalization of the Vickrey's bottleneck model. Whereas Vickrey assumed that the distribution of preferred arrival time is S-shaped, we consider more general distributions. In the second model, we have a two-route tolled network where users are continuously heterogeneous with respect to their value of time. This allows us to conduct a study on the relative efficiencies of various pricing strategy and how it is affected by the level of heterogeneity in users' value of time.Finally, a computable model is designed and corresponding solution methods are proposed. A test on the french national road network is conducted. The model is used to assess an hypothetical time-varying pricing scheme intended to ease summer traffic congestion

Nous abordons le problème de la coordination du mouvement par les deux approches continues et discrètes, afin de les comparer en vue du neurocontrôle en robotique. Dans le premier chapitre, le formalisme des champs et les équations du champ pour un tissu nerveux ont été rappelés. Ces équations incluent les mécanismes physiologiques du système nerveux pour mieux tenir compte de la réalité. Les hypothèses permettant de retrouver les modèles classiques ont été déterminées, prouvant ainsi la généralité de la théorie du champ envisagée. Dans le deuxième chapitre, la résolution numérique des équations du champ est réalisée au moyen de la technique des éléments finis et des différences finies. Il est supposé que l'équation au niveau neuronal est bidimensionnelle, alors que celle du niveau synaptique est à une dimension. A la fin de ce chapitre, nous proposons des simulations pour les réseaux multicouches classiques illustrant la propagation de l'activation. Dans le troisième chapitre, nous présentons les premiers résultats de l'application du formalisme des champs au cervelet. En particulier, les propriétés concernant l'effet de la localisation géométrique des neurones et l'effet de la hiérarchie sont déduites. Dans le quatrième chapitre, nous commençons par l'étude du modèle du cervelet en boucle fermée en analysant le rôle de la fibre grimpante dont on sait qu'elle véhicule le signal d'erreur. La méthode analytique utilisée, est fondée sur une technique classique d'optimisation. Nous généralisons ensuite notre méthode. Enfin nous abordons dans une dernière section un aspect de la robotique mobile. Nous avons considéré l'exemple simplifie d'un bipède dont il faut contrôler l'équilibre dynamique par l'accélération articulaire du tronc. On montre que notre modèle permet d'apprendre à contrôler l'équilibre dynamique du système pour les trajectoires apprises et d'anticiper celles qui sont non apprises.

The objective of this work is the formulation, development and implementation of Artificial Neural Networks (ANNs) to turbulent combustion problems, for the representation of reduced chemical kinetics. Although ANNs are general and robust tools for simulating dynamical systems within reasonable computational times, their employment in combustion has been limited. In previous studies, ANNs were trained with data collected from either the test case of interest or from a similar problem. To overcome this training drawback, in this work, ANNs are trained with samples generated from an abstract problem; the laminar flamelet equation, allowing the simulation of a wide range of problems. To achieve this, the first step is to reduce a detailed chemical mechanism to a manageable number of variables. This task is performed by the Rate-Controlled Constrained Equilibrium (RCCE) reduction method. The training data sets consist of the composition of points with random mixture fraction, recorded from flamelets with random strain rates. The training, testing and simulation of the ANNs is carried out via the Self-Organising Map - Multilayer Perceptrons (SOM-MLPs) approach. The SOM-MLPs combination takes advantage of a reference map and splits the chemical space into domains of chemical similarity, allowing the employment of a separate MLP for each sub-domain. The RCCE-ANNs tabulation is used to replace conventional chemistry integration methods in RANS computations and LES of real turbulent flames. In the context of RANS the interaction of turbulence and combustion is described by using a PDF method utilising stochastic Lagrangian particles. In LES the sub-grid PDF is represented by an ensemble of Eulerian stochastic fields. Test cases include non-premixed and partially premixed turbulent flames in both non-piloted and piloted burner configurations. The comparison between RCCE-ANNs, real-time RCCE and experimental measurements shows good overall agreement in reproducing the overall flame structure and a significant speed-up of CPU time by the RCCE-ANN method.

Made available in DSpace on 2017-07-10T18:08:10Z (GMT). No. of bitstreams: 1 Eliseu A Zanella Junior.pdf: 762348 bytes, checksum: 96525daf5cc5df21f8559d564eedfb4d (MD5) Previous issue date: 2009-02-13
In the majority of the applications of the process of ionic exchange in the chemical industry some ionic species are gifts that compete between itself for the active small sieges of the ionic exchanger. Therefore, the project of these systems requires an analysis of the coefficients of selectivity of ions gifts in the solution that determines the influence of the separation process. The data of balance of processes of ionic exchange generally are discrebed for the Law of the Action of the Masses, therefore in this boarding the no-idealists of the phases are consideret watery and solid. The calculation of Balance in systems of ionic exchange in multicomponent systems requires the resolution of a system of not linear equations, and depending on the number of involved species one high computational time cam be required. An alternative to the conventional modelin is the job of Artificial the Neural Nets. Inside of this context, the objective of the present work was to evaluate the application of Artificial the Neural Nets in the modeling of the binary and ternary data of balance in systems of ionic exchange, and also to evaluate the viability to apply Artificial the Neural Nets in the prediction of the data of balance of the ternary systems from information of the binary systems. To evaluate the efficiency of Artificial the Neural Nets in the description of the data of balance of systems of ionic exchange, the gotten results had been compared with the values calculated for the application of the Law of the Action of the Masses. Two experimental data sets of ionic exchange had been used. The first set was constituted of the binary and ternary systems of ions sulphate, chloride and nitrate and as exchanging ion the resin AMBERLITE ANGER 400, with total concentration of 0,2N 298K and had been gotten by SMITH and WOODBURN (1978). As the joint one was constituted of the binary and ternary data of ions of lead, has covered and ionic sodium and as exchanging the clinoptilolita, with 0,005 concentration eq/L and temperature of 303K, gotten for FERNANDEZ (2004). The data of entrance of the net had been the composition of íons in solution and of exit they had been the composition of the resin. The training of diverse structures of RNAs was effected. Different architectures had been tested varying the nunber of neurons of the laver of entrance and the occult layer. The nunber of neurons of the entrance layer varied of 2 up to 20 and the occult layer of 1 up to 2, searching always a structure with the lesser value of the objective function. The methods Powell and Simplex had been used to determine the weights of the net. The Law of the Action of the Masses revealed efficient in the description of the following binary systems: SO4-2-NO3-, SO4-2-Cl- e NO3 --Cl-Pb2+-Na+, Cu2+-Na+, however, the results for the system Na+-Pb2+ had not been satisfactory. In the modeling of the binary data Artificial the Neural Nets if had shown efficient in all the investigated cases. In the prediction of the ternary system the Law of the Action of the Masses only revealed efficient for systems SO42--NO3-, SO42-CI- e NO3--CI-. In the prediction of the data of ternary balance for the two evaluatede systems, using Artificial the neural Nets from the binary data generated by the Law of the Action of the Masses, one did not reveal efficient. In the ternary system (SO4²-, NO3-,CI-) the trained Artificial Neural Nets with the binary data set and the inclusion of ternary experimental data of balance (three and seven data) had obtained to represent with precision the behavior of the system. In the ternary system (Pb+². Cu+², Na+), hte nets trained from the binary data set and with the inclusion of all the experimental data of the ternary system, the gotten results had been satisfactory, because they had presented errors near by 2% to 6%. Artificial the Neural Nets had not presented predictive capacity to describe the balance in the process of ionic exchange. However, the nets present an advantage in relation the Law of the Action of the Masses, to allow that the compositions of balance of the resin are calculated explicit.
Na maioria das aplicações do processo de troca iônica na indústria química estão presentes várias espécies iônicas que competem entre si pelos sítios ativos do trocador iônico. Portanto, o projeto destes sistemas requer uma análise dos coeficientes de seletividade dos íons presentes na solução que determina a influência do processo de separação. Os dados de equilíbrio de processos de troca iônica geralmente são descritos pela Lei da Ação das Massas, pois nesta abordagem são considerados as não-idealidades das fases aquosa e sólida. O cálculo de Equilíbrio em sistemas de troca iônica em sistemas multicomponentes requer a resolução de um sistema de equações não lineares, e dependendo do número de espécies envolvidas pode-se requerer um elevado tempo computacional. Uma alternativa à modelagem convencional é o emprego das Redes Neurais Artificiais. Dentro deste contexto, o objetivo do presente trabalho foi avaliar a aplicação das Redes Neurais Artificiais na modelagem dos dados binários e ternários de equilíbrio em sistemas de troca iônica, e também avaliar a viabilidade de aplicar as Redes Neurais Artificiais na predição dos dados de equilíbrio dos sistemas ternários a partir de informações dos sistemas binários. Para avaliar a eficiência das Redes Neurais Artificiais na descrição dos dados de equilíbrio de sistemas de troca iônica, os resultados obtidos foram comparados com os valores calculados pela aplicação da Lei da Ação das Massas. Foram utilizados dois conjuntos de dados experimentais de troca iônica. O primeiro conjunto era constituído pelos sistemas binários e ternários dos íons sulfato, cloreto e nitrato e como trocador iônico a resina AMBERLITE IRA 400, com concentração total de 0,2N a 298K e foram obtidos por SMITH e WOODBURN (1978). O segundo conjunto era constituído dos dados binários e ternários dos íons de chumbo, cobre e sódio e como trocador iônico a clinoptilolita, com concentração 0,005 eq/L e temperatura de 303K, obtidos por FERNANDEZ (2004). Os dados de entrada da rede foram a composição dos íons em solução e de saída foram a composição da resina. Efetuou-se o treinamento de diversas estruturas de RNAs. Foram testadas diferentes arquiteturas variando o número de neurônios da camada de entrada e da camada oculta. O número de neurônios da camada de entrada variou de 2 até 20 e da camada oculta de 1 até 2, buscando sempre uma estrutura com o menor valor da função objetivo. Os métodos Powell e Simplex foram utilizados para determinar os pesos da rede. A Lei da Ação das Massas mostrou-se eficiente na descrição dos seguintes sistemas binários: SO42--NO3-, SO42--Cl- e NO3--Cl- Pb2+-Na+, Cu2+-Na+, entretanto, os resultados para o sistema Na+-Pb2+ não foram satisfatórios. Na modelagem dos dados binários as Redes Neurais Artificiais se mostraram eficientes em todos os casos investigados. Na predição do sistema ternário a Lei da Ação das Massas mostrou-se eficiente somente para os sistemas SO42--NO3-, SO42--Cl- e NO3--Cl-. Na predição dos dados de equilíbrio ternário para os dois sistemas avaliados, empregando as Redes neurais Artificiais a partir dos dados binários gerados pela Lei da Ação das Massas, não se mostrou eficiente. No sistema ternário (SO42-,NO3-,Cl-) as Redes Neurais Artificiais treinadas com o conjunto de dados binários e com a inclusão de dados experimentais ternários de equilíbrio (três e sete dados) conseguiram representar com precisão o comportamento do sistema. No sistema ternário (Pb+2,Cu+2,Na+), as redes treinadas a partir do conjunto de dados binários e com a inclusão de todos os dados experimentais do sistema ternário, os resultados obtidos foram satisfatórios, pois apresentaram erros na faixa de 2% a 6%. As Redes Neurais Artificiais não apresentaram capacidade preditiva de descrever o equilíbrio no processo de troca iônica. Entretanto, as redes apresenta uma vantagem em relação a Lei da Ação das Massas, permitir que sejam calculados explicitamente as composições de equilíbrio da resina.

Accurate models to reliably predict sediment and chemical transport in watershed water systems enhance the ability of environmental scientists, engineers and decision makers to analyze the impact of contamination problems and to evaluate the efficacy of alternative remediation techniques and management strategies prior to incurring expense in the field. This dissertation presents the conceptual and mathematical development of a general numerical model simulating (1) sediment and reactive chemical transport in river/stream networks of watershed systems; (2) sediment and reactive chemical transport in overland shallow water of watershed systems; and (3) reactive chemical transport in three-dimensional subsurface systems. Through the decomposition of the system of species transport equations via Gauss-Jordan column reduction of the reaction network, fast reactions and slow reactions are decoupled, which enables robust numerical integrations. Species reactive transport equations are transformed into two sets: nonlinear algebraic equations representing equilibrium reactions and transport equations of kinetic-variables in terms of kinetically controlled reaction rates. As a result, the model uses kinetic-variables instead of biogeochemical species as primary dependent variables, which reduces the number of transport equations and simplifies reaction terms in these equations. For each time step, we first solve the advective-dispersive transport of kinetic-variables. We then solve the reactive chemical system node by node to yield concentrations of all species. In order to obtain accurate, efficient and robust computations, five numerical options are provided to solve the advective-dispersive transport equations; and three coupling strategies are given to deal with the reactive chemistry. Verification examples are compared with analytical solutions to demonstrate the numerical accuracy of the code and to emphasize the need of implementing various numerical options and coupling strategies to deal with different types of problems for different application circumstances. Validation examples are presented to evaluate the ability of the model to replicate behavior observed in real systems. Hypothetical examples with complex reaction networks are employed to demonstrate the design capability of the model to handle field-scale problems involving both kinetic and equilibrium reactions. The deficiency of current practices in the water quality modeling is discussed and potential improvements over current practices using this model are addressed.
Ph.D.
Department of Civil and Environmental Engineering
Engineering and Computer Science
Civil Engineering

La stabilité des exploitations à ciel ouvert (mines ou carrières) excavées dans des massifs rocheux dépend de leur géométrie, des caractéristiques géométriques de la fracturation (orientation et espacement) du massif et des caractéristiques mécaniques du massif et des discontinuités. L'évaluation du risque d'instabilité rocheuse dépend de la quantité et la qualité des informations disponibles sur le massif rocheux et des méthodes d'analyse mises en œuvre pour évaluer le comportement mécanique de l'assemblage de blocs rocheux. Différents sites de talus naturels, d'excavation routière et de carrières et mines à ciel ouvert présentés dans le mémoire montrent que les informations connues peuvent beaucoup varier d'un site à l'autre. Les étapes de collecte et de synthèse des données, puis de modélisation afin d'analyser la stabilité d'une pente rocheuse conduisent à des incertitudes. Le mémoire traite plus particulièrement de l'influence des incertitudes géométriques et différentes méthodes de regroupement des fractures en familles sont explorées. La combinaison du travail proposé par différents auteurs conduit à une nouvelle méthode nommée PSMY. Cette méthode ainsi que la méthode spectrale ont été programmées sous l'environnement Mathematica et les résultats sont comparés à des regroupements "manuels" réalisé avec le logiciel DIPS. Les méthodes de regroupement sont présentées selon le pourcentage d'orientations classé par la méthode. L’orientation et l’espacement des familles de fractures sont ajustées à des lois statistiques. Les paramètres statistiques sont comparés en fonction de la méthode regroupement. Ces paramètres influencent la construction d'un modèle géométrique du massif rocheux appelé DFN (Discrete Fracture Network). L'influence des regroupements sur l'analyse de stabilité est étudiée à partir de modélisation stochastiques à l'équilibre limite utilisant les logiciels SWEDGE et RESOBLOK. Ces logiciels ne prennent pas en compte les mêmes incertitudes et en sortie ils fournissent des indicateurs de stabilité différents. Une analyse de sensibilité des indicateurs (nombre de blocs instables, volume moyen du bloc instable, volume total instable) aux méthodes de regroupement et aux orientations de talus est réalisée. Une analyse de variance permet de préciser l'influence de ces deux facteurs. Une évaluation de l’état de stabilité global du massif, en fonction de la cohésion et de l'angle de frottement est proposée. L'influence de la méthode de modélisation est évoquée en comparant des calculs tridimensionnels à l'équilibre limite et des modélisations bidimensionnel et tridimensionnels en éléments discrets rigides ou déformables. Un couplage entre le logiciel RESOBLOK (équilibre limite) et LMGC90 (éléments discrets) permet de comparer les analyses pour une même géométrie. Pour le cas de la déviation d'Ax-les-Thermes et certaines configurations géométriques instables l’influence des paramètres de modèles sur stabilité du talus est testée. Plusieurs variantes sont comparées. La simulation d’une excavation en 3D multi phases est réalisée et l’indice de mobilisation est étudié pour comparer les divers types de contacts dans LMGC90 par rapport à l’éventuel glissement des blocs dans les étapes successives
The stability of open-cast operations (mines or quarries), excavated in rock mass depends on their geometry, the geometrical characteristics of the rock mass fractures (orientation and spacing) and the mechanical characteristics of the rock mass and the discontinuities. The assessment of the rock instability risk depends on the quantity and quality of the available information on the rock mass and the analysis methods used for the evaluation of the mechanical behavior of the rock block collection. Different sites of natural rock slopes, of road cuts, and of open-cast mines and quarries, described in the present document, show that the known information can vary a lot from one site to the other. The steps of data gathering and analysis, then the modelling step used to analyze the stability of the rock slope, lead to uncertainties. The present work deals in particular with the influence of the geometrical uncertainties, and the different clustering methods, to define families of fractures, are examined. The combination of works suggested by several authors leads to a new method called PSMY. This method, together with the spectral method, was coded in the Mathematica platform, and the obtained results are compared with "hand-made" clusters, done with the DIPS software. The clustering methods are presented according to the ratio of classified orientations. The orientation and spacing of fracture families are fitted by statistical law. The statistical parameters are compared according to the clustering method used. These parameters have an influence on the construction of the rock mass geometrical model, called DFN (Discrete Fracture Network). The influence of the clusters on the stability analysis is studied from the stochastic models based on the limit equilibrium analysis, in the SWEDGE and RESOBLOK software. These programs do not take into account the same uncertainties, and as a result, they give different stability indicators. A sensitivity analysis of these indicators (number of unstable blocks, average volume of unstable blocks, and total volume of unstable blocks) versus the clustering methods used, and the orientation of the slopes, is carried out. A variance analysis allows an evaluation of the influence of these factors. The assessment of the global stability condition of the rock mass, depending on the cohesion and the friction angle, is proposed. The influence of the modeling method is analyzed by comparing 3D calculations using the limit equilibrium calculations, and 2D and 3D models using discrete rigid and deformable blocks. A coupling between RESOBLOK (limit equilibrium) and LMGC90 (discrete elements) allows the comparison of results on the same original geometry. For the case of Ax-les-Thermes road-cut, and for various unstable geometries, the influence of the model parameters is tested. Several cases are compared. The 3D simulation of an excavation, at different steps, is performed, and the mobilization index is studied, in order to compare several types of contacts, within the LMGC model, in relation to the possible sliding of blocks, at those different steps

Parmi tous les jeux de société joués par les humains au cours de l’histoire, le jeu de go était considéré comme l’un des plus difficiles à maîtriser par un programme informatique [Van Den Herik et al., 2002]; Jusqu’à ce que ce ne soit plus le cas [Silveret al., 2016]. Cette percée révolutionnaire [Müller, 2002, Van Den Herik et al., 2002] fût le fruit d’une combinaison sophistiquée de Recherche arborescente Monte-Carlo et de techniques d’apprentissage automatique pour évaluer les positions du jeu, mettant en lumière le grand potentiel de l’apprentissage automatique pour résoudre des jeux. L’apprentissage antagoniste, un cas particulier de l’optimisation multiobjective, est un outil de plus en plus utile dans l’apprentissage automatique. Par exemple, les jeux à deux joueurs et à somme nulle sont importants dans le domain des réseaux génératifs antagonistes [Goodfellow et al., 2014] ainsi que pour maîtriser des jeux comme le Go ou le Poker en s’entraînant contre lui-même [Silver et al., 2017, Brown andSandholm, 2017]. Un résultat classique de la théorie des jeux indique que les jeux convexes-concaves ont toujours un équilibre [Neumann, 1928]. Étonnamment, les praticiens en apprentissage automatique entrainent avec succès une seule paire de réseaux de neurones dont l’objectif est un problème de minimax non-convexe et non-concave alors que pour une telle fonction de gain, l’existence d’un équilibre de Nash n’est pas garantie en général. Ce travail est une tentative d'établir une solide base théorique pour l’apprentissage dans les jeux. La première contribution explore le théorème minimax pour une classe particulière de jeux non-convexes et non-concaves qui englobe les réseaux génératifs antagonistes. Cette classe correspond à un ensemble de jeux à deux joueurs et a somme nulle joués avec des réseaux de neurones. Les deuxième et troisième contributions étudient l’optimisation des problèmes minimax, et plus généralement, les inégalités variationnelles dans le cadre de l’apprentissage automatique. Bien que la méthode standard de descente de gradient ne parvienne pas à converger vers l’équilibre de Nash de jeux convexes-concaves simples, il existe des moyens d’utiliser des gradients pour obtenir des méthodes qui convergent. Nous étudierons plusieurs techniques telles que l’extrapolation, la moyenne et la quantité de mouvement à paramètre négatif. La quatrième contribution fournit une étude empirique du comportement pratique des réseaux génératifs antagonistes. Dans les deuxième et troisième contributions, nous diagnostiquons que la méthode du gradient échoue lorsque le champ de vecteur du jeu est fortement rotatif. Cependant, une telle situation peut décrire un pire des cas qui ne se produit pas dans la pratique. Nous fournissons de nouveaux outils de visualisation afin d’évaluer si nous pouvons détecter des rotations dans comportement pratique des réseaux génératifs antagonistes.
Among all the historical board games played by humans, the game of go was considered one of the most difficult to master by a computer program [Van Den Heriket al., 2002]; Until it was not [Silver et al., 2016]. This odds-breaking break-through [Müller, 2002, Van Den Herik et al., 2002] came from a sophisticated combination of Monte Carlo tree search and machine learning techniques to evaluate positions, shedding light upon the high potential of machine learning to solve games. Adversarial training, a special case of multiobjective optimization, is an increasingly useful tool in machine learning. For example, two-player zero-sum games are important for generative modeling (GANs) [Goodfellow et al., 2014] and mastering games like Go or Poker via self-play [Silver et al., 2017, Brown and Sandholm,2017]. A classic result in Game Theory states that convex-concave games always have an equilibrium [Neumann, 1928]. Surprisingly, machine learning practitioners successfully train a single pair of neural networks whose objective is a nonconvex-nonconcave minimax problem while for such a payoff function, the existence of a Nash equilibrium is not guaranteed in general. This work is an attempt to put learning in games on a firm theoretical foundation. The first contribution explores minimax theorems for a particular class of nonconvex-nonconcave games that encompasses generative adversarial networks. The proposed result is an approximate minimax theorem for two-player zero-sum games played with neural networks, including WGAN, StarCrat II, and Blotto game. Our findings rely on the fact that despite being nonconcave-nonconvex with respect to the neural networks parameters, the payoff of these games are concave-convex with respect to the actual functions (or distributions) parametrized by these neural networks. The second and third contributions study the optimization of minimax problems, and more generally, variational inequalities in the context of machine learning. While the standard gradient descent-ascent method fails to converge to the Nash equilibrium of simple convex-concave games, there exist ways to use gradients to obtain methods that converge. We investigate several techniques such as extrapolation, averaging and negative momentum. We explore these techniques experimentally by proposing a state-of-the-art (at the time of publication) optimizer for GANs called ExtraAdam. We also prove new convergence results for Extrapolation from the past, originally proposed by Popov [1980], as well as for gradient method with negative momentum. The fourth contribution provides an empirical study of the practical landscape of GANs. In the second and third contributions, we diagnose that the gradient method breaks when the game’s vector field is highly rotational. However, such a situation may describe a worst-case that does not occur in practice. We provide new visualization tools in order to exhibit rotations in practical GAN landscapes. In this contribution, we show empirically that the training of GANs exhibits significant rotations around Local Stable Stationary Points (LSSP), and we provide empirical evidence that GAN training converges to a stable stationary point, which is a saddle point for the generator loss, not a minimum, while still achieving excellent performance.