A variety of practical network optimization problems arising in the context of public supply and commercial transportation, emergency response and risk management, engineering design, and industrial planning are addressed in this study. The decisions to be made in these problems include the location of supply centers, the routing, allocation and scheduling of flow between supply and demand locations, and the design of links in the network. This study is concerned with the development of optimization models and the analysis of five such problems, and the subsequent design and testing of exact and heuristic algorithms for solving these various network optimization problems. The first problem addressed is the time-dependent shortest pair of disjoint paths problem. We examine computational complexity issues, models, and algorithms for the problem of finding a shortest pair of disjoint paths between two nodes of a network such that the total travel delay is minimized, given that the individual arc delays are time-dependent. It is shown that this problem, and many variations of it, are nP-Hard and a 0-1 linear programming model that can be used to solve this problem is developed. This model can accommodate various degrees of disjointedness of the pair of paths, from complete to partial with respect to specific arcs. Next, we examine a minimum-risk routing problem and pursue the development, analysis, and testing of a mathematical model for determining a route that attempts to reduce the risk of low probability-high consequence accidents related with the transportation of hazardous materials (hazmat). More specifically, the problem addressed in this study involves finding a path that minimizes the conditional expectation of a consequence, given that an accident occurs, subject to the expected value of the consequence being lesser than or equal to a specified level n, and the probability of an accident on the path being also constrained to be no more than some value h. Various insights into related modeling issues are also provided. The values n and h are user-prescribed and could be prompted by the solution of shortest path problems that minimize the respective corresponding linear risk functions. The proposed model is a discrete, fractional programming problem that is solved using a specialized branch-and-bound approach. The model is also tested using realistic data associated with a case concerned with routing hazmat through the roadways of Bethlehem, Pennsylvania. The third problem deals with the development of a resource allocation strategy for emergency and risk management. An important and novel issue addressed in modeling this problem is the effect of loss in coverage due to the non-availability of emergency response vehicles that are currently serving certain primary incidents. This is accommodated within the model by including in the objective function a term that reflects the opportunity cost for serving an additional incident that might occur probabilistically on the network. A mixed-integer programming model is formulated for the multiple incident - multiple response problem, and we show how its solution capability can be significantly enhanced by injecting a particular structure into the constraints that results in an equivalent alternative model representation. Furthermore, for certain special cases of the MIMR problem, efficient polynomial-time solution approaches are prescribed. An algorithmic module composed of these procedures, and used in concert with a computationally efficient LP-based heuristic scheme that is developed, has been incorporated into an area-wide incident management decision support system (WAIMSS) at the Center for Transportation Research, Virginia Tech. The fourth problem addressed in this study deals with the development of global optimization algorithms for designing a water distribution network, or expanding an already existing one, that satisfies specified flow demands at stated pressure head requirements. The nonlinear, nonconvex network problem is transformed into the space of certain design variables. By relaxing the nonlinear constraints in the transformed space via suitable polyhedral outer approximations and applying the Reformulation-Linearization Technique (RLT), a tight linear lower bounding problem is derived. This problem provides an enhancement and a more precise representation of previous lower bounding relaxations that use similar approximations. Computational experience on three standard test problems from the literature is provided. For all these problems, a proven global optimal solution within a tolerance of 10 -4 % and/or within 1$ of optimality is obtained. For the two larger instances dealing with the Hanoi and New York test networks that have been open for nearly three decades, the solutions derived represent significant improvements, and the global optimality has been verified at the stated level of accuracy for these problems for the very first time in the literature. A new real network design test problem based on the Town of Blacksburg Water Distribution System is also offered to be included in the available library of test cases, and related computational results on deriving global optimal solutions are presented. The final problem addressed in this study is concerned with a global optimization approach for solving capacitated Euclidean distance multifacility location-allocation problems, as well as the development of a new algorithm for solving the generalized lp distance location-allocation problem. There exists no global optimization algorithm that has been developed and tested for this class of problems, aside from a total enumeration approach. Beginning with the Euclidean distance problem, we design depth-first and best-first branch-and-bound algorithms based on a partitioning of the allocation space that finitely converges to a global optimum for this nonconvex problem. For deriving lower bounds at node subproblems in these partial enumeration schemes, we employ two types of procedures. The first approach computes a lower bound via a simple projected location space lower bounding (PLSB) subproblem. The second approach derives a significantly enhanced lower bound by using a Reformulation-Linearization Technique (RLT) to transform an equivalent representation of the original nonconvex problem into a higher dimensional linear programming relaxation. In addition, certain cut-set inequalities generated in the allocation space, objective function based cuts derived in the location space, and tangential linear supporting hyperplanes for the distance function are added to further tighten the lower bounding relaxation. The RLT procedure is then extended to the.general lp distance problem for 1 < p < 2. Various issues related to the selection of branching variables, the design of heuristics via special selective backtracking mechanisms, and the study of the sensitivity of the proposed algorithm to the value of p in the lp - norm, are computationally investigated. Computational experience is also provided on a set of test problems to investigate both the PLSB and the RLT-lower bounding schemes. The results indicate that the proposed global optimization approach using the RLT-based scheme offers a promising viable solution procedure. In fact, among the problems solved, for the only two test instances previously available in the literature for the Euclidean distance case that were posed in 1979, we report proven global optimal solutions within a tolerance of 0.1% for the first time. It is hoped that the modeling, analysis, insights, and concepts provided for these various network based problems that arise in diverse routing, location, distribution, and design contexts, will provide guidelines for studying many other problems that arise in related situations. / Ph. D.

Simulation, whether in-flight or ground-based, is an invaluable tool for testing and evaluating aircraft. Classically, a simulation model is specific to a single particular airframe, only able to model those flying characteristics. Vast information can be gained from a simulation that is able to model a wide range of aircraft, through a comparison of the performance of these aircraft. Such a variable stability simulation model was created based on 46 stability parameters, including natural frequencies, damping ratios, time constants, and gains. The simulation was obtained using transfer functions representing the aircraft state responses to control inputs. These transfer functions were converted into state space systems used to create the linear equations for the model. The model was first developed as a desktop simulation and then converted for use with the Virginia Tech's 2F122A flight simulator. This conversion required a simple dynamic inversion of the body axis force and moment terms. To reduce the error in these terms, a model following scheme was incorporated. A series of canned inputs and real-time pilot-in-the-loop tests were flown to evaluate the variable stability model. Results in this paper have demonstrated the successful creation of a variable stability simulation model. / Master of Science

The containership design problem involves optimizing a nonconvex objective function over a design space that is restricted by a set of constraints defined in terms of nonconvex functions. An application of standard nonlinear optimization methods to such a problem can at best attain a local optimum that need not be a global optimum. This thesis investigates the application of alternative modeling, approximation, and global optimization techniques for developing a multidisciplinary approach to the containership design problem. The problem involves five design variables, which prioritized according to their relative importance in the model are: design draft, depth at side, speed, overall length, and maximum beam. Five constraints are imposed on the design, viz., an equality constraint to enforce the balance between the weight and the displacement, a linear inequality constraint on the length to depth ratio that is implied by the lightship weight formulation for the design to be acceptable, an inequality constraint on the metacentric height to ensure that the design satisfies the Coast Guard wind heel criterion, an inequality on the freeboard to ensure the minimum required freeboard governed by the code of federal regulations for freeboard (46 CFR 42), and an inequality constraint on the rolling period to ensure that the design satisfies the minimum required rolling period criterion. The objective function employed is the required freight rate, expressed in dollars per metric ton per nautical mile in order to recover annualized construction and operational costs. The model also accommodates various practical issues in a manner suitable to improve its representability. For example, it takes into account the discrete container stowage issue. The carrying capacity (number of containers) is expressed as a continuous function of the principal dimensions by using a linear response surface fit that in turn makes the objective function continuous. The weight-displacement balance is maintained by including draft as a design variable and imposing an equality constraint on the weight and displacement rather than introducing an internal loop to calculate draft at each iteration. This speeds up the optimization process. Also, the weight is formulated independent of the draft to ensure independence of the weight and the displacement, which simplifies the optimization process. The time for loading and unloading containers at a given port is a function of the number of cranes available. The number of cranes is formulated as a function of the length of the ship, and the resulting expression is made continuous through a linear response surface fit. To solve this problem, we design two approaches based on employing a sequence of polynomial programming approximations, each within two alternative branch-and-bound frameworks. In the first approach, we construct a polynomial programming approximation to the containership design problem using the Response Surface Methodology (RSM) and solve this model to global optimality using the software package BARON (Branch-and-Reduce Optimization Navigator - see Sahinidis, 1996), although the Reformulation-Linearization Technique (RLT)-based procedure of Sherali and Tuncbilek (1992, 1997) offers a viable alternative (BARON itself incorporates some elements of the latter approach). The resulting solution is refined by the application of a local search method. This procedure is integrated into two alternative branch-and-bound frameworks. The motivation is that the solution of the nonconvex polynomial approximations is likely to yield solutions in the near vicinity of the true underlying global optimum, and hence, the application of a local search method initiated at such a solution has a greater prospect of detecting such a global optimum. In the second approach, we utilize a continuous-space branch-and-bound procedure based on linear programming (LP) relaxations. These relaxations are generated through an approximation scheme that first utilizes RSM to derive polynomial approximations to the objective function and the constraints, and then applies the RLT to obtain an LP relaxation. The initial stage of this lower bounding step generates a tight, nonconvex polynomial programming relaxation for the problem, and the subsequent step constructs an LP relaxation to the resulting polynomial program via a suitable RLT procedure. The underlying motivation for these two steps is to generate a tight outer approximation of the convex envelope of the objective function over the convex hull of the feasible region. The solution obtained using the polynomial approximations is treated as a lower bound. A local search method is applied to this solution to compute an upper bound. This bounding step is then integrated into two alternative branch-and-bound frameworks. The node partitioning schemes are especially designed so that the gaps resulting from these two levels of approximations are induced to approach zero in the limit, thereby ensuring convergence to a (near) global optimum. A comparison of the containership design obtained from the designed algorithmic approaches with that obtained from the application of the nonlinear optimization methods as in previous research, exhibits a significant improvement in the design parameters translating to a significant amount of annual cost savings. For a typical containership of the size pertaining to a test case addressed in this work, having a gross weight of 90,000 metric tons, an annual transportation capacity of 99,000 containers corresponding to an annual deadweight of 1,188,000 metric tons, and logging 119,000 nautical miles annually, the improvement in the prescribed design translates to an annual estimated savings of $ 1,862,784 (or approximately $ 1.86 million) and an estimated 27 % increase in the return on investment over the life of the ship. The main contribution of this research is that it develops a detailed formulation and a more precise model of the containership design problem, along with suitable response surface and global optimization methodologies for prescribing an improved modeling and algorithmic approach for the highly nonconvex containership design problem. / Master of Science

A recent development in the design of control system for a jet engine is to use a suitable, fast and accurate model running on board. Development of linear models is particularly important as most engine control designs are based on linear control theory. Engine control performance can be significantly improved by increasing the accuracy of the developed model. Current state-of-the-art is to use piecewise linear models at selected equilibrium conditions for the development of set point controllers, followed by scheduling of resulting controller gains as a function of one or more of the system states. However, arriving at an effective gain scheduler that can accommodate fast transients covering a wide range of operating points can become quite complex and involved, thus resulting in a sacrifice on controller performance for its simplicity. This thesis presents a methodology for developing a control oriented analytical linear model of a jet engine at both equilibrium and off-equilibrium conditions. This scheme requires a nonlinear engine model to run onboard in real time. The off-equilibrium analytical linear model provides improved accuracy and flexibility over the commonly used piecewise linear models developed using numerical perturbations. Linear coefficients are obtained by evaluating, at current conditions, analytical expressions which result from differentiation of simplified nonlinear expressions. Residualization of the fast dynamics states are utilized since the fast dynamics are typically outside of the primary control bandwidth. Analytical expressions based on the physics of the aerothermodynamic processes of a gas turbine engine facilitate a systematic approach to the analysis and synthesis of model based controllers. In addition, the use of analytical expressions reduces the computational effort, enabling linearization in real time at both equilibrium and off-equilibrium conditions for a more accurate capture of system dynamics during aggressive transient maneuvers. The methodology is formulated and applied to a separate flow twin-spool turbofan engine model in the Numerical Propulsion System Simulation (NPSS) platform. The fidelity of linear model is examined by validating against a detailed nonlinear engine model using time domain response, the normalized additive uncertainty and the nu-gap metric. The effects of each simplifying assumptions, which are crucial to the linear model development, on the fidelity of the linear model are analyzed in detail. A case study is performed to investigate the case when the current state (including both slow and fast states) of the system is not readily available from the nonlinear simulation model. Also, a simple model based control is used to illustrate benefits of using the proposed modeling approach.

Στη παρούσα διπλωματική μοντελοποιείται το σύστημα σφαίρας ράβδου. Εξάγεται το μοντέλο στο χώρο κατάστασης και υλοποιείται γραμμικοποίηση εισόδου κατάστασης και εισόδου εξόδου. Βρίσκονται προσεγγιστικοί νόμοι ελέγχου για το μη γραμμικό σύστημακαθώς παρουσιάζει ιδιομορφία. / This diploma thesis includes the analysis of the nonlinear system of a ball-beam and the design of different control laws. Initially, we present the physical system and we derive the mathematical model from the lagrange equation.The nonlinear system fails to be stable with the classic linear control laws and we try to stabilize it by input-state and inputoutput linearization. The ball beam system fails to be controlled by full state linearization and we proposed some approximations for inputoutput linearization. We describe in detail how we can derive the approximate control laws and through simulation we are capable of choosing the best suitable control law. We propose a switch-controller for the nonlinear system which is vital for systems with undefined relative degree and are not input state linearizable.

Many mechanical systems exhibit nonlinear movement and are subject to perturbations from a desired equilibrium state. These perturbations can greatly reduce the efficiency of the systems. It is therefore desirous to analyze the asymptotic stabilizability of an equilibrium solution of nonlinear systems; an excellent method of performing these analyses is through study of Jacobian linearization's and their properties. Two enlightening examples of nonlinear mechanical systems are the Simple Inverted Pendulum and the Inverted Pendulum on a Cart (PoC). These examples provide insight into both the feasibility and usability of Jacobian linearizations of nonlinear systems, as well as demonstrate the concepts of local stability, observability, controllability and detectability of linearized systems under varying parameters. Some examples of constant disturbances and effects are considered. The ultimate goal is to examine stabilizability, through both static and dynamic feedback controllers, of mechanical systems

Les modulations multiporteuses apparaissent aujourd'hui comme une technologie éprouvée pour la transmission de données à haut-débits sur des canaux pouvant être très perturbés. L'OFDM (Orthogonal Frequency Division Multiplexing) a d'ailleurs été choisie dans plusieurs normes de télécommunications (ADSL, Wi-Max, IEEE 802.11a/g/n, LTE, DVB,...). Cependant un des inconvénients de ce type de modulation est la forte variation de la puissance instantanée à transmettre. Cette propriété rend ces modulations très sensibles aux non-linéarités des composants analogiques, en particulier celles de l'amplificateur de puissance à l'émission. Or l'amplificateur de puissance est un élément déterminant dans une chaîne de communication dans la mesure où il a une influence prépondérante sur le bilan global de la transmission en termes de puissance, de rendement et de distorsion. Plus l'on souhaite que l'impact de ses non linéarités soit faible et plus son rendement est faible, et inversement. Il est donc nécessaire d'effectuer un compromis linéarité/rendement.L'objectif de la thèse est d'éviter cette détérioration du rendement tout en conservant de bonnes performances de linéarité, de surcroit pour des signaux OFDM. Pour ce faire nous proposons d'utiliser conjointement des méthodes de linéarisation (prédistorsion numérique en bande de base) et d'amélioration du rendement (envelope tracking) de l'amplificateur de puissance ainsi qu'une méthode de réduction de la dynamique du signal (active constellation extension). La prédistorsion numérique classique échouant aux fortes puissances, nous proposons une méthode d'amélioration de cette technique à ces puissances. Nos résultats sont validés par des mesures sur un amplificateur de puissance 50W. Nous proposons également une association des méthodes permettant d'améliorer simultanément les performances en terme de linéarité hors bande et de rendement en minimisant les dégradations des performances de taux d'erreur binaire. / Multi-carrier modulations appear as a well-tried technology for high-speed data transmission on potentially disrupted channels. OFDM (Orthogonal Frequency Division Multiplexing) has been chosen for that matter in several telecommunication standards (ADSL, Wi-Max, IEEE 802.11a/g/n, LTE, DVB,...). However one of the drawbacks of this modulation type is its high variation of the instantaneous power to transmit. This property makes these modulations very sensitive to the non-linearities of analog components, especially those related to power amplifiers. Yet the power amplifiers are critical elements in the communication chain as they have a major influence on the global assessment in terms of power, efficiency and distortion. More we want its non linearity impact is weak, more its efficiency is weak too. It is therefore necessary to make a trade-off between linearity and efficiency.The purpose of the thesis is to avoid this efficiency damage keeping at the same time the good linearity performance, moreover for OFDM signals. In this way we propose to jointly use a linearization technique (the base band digital predistortion) and a technique of efficiency improvement (the envelope tracking) for the power amplifier, together with a technique of signal dynamic reduction (the active constellation extension). The classic predistortion failing for high powers, we propose an improvement of this technique for these powers. Our results are validated by measurements on a 50W power amplifier. We also propose an association of the techniques allowing an improvement of the performance in terms of out-of-band linearity and efficiency, with smallbit error rate damages.

Les travaux de recherche de ce manuscrit se concentrent sur l’analyse des fréquences de vibrations des robots. Nos applications concernent plus particulièrement les architectures à cinématique parallèle. Dans un premier temps nous avons considéré les robots parallèles redondants en actionnement pour lesquels nous envisageons d’augmenter la fréquence de leurs oscillations en utilisant les efforts internes intrinsèques à ce type de structure. L’objectif est d’utiliser leur actionnement pour mettre en tension leur structure, et par conséquent, par analogie avec une corde vibrante, augmenter la fréquence de leurs oscillations. Nous avons étudié plusieurs robots plans redondants et nous montrons que dans le cadre de robots typiquement conçus pour être rigides,l’influence des efforts internes rajoutés n’a que peu d’importance. La suite de nos travaux soutient la proposition suivante : "les trajectoires très dynamiques influencent les fréquences des oscillations de la plateforme mobile". En effet, les robots parallèles quand ils sont conçus pour être légers, peuvent atteindre de grandes accélérations. Nous avons choisi de nous intéresser à l’étude de l’impact que peut avoir les effets dynamiques sur la fréquence des oscillations de la plateforme mobile de ces robots. Les robots considérés pour nos développements sont des robots parallèles plans, redondants en actionnement ou non. Nous proposons d’étudier cette influence en nous basant sur un développement au premier ordre du modèle dynamique. Cette linéarisation du modèle dynamique se veut plus complète que celles proposées dans la littérature. Nous expliquons et vérifions la validité de notre approche par une étude sur le lien entre accélération et vitesse et la fréquence d’oscillation pour les robots série PR (pendule sur glissière verticale) et RR (double pendule en rotation horizontale). Ensuite, nous généralisons notre modélisation au premier ordre et l’appliquons aux quatre robots PRR-2 PRR-3, PRR-4 et Dual-V pour voir si nous sommes capable d’en dégager une tendance concernant l’évolution des fréquences d’oscillation. Nous constatons que, en fonction des trajectoires, la dynamique a une influence faible mais visible, souvent positive sur l’augmentation des fréquences d’oscillation de la plateforme mobile. Cependant, les trajectoires et les lois horaires étant imposées, nous ne pouvons que subir cette influence. / The research work of this thesis manuscript focus on the analysis of the frequency of robots’ vibrations. Our applications mainly revolve around architectures with parallel kinematics. First we examined parallel robots which are redundant in actuation and for which we are considering an increase of their oscillations’ frequency using the internal forces inherent to this type of structure. The aim is to use their actuation is the tensioning of their structure, and consequently, by analogy with a vibrating-wire, to enhance theiroscillation frequency. We have studied several redundancy planar robots and we demonstrate that in the case of robots which are typically designed to be stiff, the impact of added internal forces is of low relevance. The continuation of our research supports the following proposal: “High dynamics trajectories have an impact on the oscillation frequency of the mobile platform.” Indeed parallel robots, when designed to be light, can reach greater accelerations. We chose to concentrate on the study of the impact that dynamic effects canhave on the oscillation frequency of those robots’ mobile platform. The robots examined for our developments are planar parallel robots whether they have redundant actuation or not. We offer to study this impact based on a prime order development of the dynamic model. This linearisation of the dynamic model is intended to be more complete than those suggested by literature. We explain and verify the validity of our approach with a study on the link between speed and oscillation frequency on PR robots (pendulum on a vertical sliding guide) and RR robots ( double pendulum rotating horizontally). Then we will generalize our first order model and apply it to the four robots ( PRR-2 PRR-3, PRR-4, and Dual-V) to see if we are able to identify a pattern regarding the evolution ofoscillation frequencies. We observe that, depending on the trajectories, the dynamics have a low but noticeable, and often positive, impact on the increase of oscillation frequency of the mobile platform. However, since the trajectories and speed input laws are imposed, we have no choice but to be subjected to this impact.

Servoposicionadores pneumáticos são sistemas com tecnologia limpa, pois utiliza o ar comprimido como fluido de trabalho, leves, baratos e apresentam boa relação entre a capacidade de carga e a potência fornecida. Contudo, eles apresentam não-linearidades inerentes ao sistema pneumático, como efeitos devidos à compressibilidade do ar, ao atrito e vazamentos. Por estes motivos os controladores lineares mostram-se ineficientes para estes sistemas e é necessário utilizar estratégias de controle mais elaboradas, como, por exemplo, controle por redes neurais, controle com estrutura variável, controle adaptativo ou baseado em linearização por realimentação. Neste trabalho, foi estudada a aplicação do método de linearização por realimentação aliada ao método de controle por realimentação de estados e projeto por alocação de pólos ao controle de um servoposicionador pneumático. A estratégia de linearização por realimentação utiliza as estimativas das não-linearidades do modelo pneumático para linearizar o comportamento do servoposicionador pneumático e permitir o uso tanto de controladores lineares como não-lineares. A análise e prova das características de estabilidade completa do sistema em malha fechada com parâmetros conhecidos foi realizada, obtendo-se a garantia da convergência dos erros de seguimento para zero. Também foi realizada a análise de robustez, com a análise do comportamento do sistema frente às incertezas dos parâmetros estimados. Simulações e ensaios experimentais foram realizados para avaliar o comportamento e a eficiência do controlador proposto. Os resultados do controlador mostram-se promissores, com uma redução de aproximadamente 50% no erro de posição no seguimento de trajetória e na parada precisa com relação às técnicas lineares usualmente aplicadas a estes sistemas. Entretanto, no posicionamento percebe-se a influência do atrito, indicando que a sua compensação deve ser considerada em futuros desenvolvimentos. / Pneumatic positioning systems are clean, lightweight, cheap and present a good rate between the payload and supply power. However, this system shows a highly non-linear behavior, caused mainly by the compression of the air and the friction force. Linear strategies do not present an efficiently control in this kind of system and an improved design technique of control is needed, as neural network control, adaptive control, variable structure control or feedback linearization. In this work, it is developed the application of a feedback linearization control scheme integrated with the state feedback and pole placement method to a pneumatic positioning system. The feedback linearization strategy uses the non-linearities estimation of the pneumatic model to linearize the pneumatic positioning system and allow the use of linear or non-linear controls technique to control the behavior of the servopositioner. In this work, an analysis of the convergence properties of the closed-loop errors of the system when the proposed controller is employed is provided. It is shown that if the parameters are known than the system presents asymptotic convergence of the tracking errors to zero. The robustness properties analysis of the controller is also presented and the system behavior with the uncertainty parameters is analysed. Simulation and experimental tests were performed to assess the behavior and the efficiency of the feedback linearization control. The result of the proposed control shows to be promising on the reduction of position errors in trajectory tracking and in steadystate behavior. The tests show the presence of a strong influence of the friction force and that the friction comparation in techniques must be studied in futher developments.

Cette thèse traite du problème de la stabilisation d'un système de lancement aéroporté (éventuellement non habité) pour satellites. Le lancement aéroporté consiste à ramener, à l'aide d'un avion, un satellite et son lanceur (fusée) à une certaine hauteur, et d'exécuter son lancement dans les airs (souvent en larguant la fusée). Ceci est similaire au lancement d'un missile par un avion chasseur. La plus grande différence réside dans le rapport de masse entre l'avion et le lanceur qui est beaucoup plus proche de l'unité (fusée lourde comparée à la masse de l'avion). Le système est composé de deux étages: le premier étage est dit avion porteur qui est un véhicule aérien automatisé. Il porte le lanceur qui constitue le deuxième étage (la fusée). Dans la première partie, sont proposées des approches de modélisation pour le système de largage pendant et après le largage. La première approche considère que la phase de séparation est instantanée, mais imparfaite. Par conséquent le système est vu comme un modèle d'aéroplane dont les variables d'état sont avec des larges conditions initiales dues à la séparation imparfaite. Une deuxième approche considère la séparation elle-même, représentée par une forte perturbation (un extrême cas) sur les forces et couples aérodynamiques du modèle au cours d'un intervalle de temps. Dans la deuxième partie, afin de stabiliser le système de largage après la séparation, la commande à intégrateur conditionnel modifié est développée dans un premier temps pour une classe des systèmes non-linéaires multi-entrées multi-sorties, avec comme point de départ la théorie introduite par Khalil et co-auteurs pour des systèmes mono entrée mono sortie. Cette commande a été ensuite étendue pour la commande à servo-compensateur conditionnel modifié pour une classe de systèmes non-linéaires multi-entrées multi-sorties. Les deux stratégies ont été appliquées pour stabiliser le système de largage pendant et après la phase de séparation. Ces techniques ont l'avantage d'être robustes et de pouvoir utiliser des modèles approximatifs. D'un autre côté, il était important d'examiner la possibilité d'obtenir de meilleures performances en utilisant de meilleurs modèles. Pour cette raison, la commande de linéarisation par bouclage dynamique a été étudiée. Finalement, les performances de toutes ces méthodes de commande (ainsi que certaines commandes de base additionnelles) ont été illustrées par des simulations sous Matlab/Simulink sur un modèle non-linéaire de F-16. / This thesis addresses the problem of the stabilization of an (unmanned) airlaunch system. Air launching consists in bringing a satellite and its launcher (rocket) to a certain height using an aircraft, and then launching it from the air (often by dropping the rocket), in a similar way of launching a missile from a fighter. The main difference is that the envisaged mass ratio is much closer to one (heavy rocket compared to aircraft mass). It is then composed of two stages: the first stage called carrier aircraft consists of an <unmanned> aerial vehicle that carries the launcher which constitutes the second stage (rocket). This thesis starts by introducing the problem and objectives, continues by presenting several approaches to model the airlaunch system, and ends by developing different advanced control methods to stabilize it after the launching phase. In the modeling part we propose a firstly approach called the initial condition model which assumes that the separation phase is instantaneous, and then the airlaunch system is composed of an aircraft model after the launching phase but with large initial conditions on its state variables, caused by a non-perfect split phase. A second approach assumes that the separation phase itself is modeled by a disturbance on aerodynamic forces and moments (from a worst case) during a time interval. In the control part a modified conditional integrator controller for a class of nonlinear multi-input multi-output systems is first developed starting from the conditional integrator theory developed by Khalil and co-workers. It is then extended to a modified conditional servocompensator control for a class of nonlinear multi-input multi-output systems. Both control strategies were then applied to stabilize the airlaunch system after the separation phase. They have the advantage of being very robust, and they don't depend so much on reliable models. Even if these control strategies gave good results, it was investigated in this thesis another control approach much more dependent on detailed and reliable models. This approach was based on dynamic feedback linearization theory, and the main idea is to obtain better performance in trade off better models. Finally, all proposed control methods (plus some standard ones) were compared and illustrated by simulations under Matlab/Simulink on a nonlinear F-16 model. These simulations have shown that the results were as expected, and that each control strategy was well fit for a particular situation.