Embedded system designers face multiple challenges in fulfilling the runtime requirements of programs. Effective scheduling of programs is required to extract as much parallelism as possible. These scheduling algorithms must also improve speedup after instruction-set extensions have occurred. Scheduling of dynamic code at run time is made more difficult when the static components of the program are scheduled inefficiently. This research aims to optimize a program’s static code at compile time. This is achieved with four algorithms designed to schedule code at the task and instruction level. Additionally, the algorithms improve scheduling using instruction set extended code on symmetrical homogeneous multiprocessor systems. Using these algorithms, we achieve speedups up to 3.86X over sequential execution for a 4-issue 2-processor system, and show better performance than recent heuristic techniques for small programs. Finally, the algorithms generate speedup values for a 64-point FFT that are similar to the test runs.

Monotonic optimization consists of minimizing or maximizing a monotonic objective function over a set of constraints defined by monotonic functions. Many optimization problems in economics and engineering often have monotonicity while lacking other useful properties, such as convexity. This thesis is concerned with the development and application of global optimization algorithms for monotonic optimization problems. First, we propose enhancements to an existing outer-approximation algorithm | called the Polyblock Algorithm | for monotonic optimization problems. The enhancements are shown to significantly improve the computational performance of the algorithm while retaining the convergence properties. Next, we develop a generic branch-and-bound algorithm for monotonic optimization problems. A computational study is carried out for comparing the performance of the Polyblock Algorithm and variants of the proposed branch-and-bound scheme on a family of separable polynomial programming problems. Finally, we study an important class of monotonic optimization problems | probabilistically constrained linear programs. We develop a branch-and-bound algorithm that searches for a global solution to the problem. The basic algorithm is enhanced by domain reduction and cutting plane strategies to reduce the size of the partitions and hence tighten bounds. The proposed branch-reduce-cut algorithm exploits the monotonicity properties inherent in the problem, and requires the solution of only linear programming subproblems. We provide convergence proofs for the algorithm. Some illustrative numerical results involving problems with discrete distributions are presented.

Multinomial selection problem procedures are ranking and selection techniques that aim to select the best (most probable) alternative based upon a sequence of multinomial observations. The classical formulation of the procedure design problem is to find a decision rule for terminating sampling. The decision rule should minimize the expected number of observations taken while achieving a specified indifference zone requirement on the prior probability of making a correct selection when the alternative configurations are in a particular subset of the probability space called the preference zone. We study the constrained version of the design problem in which there is a given maximum number of allowed observations. Numerous procedures have been proposed over the past 50 years, all of them suboptimal. In this thesis, we find via linear programming the optimal selection procedure for any given probability configuration. The optimal procedure turns out to be necessarily randomized in many cases. We also find via mixed integer programming the optimal non-randomized procedure. We demonstrate the performance of the methodology on a number of examples. We then reformulate the mathematical programs to make them more efficient to implement, thereby significantly expanding the range of computationally feasible problems. We prove that there exists an optimal policy which has at most one randomized decision point and we develop a procedure for finding such a policy. We also extend our formulation to replicate existing procedures. Next, we show that there is very little difference between the relative performances of the optimal randomized and non-randomized procedures. Additionally, we compare existing procedures using the optimal procedure as a benchmark, and produce updated tables for a number of those procedures. Then, we develop a methodology that guarantees the optimal randomized and non-randomized procedures for a broad class of variable observation cost functions -- the first of its kind. We examine procedure performance under a variety of cost functions, demonstrating that incorrect assumptions regarding marginal observation costs may lead to increased total costs. Finally, we investigate and challenge key assumptions concerning the indifference zone parameter and the conditional probability of correct selection, revealing some interesting implications.

Embedded system designers face multiple challenges in fulfilling the runtime requirements of programs. Effective scheduling of programs is required to extract as much parallelism as possible. These scheduling algorithms must also improve speedup after instruction-set extensions have occurred. Scheduling of dynamic code at run time is made more difficult when the static components of the program are scheduled inefficiently. This research aims to optimize a program’s static code at compile time. This is achieved with four algorithms designed to schedule code at the task and instruction level. Additionally, the algorithms improve scheduling using instruction set extended code on symmetrical homogeneous multiprocessor systems. Using these algorithms, we achieve speedups up to 3.86X over sequential execution for a 4-issue 2-processor system, and show better performance than recent heuristic techniques for small programs. Finally, the algorithms generate speedup values for a 64-point FFT that are similar to the test runs.

Submitted by Luciano Pereira Varanis (luciano_varanis@yahoo.com) on 2015-06-29T19:17:42Z No. of bitstreams: 1 Dissertação _Luciano_Varanis - Completo.pdf: 1367224 bytes, checksum: 848c44e67420ced7c19a6d2d1ab98f61 (MD5) / Approved for entry into archive by Janete de Oliveira Feitosa (janete.feitosa@fgv.br) on 2015-07-22T20:07:23Z (GMT) No. of bitstreams: 1 Dissertação _Luciano_Varanis - Completo.pdf: 1367224 bytes, checksum: 848c44e67420ced7c19a6d2d1ab98f61 (MD5) / Approved for entry into archive by Marcia Bacha (marcia.bacha@fgv.br) on 2015-07-24T17:56:00Z (GMT) No. of bitstreams: 1 Dissertação _Luciano_Varanis - Completo.pdf: 1367224 bytes, checksum: 848c44e67420ced7c19a6d2d1ab98f61 (MD5) / Made available in DSpace on 2015-07-24T17:56:24Z (GMT). No. of bitstreams: 1 Dissertação _Luciano_Varanis - Completo.pdf: 1367224 bytes, checksum: 848c44e67420ced7c19a6d2d1ab98f61 (MD5) Previous issue date: 2014-12-18 / In this dissertation we discuss models and solution methods of stochastic programs to solve ALM problmes for pension funds. We will provide the model from Drijver et al. based on Multistage Mixed-integer Stochastic Programming. A case study based on simulation for an ALM problem will be presented / Nesta dissertação discutiremos modelos e métodos de soluções de programação estocástica para resolver problemas de ALM em fundos de pensão. Apresentaremos o modelo de (Drijver et al.), baseado na programação estocástica multiestágios inteira-mista. Um estudo de caso para um problema de ALM será apresentado usando simulação de cenários.

Made available in DSpace on 2014-12-17T15:34:50Z (GMT). No. of bitstreams: 1 PriscilaTN_DISSERT.pdf: 1453426 bytes, checksum: ee7ba96d3e5a4bcadc8c78e77ad84582 (MD5) Previous issue date: 2014-02-21 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / The objectives of this research were characterizing the dairy goat production systems and model it using linear program. On the first step of this research, the model was developed using data from farms that was affiliated in the ACCOMIG/Caprileite, used a similar dairy goat production systems and have a partnership program with Universidade Federal de Minas Gerais . The data of research were from a structured questionnaire applied with farmers and monitoring of production systems during a guided visit on their farms. The results permitted identify that all farms were classified as a small and have a intensive production system. The average herd size had 63.75 dairy goats on lactation; it permits a production of 153, 38 kg of goat milk per day. It was observed that existing more than one channel of commercialization for the goat milk and their derivative products. The data obtained, on the first step of this research, was used to develop a linear program model. It was evaluated in two goat production systems, called P1 and P2. The results showed that the P1 system, with an annual birth and lactation during approximately 300 days was the best alternative for business. These results were compared with a mixed (beef and dairy) goat system in the semiarid region, which indicated merged with both systems. Therefore, to achieve profits and sustainability of the system, in all simulations it was necessary a minimum limit of funding of U.S. $ 10,000.00; this value permit earning of U.S. $ 792.00 per month and pay the investment within 5 years / Este trabalho objetivou realizar a caracteriza??o e modelagem de sistemas de produ??o de caprinos leiteiros. A primeira parte deste trabalho foi realizada no estado de Minas Gerais, onde foram caracterizadas oito propriedades, selecionadas por serem afiliadas a Associa??o dos Criadores de Caprinos e Ovinos do Estado de Minas Gerais ACCOMIG/Caprileite; possu?rem sistemas de produ??o e manejo semelhantes e terem parceria e/ou buscarem orienta??es t?cnicas na Universidade Federal de Minas Gerais. Os dados foram coletados por meio de entrevistas estruturadas e o acompanhamento dos sistemas de produ??o durante visita as suas propriedades. Com os resultados obtidos foi poss?vel identificar que todas as propriedades foram classificadas como minif?ndios e sistema intensivo de produ??o. Essas condi??es resultam em desempenho individual dos animais do rebanho, que apresentaram m?dia de 63,75 cabras em lacta??o e produzindo 153,38 litros de leite/dia; com m?dia de 2,33 ? 0,63 l/cab./dia. Observou-se tamb?m, que nessa regi?o existem mais de um canal de comercializa??o do leite fluido e derivados l?cteos. Com base nas informa??es obtidas em sistemas de produ??o de cabras leiteiras foi parametrizado um modelo de programa??o linear, segunda etapa do projeto. Este modelo foi testado em dois sistemas de produ??o, do estado do Rio Grande do Norte. A partir das informa??es obtidas em question?rios estruturados e acompanhamento dos sistemas, denominados de fazendas P1 e P2, foi rodado o modelo, usando os limites de cr?dito do Programa Nacional de Fortalecimento da Agricultura Familiar (PRONAF). Os resultados mostraram que o sistema P1, com um parto anual e lacta??o aproximada de 300 dias, foi a melhor alternativa de neg?cio. Estes resultados foram confrontados com um sistema de produ??o misto (corte e leite) no semi?rido; cujo resultado indicou mesclar os dois sistemas. No entanto, para se conseguir lucro e a sustentabilidade do sistema, em todas as simula??es realizadas foi necess?rio duplicar o limite do custeio para R$ 10.000,00, que possibilita a renda de um sal?rio m?nimo/m?s, R$ 792,00 e pagar o investimento em at? cinco anos

Orientadores: Secundino Soares Filho, Aurelio Ribeiro Leite de Oliveira / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-12T10:09:35Z (GMT). No. of bitstreams: 1 Coelho_MaykVieira_M.pdf: 4740036 bytes, checksum: d26ebcd270a5b52dd64b6186dd2720d0 (MD5) Previous issue date: 2008 / Resumo: Na eventualidade de uma contingência, com a perda de unidades de geração em um sistema de potência, podem ser verificados desequilíbrios no conjunto carga-geração. Nestas situações torna,-se necessário o emprego de medidas corretivas que eliminem estas violações operativas, reconduzindo o sistema a um ponto de operação seguro. Visando obter este nível de segurança, o método de pontos interiores primal-dual é desenvolvido para o problema de minimização das perdas na geração e transmissão do fluxo de potência ótimo CC de um sistema de potência hidrotérmico considerando restrições de reserva de potência operacional. Em outras palavras, o serviço anciliar de reserva será provido por geradores conectados à rede elétrica e sincronizados com o sistema, com objetivo de disponibilizar uma quantidade extra de potência ativa, que pode ser imediatamente utilizada durante uma situação de contingência para restabelecer o equilíbrio no conjunto carga-geração. É feita também uma comparação com o modelo sem tais restrições de reserva. / Abstract: In the eventuality of a contingency, with the loss of units of generation in a power system, unbalances cqn be verified in the group load-generation. In such situations, corrective measuresthat eliminate these operative violations are necessary in order to lead the system to a safe operation point. Seeking to obtain this leveI of safety, the primal-dual interior point method is developed for tlie problem of minimization of the losses in the generation and transmission DC power flow of a hidrotermic power system considering operational reserve restrictions. In other words, the service reselv~ anciliar will be provided by connected generators to the electric network and synchronized with the'. system, with the goal of making available an extra amount of active power, that can be immediately used during a contingency situation to reestablish the balance in the group load-generation. A comparison with the model without such reserve restrictions is aIs o performed. / Mestrado / Energia Eletrica / Mestre em Engenharia Elétrica

Embedded system designers face multiple challenges in fulfilling the runtime requirements of programs. Effective scheduling of programs is required to extract as much parallelism as possible. These scheduling algorithms must also improve speedup after instruction-set extensions have occurred. Scheduling of dynamic code at run time is made more difficult when the static components of the program are scheduled inefficiently. This research aims to optimize a program’s static code at compile time. This is achieved with four algorithms designed to schedule code at the task and instruction level. Additionally, the algorithms improve scheduling using instruction set extended code on symmetrical homogeneous multiprocessor systems. Using these algorithms, we achieve speedups up to 3.86X over sequential execution for a 4-issue 2-processor system, and show better performance than recent heuristic techniques for small programs. Finally, the algorithms generate speedup values for a 64-point FFT that are similar to the test runs.

No description available.

The evolution and deployment of fifth-generation (5G) and beyond (B5G) infrastructure will require a tremendous effort to specify the design, standards, and manufacturing. 5G is vital to modern technological evolution, including industry 4.0, automotive, entertainment, and health care. The ambitious and challenging 5G project is classified into three categories, which provide an essential supporting platform for applications associated with: Enhanced mobile broadband (eMBB) Ultra-reliable low-latency communication (URLLC) Massive machine-type communication (mMTC) The demand for URLLC grows, particularly for applications like autonomous guided vehicles (AGVs), unmanned aerial vehicles (UAVs), and factory automation, and has a strict requirement of low latency of 1 ms and high reliability of 99.999%. To meet the needs of communication-sensitive and computation-intensive applications with different quality-of-service (QoS) requirements, this evolution focuses on ultra-dense edge networks with multi-access edge computing (MEC) facilities. MEC emerges as a solution, placing resourceful servers closer to users. However, the dynamic nature of processing and interaction patterns necessitates effective network control, which is challenging due to stringent requirements on both communication and computation. In this context, we introduce a novel approach to optimally manage task offloading, considering the intricacies of heterogeneous computing and communication services. Unlike existing methods, our methodology incorporates the number of admitted service migrations and QoS upper and lower bounds as binding constraints. The comprehensive model encompasses agent positions, MEC servers, QoS requirements, edge network communication, and server computing capabilities. Formulated as a mixed-integer linear program (MILP), it provides an optimal schedule for service migrations and bandwidth allocation, addressing the challenges posed by computation-intensive and communication-sensitive applications. Moreover, tailoring to an indoor robotics environment, we explore optimization-based approaches seeking an optimal system-level architecture while considering QoS guarantees. Optimization tools, e.g., ARCHEX, prove their ability to capture cyber-physical systems (CPS) requirements and generate correct-by-construction architectural solutions. We propose an extension in ARCHEX by incorporating dynamic properties, i.e., robot trajectories, time dimension, application-specific QoS constraints, and finally, integrating the optimization tool with a discrete-event network simulator (OMNeT++). This extension automates the generation of configuration files and facilitates result analysis, ensuring a comprehensive evaluation. This part of the work focuses on the dynamism of robots, server-to-service mapping, and the integration of automated simulation. The proposed extension is validated by optimizing and analyzing various indoor robotics scenarios, emphasizing critical performance parameters such as overall throughput and end-to-end delay (E2E). This integrated approach addresses the complex interplay of computation and communication resources, providing a solution for dynamic mobility, traffic, and application patterns in edge server environments.