Thesis (Ph. D.)--Industrial and Systems Engineering, Georgia Institute of Technology, 2006. / Prausnitz, Mark, Committee Member ; Vidakovic, Brani, Committee Member ; Lee, Eva, Committee Chair ; Nemhauser, George, Committee Member ; Johnson, Ellis, Committee Member. Includes bibliographical references.

Typescript. Includes bibliographical references.

Thesis (Ph. D.)--Industrial and Systems Engineering, Georgia Institute of Technology, 2008. / Nemhauser,George, Committee Member ; Gu, Zonghao, Committee Member ; Ergun, Ozlem, Committee Member ; Sokol, Joel, Committee Co-Chair ; Ahmed, Shabbir, Committee Chair.

Thesis (Ph. D.)--University of Texas at Austin, 2006. / Vita. Includes bibliographical references.

Thesis (M.S.)--University of Texas at El Paso, 2008. / Title from title screen. Vita. CD-ROM. Includes bibliographical references. Also available online.

The models for volatility, autoregressive conditional heteroscedastic (ARCH) and generalized autoregressive conditional heteroscedastic (GARCH) are discussed. Stationarity condition and forecasting for simple ARCH(1) and GARCH(1,1) models are given. The model for discrete time series is proposed to be negative binomial integer-valued GARCH model, which is a generalization of the Poisson INGARCH model. The stationarity conditions and the autocorrelation function are given. For parameter estimation, three methodologies are presented with a focus on maximum likelihood approach. Simulation study on a sample size of 100 and 500 are carried out and the results are presented. An application of the model to a real time series with numerical example is given indicating that the proposed methodology performs better than the Poisson and double Poisson model-based methods.

Master of Agribusiness / Department of Agricultural Economics / Jason S. Bergtold / Technical support for products after the sale is commonplace in today’s businesses. Original Equipment Manufacturers (OEMs) provide technical support to their dealer channel for resolution of complex product issues. Technical support staffing levels can vary by product type, product complexity, and production volumes, and case volumes. This research seeks a better understanding of appropriate staffing levels between three product lines for one OEM. The objective of this paper is to develop a model for monthly and weekly average case volumes for the three product lines, based off of historical case volume data. This data is used to predict a product line’s (platform’s) workload based off the month of the year. The output of each platform’s monthly case volume is then used in an optimization model to determine optimal staffing levels for each platform, based off the time of the year. The models developed for each platform use a linear relationship which regresses workload on a set of binary variable for the months of the year. Each of the models developed provided statistically significant coefficients for months which contain the platform’s highest workload. The outputs from these models are used in a mixed integer nonlinear programming optimization model to determine staff level of full time equivalent (FTE) employees at each platform. Each of the three scenarios utilized in this research provide similar trends and staffing levels for each of the three product lines. Results of this research are of interest for the management of technical support staffing.

Orientador: Vinicius Amaral Armentano / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-13T19:39:00Z (GMT). No. of bitstreams: 1 Takahata_AndreKazuo_M.pdf: 1113954 bytes, checksum: 18f4c96c943dced30f100b8a56c97258 (MD5) Previous issue date: 2009 / Resumo: Este trabalho trata do desenvolvimento de heurísticas de busca genéricas para obtenção de soluções de problemas de otimização combinatória formulados como modelos de programação linear inteira, com o uso do pacote de otimização XPRESS. Este é um tema recente, em que são conjugados a flexibilidade de heurísticas e os avanços dos solvers de otimização para a obtenção de soluções de alta qualidade em tempo reduzido. As heurísticas propostas são baseadas em arredondamentos gerados a partir de raios de um cone, cujo vértice é associado à solução ótima da relaxação de programação linear, e em trajetórias factíveis e infactíveis em relação à fronteira desta relaxação. A motivação para este enfoque é dada pelo apelo geométrico e no sucesso de estratégias similares em heurísticas para problemas combinatórios. O trabalho descreve a concepção e a implementação dessas heurísticas e apresenta resultados de testes em instâncias da literatura. / Abstract: In this work we develop a set of generic search heuristics for solving combinatorial optimization problems formulated as linear integer programming models, using the XPRESS optimization package. This is a recent theme, in which efforts have been made in order to combine the flexibility offered by heuristics and the expressive advances achieved in the development of optimization solvers so as to obtain high quality solutions in a short time. The proposed heuristics are based on rounding solutions located on the rays of a cone whose vertex is associated with the optimal solution of the linear programming relaxation, and in feasible and infeasible trajectories relative to the frontier of such relaxation. This approach is motivated by its geometric appeal and by the success of similar approaches in heuristics for solving combinatorial problems. This work describes the development and implementation of the heuristics and presents computational tests on instances from literature. / Mestrado / Automação / Mestre em Engenharia Elétrica

Orientador: Cid Carvalho de Souza / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Computação / Made available in DSpace on 2018-08-12T05:36:32Z (GMT). No. of bitstreams: 1 Cavalcante_VictorFernandes_D.pdf: 1501954 bytes, checksum: afd9c038eef0eb384065875b1df282ca (MD5) Previous issue date: 2008 / Resumo: Uma das principais motivações para o estudo de Otimização Discreta reside no elevado número de problemas do nosso cotidiano representáveis através de modelos de Otimização Inteira e Combinatória. Em particular, muitos destes problemas podem ser formulados com Programação Inteira 0-1, o que desperta especial interesse em técnicas capazes de resolver tais modelos. Dentre as inúmeras formas de solução atualmente disponíveis para problemas desta natureza, os algoritmos baseados na técnica de relaxação Lagrangiana surgem como uma alternativa que tem tido grande sucesso na prática. Além disso, avanços consideráveis ocorreram na área de Programação Inteira com o advento da Combinatória Poliédrica, intensificando o interesse pelos algoritmos de planos de corte. Neste contexto, esta tese tem como principal objetivo verificar as potencialidades do uso combinado de Combinatória Poliédrica e relaxação Lagrangiana na resolução de dois problemas de otimização combinatória. Mais especificamente, o presente trabalho esta focado no desenvolvimento dos chamados algoritmos relax-and-cut para o problema de particionamento de conjuntos e ao problema do separador de vértices de um grafo. Sendo assim, são propostos algoritmos que combinam relaxação Lagrangiana e planos de cortes faciais para os dois problemas sob consideração. Em ambos os casos, os resultados obtidos com os testes computacionais realizados são comparados com os melhores resultados disponíveis na literatura. Os principais resultados alcançados na tese mostram que: (a) o uso combinado de relaxação Lagrangiana e planos de corte constitui uma alternativa bastante competitiva para solucionar o problema de particionamento de conjuntos, freqüentemente superando o desempenho dos melhores algoritmos disponíveis na literatura para o problema e, (b) no caso do problema do separador de vértices, além da combinação de técnicas Lagrangianas com o uso de planos de corte, a hibridização dos algoritmos relax-and-cut e branch-andcut leva á resolução de instâncias da literatura mais rapidamente que o melhor algoritmo exato conhecido para o problema até então. / Abstract: One of the main motivations for the study of Discrete Optimization resides in the huge number of problems from our daily life that can be represented through Integer and Combinatorial Optimization models. In particular, many of these problems can be cast as 0-1 Integer Programs, which gives rise to special interest on how to solve such models. Among the several ways currently available to solve problems of this nature, the algorithms based on Lagrangian relaxation techniques appears as an alternative that has had great success in practice. Besides, noticeable achievements occurred with the advent of Polyhedral Combinatorics, intensifying the interest on cutting plane algorithms. In this context, this thesis has as its main goal to verify the potentialities of the combined usage of Polyhedral Combinatorics and Lagrangian relaxation in the resolution of two combinatorial optimization problems. More specifically, the present work is focused on the development of the so-called relax-and-cut algorithms for the set partition problem and for the vertex separator problem on graphs. Therefore, algorithms combining Lagrangian relaxation and cutting planes are proposed for the two problems under consideration. In both cases, the results obtained in the computational tests carried out are compared with the best ones available in the literature. The main results achieved in the thesis show that: (a) the combined usage of Lagrangian relaxation and cutting planes constitutes a competitive alternative to solve the set partition problem, often outperforming the best algorithms available in the literature for the problem and, (b) in the case of the vertex separator problem, besides the combination of Lagrangian techniques and cutting planes, a hybridization of the relax-and-cut and branch-and-cut algorithms lead to the resolution of instances from the literature more rapidly than the best exact algorithm known for the problem so far. / Doutorado / Doutor em Ciência da Computação

This thesis compares different mixed integer programming (MIP) formulations for sequencing of tasks in the context of avionics scheduling. Sequencing is a key concern in many discrete optimisation problems, and there are numerous ways of accomplishing sequencing with different MIP formulations. A scheduling tool for avionic systems has previously been developed in a collaboration between Saab and Linköping University. This tool includes a MIP formulation of the scheduling problem where one of the model components has the purpose to sequence tasks. In this thesis, this sequencing component is replaced with other MIP formulations in order to study whether the computational performance of the scheduling tool can be improved. Different scheduling instances and objective functions have been used when performing the tests aiming to evaluate the performances, with the computational times of the entire avionic scheduling model determining the success of the different MIP formulations for sequencing. The results show that the choice of MIP formulation makes a considerable impact on the computational performance and that a significant improvement can be achieved by choosing the most suitable one.