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

Papers by rodney saldanha

2005 IEEE Congress on Evolutionary Computation

This paper presents a new operator for genetic algorithms that enhances their convergence in the ... more This paper presents a new operator for genetic algorithms that enhances their convergence in the case of nonlinear problems with nonlinear equality constraints. The proposed operator, named CQA (Constraint Quadratic Approximation), can be interpreted as both a local search engine (that employs quadratic approximations of both objective and constraint functions for guessing a solution estimate) and a kind of elitism operator that plays the role of "fixing" the best estimate of the feasible set. The proposed operator has the advantage of not requiring any additional function evaluation per algorithm iteration, solely making use of the information that would be already obtained in the course of the usual Genetic Algorithm iterations. The test cases that were performed suggest that the new operator can enhance both the convergence speed (in terms of the number of function evaluations) and the accuracy of the final result.

Energy Conversion and Management, 2015

This paper analyzes the mathematical properties related to the hydro-power generation optimizatio... more This paper analyzes the mathematical properties related to the hydro-power generation optimization problem. From the problem's physical nature, it can be concluded that the hydro-power generation is a function of the plant productivity and the water discharge. Since the plant productivity must be an increasing function of the net height, it was shown, under mild conditions, that the resulting generation is a strongly increasing function. Given that, it was established some mathematical properties that guarantee a unique maximum inside the feasible set and, therefore, global optimality. Under this analysis it is possible to derive optimization problems with optimality guarantees. Finally, a problem concerning the minimization of the deficit between supply and demand during a time window for a single power plant is explored. A numerical example based on a real case is given. Other existing formulations could also take advantage of our main results to confirm global optimality or to prove multimodality. These results are also useful to define the best optimization strategies for a given optimization problem.

A fast power flow method for radial networks with linear storage and no matrix inversions

ABSTRACT This paper presents a modified direct approach for the forward/backward sweep power flow... more ABSTRACT This paper presents a modified direct approach for the forward/backward sweep power flow method. Taking advantage of the special topological characteristics of the radial network, an algorithm with linear storage complexity is defined. These features are summarized in the incidence matrix, which becomes a lower triangular matrix after the vertex ordering. This new formulation allows to solve linear systems of equations instead of explicitly inverting matrices during the iterative process, leading to a lower computational burden. Therefore, the proposed method is time and memory-efficient. The results show that the proposed method improves the storage and time complexity without any loss of accuracy, making it a robust and efficient method.

Neurocomputing, 2009

This paper applies the minimum gradient method (MGM) to denoise signals in engineering problems. ... more This paper applies the minimum gradient method (MGM) to denoise signals in engineering problems. The MGM is a novel technique based on the complexity control, which defines the learning as a bi-objective problem in such a way to find the best trade-off between the empirical risk and the machine complexity. A neural network trained with this method can be used to pre-process data aiming at increasing the signal-to-noise ratio (SNR). After training, the neural network behaves as an adaptive filter which minimizes the cross-validation error. By applying the general singular value decomposition (GSVD), we show the relation between the proposed approach and the Wiener filter. Some results are presented, including a toy example and two complex engineering problems, which prove the effectiveness of the proposed approach.

IEEE Transactions on Magnetics, 2004

This paper presents Optimise, a computational optimization environment tool for education in elec... more This paper presents Optimise, a computational optimization environment tool for education in electrical engineering. Optimise has been developed using software engineering process and object-oriented programming philosophy. This educational tool incorporates a set of deterministic and stochastic methods and also a set of computational intelligence techniques. Optimise offers a friendly interface that allows the students to practice the theory learned, and also to verify and compare the features of the optimization methods. A general view of the software is presented, describing its modules and class libraries. Some optimization problems are discussed to illustrate the flexibility and power of Optimise as an educational tool.

Adaptive deep-cut method in ellipsoidal optimization for electromagnetic design

IEEE Transactions on Magnetics, 1999

This paper presents an adaptive deep cut algorithm for the ellipsoid method. The formulation prop... more This paper presents an adaptive deep cut algorithm for the ellipsoid method. The formulation proposed employs a variable cut depth which depends on an estimate of the current ellipsoid center distance to the solution. These adaptations in the formulation provide an enhanced convergence in the algorithm without losing robustness. The results of an analytical problem and benchmark problem 22 indicate

IEEE Transactions on Magnetics, 2006

IEEE Transactions on Magnetics, 1997

In this paper, a hybrid technique for global based on the genetic algorithm and a determinktic me... more In this paper, a hybrid technique for global based on the genetic algorithm and a determinktic method is presented. A potential advantage of the hybrid method compared to the genetic algorithm is that global o p t h b t i ( ~1 can be performed more efficiently. An intrinsic pmb1em of the hybrid tmKques is rehted to the moment of stopping the stochastic routin; to launch the one. This is investigated using some natural criteria for the commutation between the two methods. The results show that it is possible to gain in efficiency and in accuracy but the criterion is usually problem dependent. Finally, to show the solution of a real problem, the hybrid algoritlann is coupled to a 2D cade based on the boundary element a conneetor of 145 kV GIs.

IEEE Transactions on Magnetics, 2001

This paper presents an exhaustive study of the Simple Genetic Algorithm (SGA), Steady State Genet... more This paper presents an exhaustive study of the Simple Genetic Algorithm (SGA), Steady State Genetic Algorithm (SSGA) and Replacement Genetic Algorithm (RGA). The performance of each method is analyzed in relation to several operators types of crossover, selection and mutation, as well as in relation to the probabilities of crossover and mutation with and without dynamic change of its values during the optimization process. In addition, the space reduction of the design variables and global elitism are analyzed. All GAs are effective when used with its best operations and values of parameters. For each GA, both sets of best operation types and parameters are found. The dynamic change of crossover and mutation probabilities, the space reduction and the global elitism during the evolution process show that great improvement can be achieved for all GA types. These GAs are applied to TEAM benchmark problem 22.

IEEE Transactions on Magnetics, 2005

This paper presents a three-dimensional computational model for nonlinear microwave heating probl... more This paper presents a three-dimensional computational model for nonlinear microwave heating problems. In order to reduce the computational costs, a new finite element method adaptive time-stepping scheme is proposed. Numerical results are compared with experimental measurements on a problem of dielectric heating of water.

IEEE Transactions on Magnetics, 2003

This work presents a multiobjective genetic algorithm with a novel feature, the real biased cross... more This work presents a multiobjective genetic algorithm with a novel feature, the real biased crossover operator. This operator takes into account the function values of the two parents, defining a nonuniform probability for the new individuals' locations that biases them toward the best parents' locations. The procedure leads to better estimates of the Pareto set. The proposed algorithm is applied to the optimization of a Yagi-Uda antenna in a wide frequency range with several simultaneous performance specifications, providing antenna geometries with good performance, compared to those presented in the available literature.

IEEE Transactions on Magnetics, 2005

Typical formulations for fuzzy optimization utilize piecewise linear membership functions, which ... more Typical formulations for fuzzy optimization utilize piecewise linear membership functions, which introduces regions of no differentiability and prevents the employment of deterministic methods. We propose the utilization of continuously differentiable membership functions, which permit the use of gradient-based methods. We present a general expression for the gradient of the decision degree function. Then, a stochastic method can be used to find a good starting point for the deterministic technique, in a hybrid approach. The formulation is applied to the optimization of an electrostatic micromotor. The average torque is maximized subject to a fuzzy constraint for the torque ripple. The results show the validity of the methodology in electromagnetic design.

IEEE Transactions on Magnetics, 2006

The TEAM benchmark problem 22 is an important optimization problem in electromagnetic design, whi... more The TEAM benchmark problem 22 is an important optimization problem in electromagnetic design, which can be formulated as a constrained mono-objective problem or a multiobjective one with two objectives. In this paper, we propose a multiobjective version with three objectives, whose third objective is related to the quench constraint and the better use of the superconducting material. The formulation proposed yields results that provide new alternatives to the designer. We solved the formulation proposed using the multiobjective clonal selection algorithm. After that, we selected a particular solution using a simple decision making procedure.

IEEE Transactions on Magnetics, 2001

Neurocomputing, 2015

This paper presents a prediction algorithm for features detection in Ground Penetrating Radar (GP... more This paper presents a prediction algorithm for features detection in Ground Penetrating Radar (GPR) based surveys. Based on signal processing and soft-computing techniques, the coupled use of principalcomponent analysis and neural networks enable a definition of an efficient method for analyzing GPR electromagnetic data. To guarantee a low error rate, a study of the algorithm main numerical parameters was performed by means of electromagnetic synthetic-data models. Results for detecting features of geological layers demonstrate not only the method predictions accuracy but also the simple interpretation of its output through scenarios reconstructed images.

Decisior implementation in neural model selection by multi-objective optimization

Vii Brazilian Symposium on Neural Networks, Proceedings, 2002

This work presents a new learning scheme for improving the generalization of multilayer perceptro... more This work presents a new learning scheme for improving the generalization of multilayer perceptrons (MLPs). The proposed multiobjective algorithm approach minimizes both the sum of squared error and the norm of network weight vectors to obtain the Pareto-optimal solutions. Since the Pareto-optimal solutions are not unique, we need a decision phase ("decisor") in order to choose the best one as

International Journal of Numerical Modelling: Electronic Networks, Devices and Fields, 2014

The solution of electromagnetic wave propagation problems in time domain using an explicit method... more The solution of electromagnetic wave propagation problems in time domain using an explicit method requires the inversion of Hodge matrices. This paper proposes an approximation to obtain a sparse inverse via the sparsity pattern of the original matrix. It is also shown the application of the algorithm Cuthill-McKee on Hodge matrices in order to reduce their bandwidth and thus speed up the method of recursive sparsification.

Improving helical antennas using a deep-cut ellipsoidal algorithm

2013 SBMO/IEEE MTT-S International Microwave & Optoelectronics Conference (IMOC), 2013

ABSTRACT In this paper, uniform helical antennas are used as starting point to generate optimized... more ABSTRACT In this paper, uniform helical antennas are used as starting point to generate optimized non-uniform ones. The improvements in the input impedance and gain are obtained varying the shape of the helix using a second order polynomial function. The optimal polynomial parameters are found using the deep-cut version of the ellipsoidal optimization algorithm. The use of the well known formulae and diagrams for the design of uniform helical antennas as a initial point of the deterministic method can reduce the search domain and provides a fast convergence rate. Results are presented for WiFi application using a center frequency of 2.45 GHz. It shows that the input impedance and the antenna length can be considerably improved using a simple and efficient parametrization.

IEEE Transactions on Magnetics, 1996

The optimal design of a ridged waveguide is considered in this paper. The geometric shape of the ... more The optimal design of a ridged waveguide is considered in this paper. The geometric shape of the waveguide is computed so that two prescribed lowest modes of the transverse magnetic fields are obtained. A general formulation of the problem is presented: the electromagnetic problem i s solved by a finite-difference scheme, and the Ellipsoid and the Tabu Search Algorithms are used to solve the optimization problem. This formulation should enable the design of guides to meet most practical specifications. Some computational results are given.

A new constrained ellipsoidal algorithm for nonlinear optimization with equality constraints

IEEE Transactions on Magnetics, 2003

This paper presents a new algorithm for nonlinear optimization, the cone ellipsoidal algorithm (C... more This paper presents a new algorithm for nonlinear optimization, the cone ellipsoidal algorithm (CEA), that is suitable for dealing with equality constraints and deterministically converges to the global solution in convex problems. The algorithm is based on the traditional ellipsoidal algorithm and on some new cone conditions. CEA simultaneously searches the objective function minimum and the problem feasible region. A

2005 IEEE Congress on Evolutionary Computation

This paper presents a new operator for genetic algorithms that enhances their convergence in the ... more This paper presents a new operator for genetic algorithms that enhances their convergence in the case of nonlinear problems with nonlinear equality constraints. The proposed operator, named CQA (Constraint Quadratic Approximation), can be interpreted as both a local search engine (that employs quadratic approximations of both objective and constraint functions for guessing a solution estimate) and a kind of elitism operator that plays the role of "fixing" the best estimate of the feasible set. The proposed operator has the advantage of not requiring any additional function evaluation per algorithm iteration, solely making use of the information that would be already obtained in the course of the usual Genetic Algorithm iterations. The test cases that were performed suggest that the new operator can enhance both the convergence speed (in terms of the number of function evaluations) and the accuracy of the final result.

Energy Conversion and Management, 2015

This paper analyzes the mathematical properties related to the hydro-power generation optimizatio... more This paper analyzes the mathematical properties related to the hydro-power generation optimization problem. From the problem's physical nature, it can be concluded that the hydro-power generation is a function of the plant productivity and the water discharge. Since the plant productivity must be an increasing function of the net height, it was shown, under mild conditions, that the resulting generation is a strongly increasing function. Given that, it was established some mathematical properties that guarantee a unique maximum inside the feasible set and, therefore, global optimality. Under this analysis it is possible to derive optimization problems with optimality guarantees. Finally, a problem concerning the minimization of the deficit between supply and demand during a time window for a single power plant is explored. A numerical example based on a real case is given. Other existing formulations could also take advantage of our main results to confirm global optimality or to prove multimodality. These results are also useful to define the best optimization strategies for a given optimization problem.

A fast power flow method for radial networks with linear storage and no matrix inversions

ABSTRACT This paper presents a modified direct approach for the forward/backward sweep power flow... more ABSTRACT This paper presents a modified direct approach for the forward/backward sweep power flow method. Taking advantage of the special topological characteristics of the radial network, an algorithm with linear storage complexity is defined. These features are summarized in the incidence matrix, which becomes a lower triangular matrix after the vertex ordering. This new formulation allows to solve linear systems of equations instead of explicitly inverting matrices during the iterative process, leading to a lower computational burden. Therefore, the proposed method is time and memory-efficient. The results show that the proposed method improves the storage and time complexity without any loss of accuracy, making it a robust and efficient method.

Neurocomputing, 2009

This paper applies the minimum gradient method (MGM) to denoise signals in engineering problems. ... more This paper applies the minimum gradient method (MGM) to denoise signals in engineering problems. The MGM is a novel technique based on the complexity control, which defines the learning as a bi-objective problem in such a way to find the best trade-off between the empirical risk and the machine complexity. A neural network trained with this method can be used to pre-process data aiming at increasing the signal-to-noise ratio (SNR). After training, the neural network behaves as an adaptive filter which minimizes the cross-validation error. By applying the general singular value decomposition (GSVD), we show the relation between the proposed approach and the Wiener filter. Some results are presented, including a toy example and two complex engineering problems, which prove the effectiveness of the proposed approach.

IEEE Transactions on Magnetics, 2004

This paper presents Optimise, a computational optimization environment tool for education in elec... more This paper presents Optimise, a computational optimization environment tool for education in electrical engineering. Optimise has been developed using software engineering process and object-oriented programming philosophy. This educational tool incorporates a set of deterministic and stochastic methods and also a set of computational intelligence techniques. Optimise offers a friendly interface that allows the students to practice the theory learned, and also to verify and compare the features of the optimization methods. A general view of the software is presented, describing its modules and class libraries. Some optimization problems are discussed to illustrate the flexibility and power of Optimise as an educational tool.

Adaptive deep-cut method in ellipsoidal optimization for electromagnetic design

IEEE Transactions on Magnetics, 1999

This paper presents an adaptive deep cut algorithm for the ellipsoid method. The formulation prop... more This paper presents an adaptive deep cut algorithm for the ellipsoid method. The formulation proposed employs a variable cut depth which depends on an estimate of the current ellipsoid center distance to the solution. These adaptations in the formulation provide an enhanced convergence in the algorithm without losing robustness. The results of an analytical problem and benchmark problem 22 indicate

IEEE Transactions on Magnetics, 2006

IEEE Transactions on Magnetics, 1997

In this paper, a hybrid technique for global based on the genetic algorithm and a determinktic me... more In this paper, a hybrid technique for global based on the genetic algorithm and a determinktic method is presented. A potential advantage of the hybrid method compared to the genetic algorithm is that global o p t h b t i ( ~1 can be performed more efficiently. An intrinsic pmb1em of the hybrid tmKques is rehted to the moment of stopping the stochastic routin; to launch the one. This is investigated using some natural criteria for the commutation between the two methods. The results show that it is possible to gain in efficiency and in accuracy but the criterion is usually problem dependent. Finally, to show the solution of a real problem, the hybrid algoritlann is coupled to a 2D cade based on the boundary element a conneetor of 145 kV GIs.

IEEE Transactions on Magnetics, 2001

This paper presents an exhaustive study of the Simple Genetic Algorithm (SGA), Steady State Genet... more This paper presents an exhaustive study of the Simple Genetic Algorithm (SGA), Steady State Genetic Algorithm (SSGA) and Replacement Genetic Algorithm (RGA). The performance of each method is analyzed in relation to several operators types of crossover, selection and mutation, as well as in relation to the probabilities of crossover and mutation with and without dynamic change of its values during the optimization process. In addition, the space reduction of the design variables and global elitism are analyzed. All GAs are effective when used with its best operations and values of parameters. For each GA, both sets of best operation types and parameters are found. The dynamic change of crossover and mutation probabilities, the space reduction and the global elitism during the evolution process show that great improvement can be achieved for all GA types. These GAs are applied to TEAM benchmark problem 22.

IEEE Transactions on Magnetics, 2005

This paper presents a three-dimensional computational model for nonlinear microwave heating probl... more This paper presents a three-dimensional computational model for nonlinear microwave heating problems. In order to reduce the computational costs, a new finite element method adaptive time-stepping scheme is proposed. Numerical results are compared with experimental measurements on a problem of dielectric heating of water.

IEEE Transactions on Magnetics, 2003

This work presents a multiobjective genetic algorithm with a novel feature, the real biased cross... more This work presents a multiobjective genetic algorithm with a novel feature, the real biased crossover operator. This operator takes into account the function values of the two parents, defining a nonuniform probability for the new individuals' locations that biases them toward the best parents' locations. The procedure leads to better estimates of the Pareto set. The proposed algorithm is applied to the optimization of a Yagi-Uda antenna in a wide frequency range with several simultaneous performance specifications, providing antenna geometries with good performance, compared to those presented in the available literature.

IEEE Transactions on Magnetics, 2005

Typical formulations for fuzzy optimization utilize piecewise linear membership functions, which ... more Typical formulations for fuzzy optimization utilize piecewise linear membership functions, which introduces regions of no differentiability and prevents the employment of deterministic methods. We propose the utilization of continuously differentiable membership functions, which permit the use of gradient-based methods. We present a general expression for the gradient of the decision degree function. Then, a stochastic method can be used to find a good starting point for the deterministic technique, in a hybrid approach. The formulation is applied to the optimization of an electrostatic micromotor. The average torque is maximized subject to a fuzzy constraint for the torque ripple. The results show the validity of the methodology in electromagnetic design.

IEEE Transactions on Magnetics, 2006

The TEAM benchmark problem 22 is an important optimization problem in electromagnetic design, whi... more The TEAM benchmark problem 22 is an important optimization problem in electromagnetic design, which can be formulated as a constrained mono-objective problem or a multiobjective one with two objectives. In this paper, we propose a multiobjective version with three objectives, whose third objective is related to the quench constraint and the better use of the superconducting material. The formulation proposed yields results that provide new alternatives to the designer. We solved the formulation proposed using the multiobjective clonal selection algorithm. After that, we selected a particular solution using a simple decision making procedure.

IEEE Transactions on Magnetics, 2001

Neurocomputing, 2015

This paper presents a prediction algorithm for features detection in Ground Penetrating Radar (GP... more This paper presents a prediction algorithm for features detection in Ground Penetrating Radar (GPR) based surveys. Based on signal processing and soft-computing techniques, the coupled use of principalcomponent analysis and neural networks enable a definition of an efficient method for analyzing GPR electromagnetic data. To guarantee a low error rate, a study of the algorithm main numerical parameters was performed by means of electromagnetic synthetic-data models. Results for detecting features of geological layers demonstrate not only the method predictions accuracy but also the simple interpretation of its output through scenarios reconstructed images.

Decisior implementation in neural model selection by multi-objective optimization

Vii Brazilian Symposium on Neural Networks, Proceedings, 2002

This work presents a new learning scheme for improving the generalization of multilayer perceptro... more This work presents a new learning scheme for improving the generalization of multilayer perceptrons (MLPs). The proposed multiobjective algorithm approach minimizes both the sum of squared error and the norm of network weight vectors to obtain the Pareto-optimal solutions. Since the Pareto-optimal solutions are not unique, we need a decision phase ("decisor") in order to choose the best one as

International Journal of Numerical Modelling: Electronic Networks, Devices and Fields, 2014

The solution of electromagnetic wave propagation problems in time domain using an explicit method... more The solution of electromagnetic wave propagation problems in time domain using an explicit method requires the inversion of Hodge matrices. This paper proposes an approximation to obtain a sparse inverse via the sparsity pattern of the original matrix. It is also shown the application of the algorithm Cuthill-McKee on Hodge matrices in order to reduce their bandwidth and thus speed up the method of recursive sparsification.

Improving helical antennas using a deep-cut ellipsoidal algorithm

2013 SBMO/IEEE MTT-S International Microwave & Optoelectronics Conference (IMOC), 2013

ABSTRACT In this paper, uniform helical antennas are used as starting point to generate optimized... more ABSTRACT In this paper, uniform helical antennas are used as starting point to generate optimized non-uniform ones. The improvements in the input impedance and gain are obtained varying the shape of the helix using a second order polynomial function. The optimal polynomial parameters are found using the deep-cut version of the ellipsoidal optimization algorithm. The use of the well known formulae and diagrams for the design of uniform helical antennas as a initial point of the deterministic method can reduce the search domain and provides a fast convergence rate. Results are presented for WiFi application using a center frequency of 2.45 GHz. It shows that the input impedance and the antenna length can be considerably improved using a simple and efficient parametrization.

IEEE Transactions on Magnetics, 1996

The optimal design of a ridged waveguide is considered in this paper. The geometric shape of the ... more The optimal design of a ridged waveguide is considered in this paper. The geometric shape of the waveguide is computed so that two prescribed lowest modes of the transverse magnetic fields are obtained. A general formulation of the problem is presented: the electromagnetic problem i s solved by a finite-difference scheme, and the Ellipsoid and the Tabu Search Algorithms are used to solve the optimization problem. This formulation should enable the design of guides to meet most practical specifications. Some computational results are given.

A new constrained ellipsoidal algorithm for nonlinear optimization with equality constraints

IEEE Transactions on Magnetics, 2003

This paper presents a new algorithm for nonlinear optimization, the cone ellipsoidal algorithm (C... more This paper presents a new algorithm for nonlinear optimization, the cone ellipsoidal algorithm (CEA), that is suitable for dealing with equality constraints and deterministically converges to the global solution in convex problems. The algorithm is based on the traditional ellipsoidal algorithm and on some new cone conditions. CEA simultaneously searches the objective function minimum and the problem feasible region. A