rodney saldanha - Academia.edu (original) (raw)

Papers by rodney saldanha

Learning and Nonlinear Models, 2007

In this paper stochastic methods for solving redundancy-reliability allocation problems are emplo... more In this paper stochastic methods for solving redundancy-reliability allocation problems are employed. In order to understand the major issues on solving those problems, three different designs are considered and redundancy allocation problems are formulated for each of them. The problems are solved in two stages, one stochastic and another deterministic. Genetic and the immune algorithms are implemented.

Machine Learning, 2010

Nowadays, neural networks (NNs) are widely applied in the solution of several real world problems... more Nowadays, neural networks (NNs) are widely applied in the solution of several real world problems. They have been successfully used in many fields such as chemistry, physics, engineering, and bio-informatics among others. However, their use often relies on some handcrafted settings, such as the number of layers and neurons. This chapter will discuss the Structural Risk Minimization (SRM) problem using some multiobjective optimization concepts. Both are closely related to the classical Tikhonov's regularization scheme, and, it is also exploited in this work. A neural network is a learning machine capable to describe, to the input x, the set of functions F = { f (x, w) : x ∈ X, w ∈ W}, where W is the space of possible weights. Given a supervisor which defines an output vector y ∈ Y (desired output), for a given input x, according to the conditional distribution F(y|x), the ultimate goal in the learning problem is to find w ∈ W that best approximates the supervisor answer given some measure. To some loss function L(.), the expected risk (error) can be defined as, Vapnik (1998): R(w) = L (y, f (x, w)) dF(x, y). (1) This approach was considered self-evident for many years and the main milestone was to find better algorithms to solve (2). However, the non self-evident overfitting phenomenon has appeared. This would imply that w * = w 0. One way to characterize it is by the bias and

Satellite Communications, 2010

Neurocomputing, 2000

This paper presents a new learning scheme for improving generalization of multilayer perceptrons.... more This paper presents a new learning scheme for improving generalization of multilayer perceptrons. The algorithm uses a multi-objective optimization approach to balance between the error of the training data and the norm of network weight vectors to avoid ...

IEEE Transactions on Magnetics, 1990

The problem to be solved is the determination of the optimum profile and location of shielding ut... more The problem to be solved is the determination of the optimum profile and location of shielding utilized by high-voltage equipment. The aim is to obtain a linear voltage distribution along the axis of symmetry. The calculation of electrostatic fields requires the solution of Laplace's or Poisson's equation with boundary conditions satisfied. This problem can be solved either by analytical or numerical methods such as the finite difference techniques, the finite-element method, the Monte Carlo method and the charge simulation method. The last is very simple when applied to the proposed system. It can be shown that the application of the charge simulation to the determination of the optimum geometry of the shielding electrodes leads to an unconstrained minimization problem. Some algorithms are discussed and computational results are presented.

IEEE Transactions on Magnetics, 2010

This paper proposes a novel ellipsoid method for the optimization of electromagnetic constrained ... more This paper proposes a novel ellipsoid method for the optimization of electromagnetic constrained problems. Unlike the classical method, which can apply only one cut per iteration, this novel algorithm can employ multiple cuts simultaneously. This improves the convergence rate while preserving all theoretical guarantees of the original method. The design of modelled reflector antennas for optimal coverage of the Brazilian, Chinese, and American territories is presented. These problems have 38 control variables and numerous constraints (18, 11, and 12). Several results are presented for the enhanced and classical ellipsoid methods, as well as for stochastic algorithms. They assert the efficiency of the introduced technique.

Energies, 2018

This paper proposes a revaluation of the Brazilian wind energy policy framework and the energy au... more This paper proposes a revaluation of the Brazilian wind energy policy framework and the energy auction requirements. The proposed model deals with the four major issues associated with the wind policy framework that are: long-term wind speed sampling, wind speed forecasting reliability, energy commercialization, and the wind farm profitability. Brazilian wind policy, cross-checked against other countries policies, showed to be too restrictive and outdated. This paper proposes its renewal, through the adoption of international standards by Brazilian policy-makers, reducing the wind time sampling necessary to implement wind farms. To support such a policy change, a new wind forecasting method is designed. The method is based on fuzzy time series shaped with a statistical significance approach. It can be used to forecast wind behavior, by drawing the most-likely wind energy generation intervals given a confidence degree. The proposed method is useful to evaluate a wind farm profitabili...

IEEE Transactions on Magnetics, 2016

A possible strategy for avoiding singular material parameters in a transformation-based invisibil... more A possible strategy for avoiding singular material parameters in a transformation-based invisibility cloak involves an out-of-plane stretching, calculated to compensate the in-plane singular transformation. In this paper, we used numerical simulations to analyze the relation between the out-of-plane transformation, the resulting material anisotropy and the total scattering cross width. Moreover, we show how this information can be used to optimize a cloak with homogeneous anisotropic layers.

This paper deals with the problem of detecting the occurrence of a car accident in an urban envir... more This paper deals with the problem of detecting the occurrence of a car accident in an urban environment. For this purpose, machine learning techniques are trained with the traffic flow measurements considering both the normal and the situation in which the accident caused a partial closure of the lanes. Several machine learning techniques results are presented to several car breaking scenarios.

IEEE Transactions on Magnetics, 1994

IEEE Transactions on Magnetics, 2000

IEEE Transactions on Magnetics, 2006

IEEE Transactions on Magnetics, 2015

Reducing the scattering and enhancing the bandwidth (BW) are two desirable aspects in cloak appli... more Reducing the scattering and enhancing the bandwidth (BW) are two desirable aspects in cloak applications. In this paper, we propose their optimization using a simplified objective function, which avoids a costly detailed frequency sweep to get the BW at each iteration. The multi-objective characteristic of our optimization permits us to obtain a set of possible solutions (the Pareto-Front), which represents different compromises between the two objectives. Two examples of dielectric cloaks are presented and discussed.

Digests of the 2010 14th Biennial IEEE Conference on Electromagnetic Field Computation, CEFC 2010, 2010

ABSTRACT This paper presents an optimization problem formulation to design broadband antennas wit... more ABSTRACT This paper presents an optimization problem formulation to design broadband antennas with optimal impedance matching considering that the source impedance is tunable inside a range. Usually an antenna is designed to match a given source impedance, which can be too constraining. The new optimization problem formulation is introduced to answer a less strict query: which impedance is optimal for the behavior inside a frequency range of a given antenna? The formulation is used to investigate Archimedean spiral antennas.

ABSTRACT This paper presents a real-world application of neural networks. This application consid... more ABSTRACT This paper presents a real-world application of neural networks. This application considers the estimation of the convection heat transfer coefficient of a run-out cooling table in a steel-making process. Firstly, data of several runs were collected considering the cooling table variables and the reached temperatures. Afterwards, using numerical models and optimization, the equivalent heat transfer coefficient is evaluated for each run. Finally, a neural network is applied to define the relationships between the process variables (thickness, water flow, among others) and the estimated heat transfer coefficient. The results are compared with some models derived from the process physics.

2006 12th Biennial IEEE Conference on Electromagnetic Field Computation, 2006

ABSTRACT Adaptive time stepping schemes can reduce the computational costs in 3D microwave heatin... more ABSTRACT Adaptive time stepping schemes can reduce the computational costs in 3D microwave heating models. In this work, the temperature-permittivity characteristic of different materials are used as criterion to reduce the number of iterations without compromise the accuracy