Convexidade generalizada em problemas de controle ótimo com tempo livre

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Condições de otimidade para problemas de controle ótimo com condições de contorno funcionais

Pós-graduação em Matemática - IBILCE

Incerteza intervalar em otimização e controle

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

On Properness of Impulsive Extension

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Estudo da dinâmica de populações utilizando a técnica de polinômios de colocação

Pós-graduação em Engenharia Elétrica - FEB

Switching-off in a minimal time a fitzhugh-nagumo system

This paper deals with a system that describes an electrical circuitcomposed by a linear system coupled to a nonlinear one involving a tunneldiode in a flush-and-fill circuit. One of the most comprehensive models for thiskind of circuits was introduced by R. Fitzhugh in 1961, when taking on carebiological tasks. The equation has in its phase plane only two periodic solutions,namely, the unstable singular point S0 and the stable cycle Γ. If the system isat rest on S0, the natural flow of orbits seeks to switch-on the process by going- as time goes by - toward its steady-state, Γ. By using suitable controls it ispossible to reverse such natural tendency going in a minimal time from Γ toS0, switching-off in this way the system. To achieve this goal it is mandatorya minimal enough strength on controls. These facts will be shown by means ofconsiderations on the null control sets in the process.

The dynamic behavior of a parametrically excited time-periodic MEMS taking into account parametric errors

Micro-electromechanical systems (MEMS) are micro scale devices that are able to convert electrical energy into mechanical energy or vice versa. In this paper, the mathematical model of an electronic circuit of a resonant MEMS mass sensor, with time-periodic parametric excitation, was analyzed and controlled by Chebyshev polynomial expansion of the Picard interaction and Lyapunov-Floquet transformation, and by Optimal Linear Feedback Control (OLFC). Both controls consider the union of feedback and feedforward controls. The feedback control obtained by Picard interaction and Lyapunov-Floquet transformation is the first strategy and the optimal control theory the second strategy. Numerical simulations show the efficiency of the two control methods, as well as the sensitivity of each control strategy to parametric errors. Without parametric errors, both control strategies were effective in maintaining the system in the desired orbit. On the other hand, in the presence of parametric errors, the OLFC technique was more robust.

Aproximação para Problema de Controle Ótimo Impulsivo e Problema de Tempo Mínimo sobre Domínios Estratificados

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Condições de otimalidade para controle ótimo via formalismo de Dubovitskii-Milyutin

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Estudo e desenvolvimento de filmes espessos de CeO2 obtidos por tape casting

Tape Casting proved to be an effective method for the production of thick films of CeO2 pure and doped with La. For this study, the nanoparticles used to form the slurry were synthesizes by the H-M method, at 100°C for 8 minutes, using KOH mineralizer. The slurry was made in aqueous solvent, requiring optimal control of surroundings conditions so that the produced tape has conditions to be studied. However, there's no toxicity or flammability in the film made by such solvent, being pleasing to the environment. The structural, optical and electrical properties of the films obtained by the Tape Casting process were studied by the methods of X-ray diffraction, scanning electron microscopy, specific surface area, Ultraviolet-visible spectroscopy and voltage-current measures, varying the electric field and frequency. From the results obtained by laboratory experiments, based on the literature, it was possible to reveal and understand some CeO2 features pure and doped with La

Estudo e desenvolvimento de filmes espessos de CeO2 obtidos por tape casting

Tape Casting proved to be an effective method for the production of thick films of CeO2 pure and doped with La. For this study, the nanoparticles used to form the slurry were synthesizes by the H-M method, at 100°C for 8 minutes, using KOH mineralizer. The slurry was made in aqueous solvent, requiring optimal control of surroundings conditions so that the produced tape has conditions to be studied. However, there's no toxicity or flammability in the film made by such solvent, being pleasing to the environment. The structural, optical and electrical properties of the films obtained by the Tape Casting process were studied by the methods of X-ray diffraction, scanning electron microscopy, specific surface area, Ultraviolet-visible spectroscopy and voltage-current measures, varying the electric field and frequency. From the results obtained by laboratory experiments, based on the literature, it was possible to reveal and understand some CeO2 features pure and doped with La

A general framework for multi-compartmental analysis of drug chemotherapy dynamics in human immunodeficiency virus type-1 infected individuals

An optimal control strategy for the highly active antiretroviral therapy associated to the acquired immunodeficiency syndrome should be designed regarding a comprehensive analysis of the drug chemotherapy behavior in the host tissues, from major viral replication sites to viral sanctuary compartments. Such approach is critical in order to efficiently explore synergistic, competitive and prohibitive relationships among drugs and, hence, therapy costs and side-effect minimization. In this paper, a novel mathematical model for HIV-1 drug chemotherapy dynamics in distinct host anatomic compartments is proposed and theoretically evaluated on fifteen conventional anti-retroviral drugs. Rather than interdependence between drug type and its concentration profile in a host tissue, simulated results suggest that such profile is importantly correlated with the host tissue under consideration. Furthermore, the drug accumulative dynamics are drastically affected by low patient compliance with pharmacotherapy, even when a single dose lacks. (C) 2012 Elsevier Inc. All rights reserved.

Management of diabetes mellitus and associated cardiovascular risk factors in Brazil – the Brazilian study on the practice of diabetes care

Abstract Background The Brazilian Study on the Practice of Diabetes Care main objective was to provide an epidemiological profile of individuals with type 1 and 2 diabetes mellitus (DM) in Brazil, concerning therapy and adherence to international guidelines in the medical practice. Methods This observational, cross-sectional, multicenter study collected and analyzed data from individuals with type 1 and 2 DM attending public or private clinics in Brazil. Each investigator included the first 10 patients with type 2 DM who visited his/her office, and the first 5 patients with type 1 DM. Results A total of 1,358 patients were analyzed; 375 (27.6%) had type 1 and 983 (72.4%) had type 2 DM. Most individuals were women, Caucasian, and private health care users. High prevalence rates of hypertension, dyslipidemia and central obesity were observed, particularly in type 2 DM. Only 7.3% and 5.1% of the individuals with types 1 and 2 DM, respectively, had optimal control of blood pressure, plasma glucose and lipids. The absence of hypertension and female sex were associated with better control of type 1 DM and other cardiovascular risk factors. In type 2 DM, older age was also associated with better control. Conclusions Female sex, older age, and absence of hypertension were associated with better metabolic control. An optimal control of plasma glucose and other cardiovascular risk factors are obtained only in a minority of individuals with diabetes. Local numbers, compared to those from other countries are worse.

Black-scholes type equations

In this work we are concerned with the analysis and numerical solution of Black-Scholes type equations arising in the modeling of incomplete financial markets and an inverse problem of determining the local volatility function in a generalized Black-Scholes model from observed option prices. In the first chapter a fully nonlinear Black-Scholes equation which models transaction costs arising in option pricing is discretized by a new high order compact scheme. The compact scheme is proved to be unconditionally stable and non-oscillatory and is very efficient compared to classical schemes. Moreover, it is shown that the finite difference solution converges locally uniformly to the unique viscosity solution of the continuous equation. In the next chapter we turn to the calibration problem of computing local volatility functions from market data in a generalized Black-Scholes setting. We follow an optimal control approach in a Lagrangian framework. We show the existence of a global solution and study first- and second-order optimality conditions. Furthermore, we propose an algorithm that is based on a globalized sequential quadratic programming method and a primal-dual active set strategy, and present numerical results. In the last chapter we consider a quasilinear parabolic equation with quadratic gradient terms, which arises in the modeling of an optimal portfolio in incomplete markets. The existence of weak solutions is shown by considering a sequence of approximate solutions. The main difficulty of the proof is to infer the strong convergence of the sequence. Furthermore, we prove the uniqueness of weak solutions under a smallness condition on the derivatives of the covariance matrices with respect to the solution, but without additional regularity assumptions on the solution. The results are illustrated by a numerical example.

Reinforcement Lernen mit Regularisierungsnetzwerken

Die Arbeit behandelt das Problem der Skalierbarkeit von Reinforcement Lernen auf hochdimensionale und komplexe Aufgabenstellungen. Unter Reinforcement Lernen versteht man dabei eine auf approximativem Dynamischen Programmieren basierende Klasse von Lernverfahren, die speziell Anwendung in der Künstlichen Intelligenz findet und zur autonomen Steuerung simulierter Agenten oder realer Hardwareroboter in dynamischen und unwägbaren Umwelten genutzt werden kann. Dazu wird mittels Regression aus Stichproben eine Funktion bestimmt, die die Lösung einer "Optimalitätsgleichung" (Bellman) ist und aus der sich näherungsweise optimale Entscheidungen ableiten lassen. Eine große Hürde stellt dabei die Dimensionalität des Zustandsraums dar, die häufig hoch und daher traditionellen gitterbasierten Approximationsverfahren wenig zugänglich ist. Das Ziel dieser Arbeit ist es, Reinforcement Lernen durch nichtparametrisierte Funktionsapproximation (genauer, Regularisierungsnetze) auf -- im Prinzip beliebig -- hochdimensionale Probleme anwendbar zu machen. Regularisierungsnetze sind eine Verallgemeinerung von gewöhnlichen Basisfunktionsnetzen, die die gesuchte Lösung durch die Daten parametrisieren, wodurch die explizite Wahl von Knoten/Basisfunktionen entfällt und so bei hochdimensionalen Eingaben der "Fluch der Dimension" umgangen werden kann. Gleichzeitig sind Regularisierungsnetze aber auch lineare Approximatoren, die technisch einfach handhabbar sind und für die die bestehenden Konvergenzaussagen von Reinforcement Lernen Gültigkeit behalten (anders als etwa bei Feed-Forward Neuronalen Netzen). Allen diesen theoretischen Vorteilen gegenüber steht allerdings ein sehr praktisches Problem: der Rechenaufwand bei der Verwendung von Regularisierungsnetzen skaliert von Natur aus wie O(n**3), wobei n die Anzahl der Daten ist. Das ist besonders deswegen problematisch, weil bei Reinforcement Lernen der Lernprozeß online erfolgt -- die Stichproben werden von einem Agenten/Roboter erzeugt, während er mit der Umwelt interagiert. Anpassungen an der Lösung müssen daher sofort und mit wenig Rechenaufwand vorgenommen werden. Der Beitrag dieser Arbeit gliedert sich daher in zwei Teile: Im ersten Teil der Arbeit formulieren wir für Regularisierungsnetze einen effizienten Lernalgorithmus zum Lösen allgemeiner Regressionsaufgaben, der speziell auf die Anforderungen von Online-Lernen zugeschnitten ist. Unser Ansatz basiert auf der Vorgehensweise von Recursive Least-Squares, kann aber mit konstantem Zeitaufwand nicht nur neue Daten sondern auch neue Basisfunktionen in das bestehende Modell einfügen. Ermöglicht wird das durch die "Subset of Regressors" Approximation, wodurch der Kern durch eine stark reduzierte Auswahl von Trainingsdaten approximiert wird, und einer gierigen Auswahlwahlprozedur, die diese Basiselemente direkt aus dem Datenstrom zur Laufzeit selektiert. Im zweiten Teil übertragen wir diesen Algorithmus auf approximative Politik-Evaluation mittels Least-Squares basiertem Temporal-Difference Lernen, und integrieren diesen Baustein in ein Gesamtsystem zum autonomen Lernen von optimalem Verhalten. Insgesamt entwickeln wir ein in hohem Maße dateneffizientes Verfahren, das insbesondere für Lernprobleme aus der Robotik mit kontinuierlichen und hochdimensionalen Zustandsräumen sowie stochastischen Zustandsübergängen geeignet ist. Dabei sind wir nicht auf ein Modell der Umwelt angewiesen, arbeiten weitestgehend unabhängig von der Dimension des Zustandsraums, erzielen Konvergenz bereits mit relativ wenigen Agent-Umwelt Interaktionen, und können dank des effizienten Online-Algorithmus auch im Kontext zeitkritischer Echtzeitanwendungen operieren. Wir demonstrieren die Leistungsfähigkeit unseres Ansatzes anhand von zwei realistischen und komplexen Anwendungsbeispielen: dem Problem RoboCup-Keepaway, sowie der Steuerung eines (simulierten) Oktopus-Tentakels.