La théorie des grandes déviations traite des comportements asymptotiques d'évènements rares. C'est le langage moderne de la physique statistique d'équilibre, qui semble offrir un cadre naturel pour une extension hors équilibre. Nous présentons dans cette thèse plusieurs applications, analytiques et numériques, de cette théorie dans différents contextes. D'abord, nous montrons comment localiser numériquement des trajectoires de chaoticité atypique de systèmes dynamiques complexes. Nous étendons ensuite l'algorithme présenté à une classe de systèmes et d'observables plus large. La deuxième partie de cette thèse montre sur un exemple comment le calcul de fonctions de grandes déviations d'un système hors équilibre peut parfois être ramené à un calcul d'équilibre. La dernière partie traite des chemins de réactions en chimie et de leur détermination numérique. Le formalisme introduit repose sur la supersymétrie de l'équation de Fokker-Planck et redonne naturellement la théorie de Morse.

The Wiggins-Holmes extension of the generalized Melnikov method (GMM) is applied to weakly damped parametrically excited cross waves with surface tension in a long rectangular wave channel in order to determine if these cross waves are chaotic. The Lagrangian density function for surface waves with surface tension is simplified by transforming the volume integrals to surface integrals and by subtracting the zero variation integrals. The Lagrangian is written in terms of the three generalized coordinates (or, equivalently the three degrees of freedom) that are the time-dependent components of the velocity potential. A generalized dissipation function is assumed to be proportional to the Stokes material derivative of the free surface. The generalized momenta are calculated from the Lagrangian and the Hamiltonian is determined from a Legendre transformation of the Lagrangian. The first order ordinary differential equations derived from the Hamiltonian are usually suitable for the application of the GMM. However, the cross wave equations of motion must be transformed in order to obtain a suspended system for the application of the GMM. Only three canonical transformations that preserve the dynamics of the cross wave equations of motion are made because of an extension of the Herglotz algorithm to nonautonomous systems. This extension includes two distinct types of the generalized Herglotz algorithm (GHA). The system of nonlinear nonautonomous evolution equations determined from Hamilton's equations of motion of the second kind are averaged in order to obtain an autonomous system. The unperturbed system is analyzed to determine hyperbolic saddle points that are connected by heteroclinic orbits The perturbed Hamiltonian system that includes surface tension satisfies the KAM nondegeneracy requirements; and the Melnikov integral is calculated to demonstrate that the motion is chaotic. For the perturbed dissipative system with surface tension, the Melnikov integral is identically zero implying that a higher dimensional GMM is necessary in order to demonstrate by the GMM that the motion is chaotic. However, numerical calculations of the largest Liapunov characteristic exponent demonstrate that the perturbed dissipative system with surface tension is also chaotic. A chaos diagram is computed in order to search for possible regions of the damping parameter and the Floquet parametric forcing parameter where chaotic motions may exist. / Graduation date: 1999

This thesis presents an effective control design methodology using a one-step-ahead prediction adaptive control law and an adaptive control law based on a Lyapunov function. These control law were applied to a highly maneuverable high performance aircraft, in particular, a modified F/A-18. An adaptive controller is developed to maneuver an aircraft at a high angle of attack even if the aircraft is required to fly over a highly nonlinear flight regime. The adaptive controller presented in this thesis is based on linear, bilinear, and nonlinear prediction models with input constraints. It is shown that the linear, bilinear, and nonlinear adaptive controllers can be constructed to minimize the given cost function or Lyapunov function with respect to the control input at each step. The control is calculated such that the system follows the reference trajectory, and such that control signal remains within its constraints. From several simulation results, the nonlinear controller is controller is better than the linear controller. A nonlinear adaptive control law based on a Lyapunov function is designed such that control inputs are smoother than for the one-step-ahead prediction adaptive controller. / Graduation date: 1994

In order to understand water quality in the coastal ocean and its effects on human health, the necessity arises to locate the sources of contaminants and track their transport throughout the ocean. Dynamical systems methods are applied to the study of transport of enterococci as an indicator of microbial concentration in the vicinity of Hobie Beach, an urban, subtropical beach in Miami, FL that is used for recreation and bathing on a daily basis. Previous studies on water quality have shown that Hobie Beach has high microbial levels despite having no known point source. To investigate the cause of these high microbial levels, a combination of measured surface drifter trajectories and numerically simulated flows in the vicinity of Hobie Beach is used. The numerically simulated flows are used to identify Lagrangian Coherent Structures (LCSs), which provide a template for transport in the study area. Surface drifter trajectories are shown to be consistent with the simulated flows and the LCS structure. LCSs are then used to explain the persistent water contamination and unusually high concentrations of microbes in the water off of this beach as compared with its neighboring beaches. From the drifter simulations, as well as field experiments, one can see that passive tracers are trapped in the area along the coastline by LCS. The Lagrangian circulation of Hobie Beach, influenced primarily by tide and land geometry causes a high retention rate of water near the shore, and can be used to explain the elevated levels of enterococci in the water.

An investigation focusing on methods of estimation of robustness of nominally linear dynamic systems with unstructured uncertainties was performed. The algorithm proposed involves the consideration of an associated system, selection, and subsequent development, of Liapunov function candidate and integration of their derivatives along the solution trajectory. A nominally linear multi-dimensional dynamic system is considered with unstructured, nonlinear, time-varying and bounded perturbations. The examples illustrate the success of the method: better estimates of the bounds, than those which results from traditional approaches were obtained. Robustness of linear, time-invariant systems subject to nonlinear, time-varying perturbations has been a matter of considerable research interest recently. Design of conventional state-feedback controllers requires knowledge of the bounds for disturbances. The knowledge of disturbance bounds is also important in adaptive control and control of nonlinear & uncertain systems. Numerous applications can be found in the fields of automation, aircraft control, manipulator trajectory control, etc. The technique for the determination of robust stability bounds proposed in this paper can be utilized effectively in computerized robust control system design. / Graduation date: 1992

It is known that a highly ordered medium allows certain wave functions to move unhindered throughout and in this manner achieve delocalization. It is also known that if one introduces disorder into a medium, wave packets will not be able to move as freely and will instead be trapped or localized. In this thesis, I have simulated a medium in which the amount of disorder can be modified and using this I have shown that the shape of the localization can be altered.

The search for deterministic chaos in economic and financial time series has attracted much interest over the past decade. However, clear evidence of chaotic structures is usually prevented by large random components in the time series. In the first part of this paper we show that even if a sophisticated algorithm estimating and testing the positivity of the largest Lyapunov exponent is applied to time series generated by a stochastic dynamical system or a return series of a stock index, the results are difficult to interpret. We conclude that the notion of sensitive dependence on initial conditions as it has been developed for deterministic dynamics, can hardly be transfered into a stochastic context. Therefore, in the second part of the paper our starting point for measuring dependencies for stochastic dynamics is a distributional characterization of the dynamics, e.g. by heteroskedastic models for economic and financial time series. We adopt a sensitivity measure proposed in the literature which is an information-theoretic measure of the distance between probability density functions. This sensitivity measure is well defined for stochastic dynamics, and it can be calculated analytically for the classes of stochastic dynamics with conditional normal distributions of constant and state-dependent variance. In particular, heteroskedastic return series models such as ARCH and GARCH models are investigated. (author's abstract) / Series: Report Series SFB "Adaptive Information Systems and Modelling in Economics and Management Science"

<p>Noise-induced stabilization occurs when an unstable deterministic system is stabilized by the addition of white noise. Proving that this phenomenon occurs for a particular system is often manifested through the construction of a global Lyapunov function. However, the procedure for constructing a Lyapunov function is often quite ad hoc, involving much time and tedium. In this thesis, a systematic algorithm for the construction of a global Lyapunov function for planar systems is presented. The general methodology is to construct a sequence of local Lyapunov functions in different regions of the plane, where the regions are delineated by different behaviors of the deterministic dynamics. A priming region, where the deterministic drift is directed inward, is first identified where there is an obvious choice for a local Lyapunov function. This priming Lyapunov function is then propagated to the other regions through a series of Poisson equations. The local Lyapunov functions are lastly patched together to form one smooth global Lyapunov function.</p><p>The algorithm is applied to a model problem which displays finite time blow up in the deterministic setting in order to prove that the system exhibits noise-induced stabilization. Moreover, the Lyapunov function constructed is in fact what we define to be a super Lyapunov function. We prove that the existence of a super Lyapunov function, along with a minorization condition, implies that the corresponding system converges to a unique invariant probability measure at an exponential rate that is independent of the initial condition.</p> / Dissertation

Based on the Lyapunov stability theorem, a modified stability analysis as well as a modified observer is proposed in this thesis for a class of uncertain nonlinear systems with an existent high gain observer. By assuming that the first two state variables are indirectly measurable, reanalyzing the stability of the error dynamics is presented first. The advantage of this modified analytic method is that the upper bound of the disturbance distribution functions is not required to be known in advance, and the asymptotic stability is still guaranteed. Next, based on this existent observer, a slightly modified observer is presented for systems with disturbances whose upper bound is unknown. An adaptive mechanism is embedded in the proposed observer, so that the upper bound of perturbations is not required to be known beforehand. The resultant dynamics of estimation errors can be driven into the sliding surface in a finite time, and guarantee asymptotic stability. A numerical example and a practical example are given to demonstrate the feasibility of the proposed observer.

The rapid growth of peer-to-peer (P2P) networks in the past few years has brought with it increases in transit cost to Internet Service Providers (ISPs), as peers exchange large amounts of traffic across ISP boundaries. This ISP oblivious behavior has resulted in misalignment of incentives between P2P networks|that seek to maximize user quality|and ISPs|that would seek to minimize costs. Can we design a P2P overlay that accounts for both ISP costs as well as quality of service, and attains a desired tradeoff between the two? We design a system, which we call MultiTrack, that consists of an overlay of multiple kinds of Trackers whose purpose it is to align these goals. We have mTrackers that form an overlay network among themselves, and split demand from users among different ISP domains while trying to minimize their individual costs (delay plus transit cost) in their ISP domain. We design the signals in this overlay of mTrackers in such a way that potentially competitive individual optimization goals are aligned across the mTrackers. The system could also have a tTracker that acts as a gateway into the system, and ensures that users who are from different ISP domains have a fair chance of being admitted into the system, while keeping costs in check. We prove analytically that our system is stable and achieves maximum utility with minimum cost. We validated our system design using Matlab simulations, and implemented the system on ns-2 in order to conduct more realistic experiments. We showed that our system significantly outperforms two types of systems, one in which user delay is the only control dimension (forwarding traffic without considering the transit prices) and a second system in which transit prices are the only control dimension (localized traffic only). Thus, we conclude that our system, that operates in two dimensions: (1) user delay and (2) transit prices, results in minimum cost and maximum utility for fixed capacity of the system.