Epidemics have shaped, sometimes more than wars and natural disasters, demo- graphic aspects of human populations around the world, their health habits and their economies. Ebola and the Middle East Respiratory Syndrome (MERS) are clear and current examples of potential hazards at planetary scale. During the spread of an epidemic disease, there are phenomena, like the sudden extinction of the epidemic, that can not be captured by deterministic models. As a consequence, stochastic models have been proposed during the last decades. A typical forward problem in the stochastic setting could be the approximation of the expected number of infected individuals found in one month from now. On the other hand, a typical inverse problem could be, given a discretely observed set of epidemiological data, infer the transmission rate of the epidemic or its basic reproduction number. Markovian epidemic models are stochastic models belonging to a wide class of pure jump processes known as Stochastic Reaction Networks (SRNs), that are intended to describe the time evolution of interacting particle systems where one particle interacts with the others through a finite set of reaction channels. SRNs have been mainly developed to model biochemical reactions but they also have applications in neural networks, virus kinetics, and dynamics of social networks, among others. 4 This PhD thesis is focused on novel fast simulation algorithms and statistical inference methods for SRNs. Our novel Multi-level Monte Carlo (MLMC) hybrid simulation algorithms provide accurate estimates of expected values of a given observable of SRNs at a prescribed final time. They are designed to control the global approximation error up to a user-selected accuracy and up to a certain confidence level, and with near optimal computational work. We also present novel dual-weighted residual expansions for fast estimation of weak and strong errors arising from the MLMC methodology. Regarding the statistical inference aspect, we first mention an innovative multi- scale approach, where we introduce a deterministic systematic way of using up-scaled likelihoods for parameter estimation while the statistical fittings are done in the base model through the use of the Master Equation. In a di↵erent approach, we derive a new forward-reverse representation for simulating stochastic bridges between con- secutive observations. This allows us to use the well-known EM Algorithm to infer the reaction rates. The forward-reverse methodology is boosted by an initial phase where, using multi-scale approximation techniques, we provide initial values for the EM Algorithm.

No description available.

In this dissertation we study several non-convex and stochastic optimization problems. The common theme is the use of mixed-integer programming (MIP) techniques including valid inequalities and reformulation to solve these problems. We first study a strategic capacity planning model which captures the trade-off between the incentive to delay capacity installation to wait for improved technology and the need for some capacity to be installed to meet current demands. This problem is naturally formulated as a MIP with a bilinear objective. We develop several linear MIP formulations, along with classes of strong valid inequalities. We also present a specialized branch-and-cut algorithm to solve a compact concave formulation. Computational results indicate that these formulations can be used to solve large-scale instances. We next study methods for optimization with joint probabilistic constraints. These problems are challenging because evaluating solution feasibility requires multidimensional integration and the feasible region is not convex. We propose and analyze a Monte Carlo sampling scheme to simplify the probabilistic structure of such problems. Computational tests of the approach indicate that it can yield good feasible solutions and reasonable bounds on their quality. Next, we study a MIP formulation of the non-convex sample approximation problem. We obtain two strengthened formulations. As a byproduct of this analysis, we obtain new results for the previously studied mixing set, subject to an additional knapsack inequality. Computational results indicate that large-scale instances can be solved using the strengthened formulations. Finally, we study optimization problems with stochastic dominance constraints. A stochastic dominance constraint states that a random outcome which depends on the decision variables should stochastically dominate a given random variable. We present new formulations for both first and second order stochastic dominance which are significantly more compact than existing formulations. Computational tests illustrate the benefits of the new formulations.

Thesis (M.S.)--University of North Carolina Wilmington, 2009. / Title from PDF title page (February 16, 2010) Includes bibliographical references (p. 34-35)

Orientador: Paulo Regis Caron Ruffino / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Matemática Estatística e Computação Científica / Made available in DSpace on 2018-08-27T10:29:03Z (GMT). No. of bitstreams: 1 Morgado_LeandroBatista_D.pdf: 1320837 bytes, checksum: db1015f01556b3de2b1f7ca1c6bf33d3 (MD5) Previous issue date: 2015 / Resumo: Este trabalho aborda alguns aspectos da teoria de equações diferenciais estocásticas em relação a semimartingales com saltos, suas aplicações na decomposição de fluxos estocásticos em variedades, bem como algumas implicações de natureza geométrica. Inicialmente, em uma variedade munida de distribuições complementares, discutimos o problema da decomposição de fluxos estocásticos contínuos, isto é, gerados por EDE em relação ao movimento Browniano. Resultados anteriores garantem a existência de uma decomposição em difeomorfismos que preservam as distribuições até um tempo de parada. Usando a assim denominada equação de Marcus, bem como uma técnica que denominamos equação 'stop and go', vamos construir um fluxo estocástico próximo ao original, com a propriedade adicional que o fluxo construído pode ser decomposto além do tempo de parada inicial. Em seguida, trataremos da decomposição de fluxos estocásticos no caso descontínuo, isto é, para processos gerados por uma EDE em relação a um semimartingale com saltos. Após uma discussão sobre a existência da decomposição, obtemos as EDEs para as componentes respectivas, a partir de uma extensão que propomos da fórmula de Itô-Ventzel-Kunita. Finalmente, propomos um modelo de equações diferenciais estocásticas com retardo incluindo saltos. A ideia é modelar certos fenômenos em que a informação pode chegar ao receptor por diferentes canais: de forma contínua, mas com retardo, e em tempos discretos, de forma instantânea. Vamos abordar aspectos geométricos relacionados ao tema: transporte paralelo em curvas diferenciáveis com saltos, bem como possibilidade de levantamento de uma solução do nosso modelo de equação para o fibrado de bases de uma variedade diferenciável / Abstract: The main subject of this thesis is the theory of stochastic differential equations driven by semimartingales with jumps. We consider applications in the decomposition of stochastic flows in differentiable manifolds, and geometrical aspects about these equations. Initially, in a differentiable manifold endowed with a pair of complementary distributions, we discuss the decomposition of continuous stochastic flows, that is, flows generated by SDEs driven by Brownian motion. Previous results guarantee that, under some assumptions, there exists a decomposition in diffeomorphisms that preserves the distributions up to a stopping time. Using the so called Marcus equation, and a technique that we call 'stop and go' equation, we construct a stochastic flow close to the original one, with the property that the constructed flow can be decomposed further on the stopping time. After, we deal with the decomposition of stochastic flows in the discontinuous case, that is, processes generated by SDEs driven by semimartingales with jumps. We discuss the existence of this decomposition, and obtain the SDEs for the respective components, using an extension of the Itô-Ventzel-Kunita formula. Finally, we propose a model of stochastic differential equations including delay and jumps. The idea is to describe some phenomena such that the information comes to the receptor by different channels: continuously, with some delay, and in discrete times, instantaneously. We deal with geometrical aspects related with this subject: parallel transport in càdlàg curves, and lifting of solutions of these equations to the linear frame bundle of a differentiable manifold / Doutorado / Matematica / Doutor em Matemática

<p>This dissertation investigates the use of sampling methods for solving stochastic optimization problems using iterative algorithms. Two sampling paradigms are considered: (i) adaptive sampling, where, before each iterate update, the sample size for estimating the objective function and the gradient is adaptively chosen; and (ii) retrospective approximation (RA), where, iterate updates are performed using a chosen fixed sample size for as long as progress is deemed statistically significant, at which time the sample size is increased. We investigate adaptive sampling within the context of a trust-region framework for solving stochastic optimization problems in $\mathbb{R}^d$, and retrospective approximation within the broader context of solving stochastic optimization problems on a Hilbert space. In the first part of the dissertation, we propose Adaptive Sampling Trust-Region Optimization (ASTRO), a class of derivative-based stochastic trust-region (TR) algorithms developed to solve smooth stochastic unconstrained optimization problems in $\mathbb{R}^{d}$ where the objective function and its gradient are observable only through a noisy oracle or using a large dataset. Efficiency in ASTRO stems from two key aspects: (i) adaptive sampling to ensure that the objective function and its gradient are sampled only to the extent needed, so that small sample sizes are chosen when the iterates are far from a critical point and large sample sizes are chosen when iterates are near a critical point; and (ii) quasi-Newton Hessian updates using BFGS. We prove three main results for ASTRO and for general stochastic trust-region methods that estimate function and gradient values adaptively, using sample sizes that are stopping times with respect to the sigma algebra of the generated observations. The first asserts strong consistency when the adaptive sample sizes have a mild logarithmic lower bound, assuming that the oracle errors are light-tailed. The second and third results characterize the iteration and oracle complexities in terms of certain risk functions. Specifically, the second result asserts that the best achievable $\mathcal{O}(\epsilon^{-1})$ iteration complexity (of squared gradient norm) is attained when the total relative risk associated with the adaptive sample size sequence is finite; and the third result characterizes the corresponding oracle complexity in terms of the total generalized risk associated with the adaptive sample size sequence. We report encouraging numerical results in certain settings. In the second part of this dissertation, we consider the use of RA as an alternate adaptive sampling paradigm to solve smooth stochastic constrained optimization problems in infinite-dimensional Hilbert spaces. RA generates a sequence of subsampled deterministic infinite-dimensional problems that are approximately solved within a dynamic error tolerance. The bottleneck in RA becomes solving this sequence of problems efficiently. To this end, we propose a progressive subspace expansion (PSE) framework to solve smooth deterministic optimization problems in infinite-dimensional Hilbert spaces with a TR Sequential Quadratic Programming (SQP) solver. The infinite-dimensional optimization problem is discretized, and a sequence of finite-dimensional problems are solved where the problem dimension is progressively increased. Additionally, (i) we solve this sequence of finite-dimensional problems only to the extent necessary, i.e., we spend just enough computational work needed to solve each problem within a dynamic error tolerance, and (ii) we use the solution of the current optimization problem as the initial guess for the subsequent problem. We prove two main results for PSE. The first assesses convergence to a first-order critical point of a subsequence of iterates generated by the PSE TR-SQP algorithm. The second characterizes the relationship between the error tolerance and the problem dimension, and provides an oracle complexity result for the total amount of computational work incurred by PSE. This amount of computational work is closely connected to three quantities: the convergence rate of the finite-dimensional spaces to the infinite-dimensional space, the rate of increase of the cost of making oracle calls in finite-dimensional spaces, and the convergence rate of the solution method used. We also show encouraging numerical results on an optimal control problem supporting our theoretical findings.</p> <p>  </p>

In this study we are interested on the following 4-dimensional system of stochastic differential equations.dS=(-βSI+μ(K-S)+αI+ζR)dt-σ_1 SIF_1 (S,E,I,R)dW_1+σ_4 RF_4 (S,E,I,R)dW_4 dE=(βSI-(μ+η)E)dt+σ_1 SIF_1 (S,E,I,R)dW_1-σ_2 EF_2 (S,E,I,R)dW_2 dI=(ηE-(α+γ+μ)I)dt+σ_2 EF_2 (S,E,I,R)dW_2-σ_3 IF_3 (S,E,I,R)dW_3 dR=(γI-(μ+ζ)R)dt+σ_3 IF_3 (S,E,I,R)dW_3-σ_4 RF_4 (S,E,I,R)dW_4 with variance parameters σ_i≥0 and constants α,β,η,γ,μ ζ≥0. This system may be used to model the dynamics of susceptible, exposed, infected and recovering individuals subject to a present virus with state-dependent random transitions. Our main goal is to prove the existence of a bounded, unique, strong (pathwise), global solution to this system, and to discuss asymptotic stochastic and moment stability of the two equilibrium points, namely the disease free and the endemic equilibria. In this model, as suggested by our advisor, diffusion coefficients can be any local Lipschitz continuous functions on bounded domain D={(S,E,I,R)∈R_+^4:00 of maximum carrying capacity and W_i are independent and identical Wiener processes defined on a complete probability space (Ω,F,{F_t }_(t≥0),P). At the end we carry out some simulations to illustrate our results.

The stochastic SEIR(S) model with random total population is given by the system of stochastic differential equations:dS=(-βSI+μ(K-S)+αI+ζR)dt-σ_1 SIF_1 (S,E,I,R)dW_1+σ_4 RF_4 (S,E,I,R)dW_4+σ_5 S(K-N)dW_5\\ dE=(βSI-(μ+η)E)dt+σ_1 SIF_1 (S,E,I,R)dW_1-σ_2 EF_2 (S,E,I,R)dW_2+σ_5 E(K-N)dW_5 \\ dI=(ηE-(α+γ+μ)I)dt+σ_2 EF_2 (S,E,I,R)dW_2-σ_3 IF_3 (S,E,I,R)dW_3+σ_5 I(K-N)dW_5 \\ dR=(γI-(μ+ζ)R)dt+σ_3 IF_3 (S,E,I,R)dW_3-σ_4 RF_4 (S,E,I,R)dW_4+σ_5 R(K-N)dW_5, where σ_i>0 and constants α, β, η, γ, ζ, μ≥0. K represents the maximum carrying capacity for the total population and W_k=(W_k (t))_(t≥0) are independent, standard Wiener processes on a complete probability space (Ω,F,(F_t )_(t≥0),P). The SDE for the total population N=S+E+I+R has the form dN(t)=μ(K-N)dt+σ_5 N(K-N)dW_5 on D_0=(0,K). The goal of our study is to prove the existence of unique, Markovian, continuous time solutions on the 4D prism D={(S,E,I,R)∈R_+^4:0≤S, E,I,R≤K, S+E+I+R≤K}. Then using the method of Lyapunov functions we prove the asymptotic stochastic and moment stability of disease-free and endemic equilibria. Finally, we use numerical simulations to illustrate our results.

In the operation of stochastic systems, and of queues in particular, it is important to recognize quickly the development in time of situations not compatible with their design criteria. Once such an anomalous condition is detected, it has to be decided, if the occurrence of that event can be attributed to chance or is due to a change in the parameters governing the system. This procedure of tracking the system is defined as monitoring. The design of a monitor and the selection of suitable threshold regions for monitoring a single server queue are the subjects of this thesis. The notion of profile curves, useful in formalizing monitoring schemes for queues, is also discussed. Finally, some numerical examples are presented to illustrate the performance of the monitor designed.

published_or_final_version / Statistics and Actuarial Science / Doctoral / Doctor of Philosophy