João A. Vasconcelos

Improvements in genetic algorithms

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.

A multiobjective methodology for evaluating genetic operators

This paper is concerned with the problem of evaluating genetic algorithm (GA) operator combinations. Each GA operator, like crossover or mutation, can be implemented according to several different formulations. This paper shows that: 1) the performances of different operators are not independent and 2) different merit figures for measuring a GA performance are conflicting. In order to account for this problem structure, a multiobjective analysis methodology is proposed. This methodology is employed for the evaluation of a new crossover operator (real-biased crossover) that is shown to bring a performance enhancement. A GA that was found by the proposed methodology is applied in an electromagnetic (EM) benchmark problem.

Treating constraints as objectives in multiobjective optimization problems using niched Pareto genetic algorithm

In this paper, the constraints, in multiobjective optimization problems, are treated as objectives. The constraints are transformed in two new objectives: one is based on a penalty function and the other is made equal to the number of violated constraints. To ensure the convergence to a feasible Pareto optimal front, the constrained individuals are eliminated during the elitist process. The treatment of infeasible individuals required some relevant modifications in the standard Parks and Miller elitist technique. Analytical and electromagnetic problems are presented and the results suggest the effectiveness of the proposed approach.

Genetic algorithm coupled with a deterministic method for optimization in electromagnetics

In this paper, a hybrid technique for global optimization based on the genetic algorithm and a deterministic method is presented. A potential advantage of the hybrid method compared to the genetic algorithm is that global optimization can be performed more efficiently. An intrinsic problem of the hybrid techniques is related to the moment of stopping the stochastic routine to launch the deterministic 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 algorithm is coupled to a 2D code based on the boundary element method to optimize a connector of 145 kV GIS.

Optimization of an offset reflector antenna using genetic algorithms

This paper presents the application of genetic algorithms in the optimization of an offset reflector antenna. The antenna shape is designed in order to obtain a uniform radiation pattern on the Brazilian territory. Modified genetic operators are proposed with the aim to increase the efficiency of the real coded genetic algorithms used here.

The

This paper presents a novel approach for dealing with the structural risk minimization (SRM) applied to a general setting of the machine learning problem. The formulation is based on the fundamental concept that supervised learning is a bi-objective optimization problem in which two conflicting objectives should be minimized. The objectives are related to the empirical training error and the machine complexity. In this paper, one general Q-norm method to compute the machine complexity is presented, and, as a particular practical case, the minimum gradient method (MGM) is derived relying on the definition of the fat-shattering dimension. A practical mechanism for parallel layer perceptron (PLP) network training, involving only quasi-convex functions, is generated using the aforementioned definitions. Experimental results on 15 different benchmarks are presented, which show the potential of the proposed ideas.

Q

The method Promethee II has produced attractive results in the choice of the most satisfactory optimal solution of convex multiobjective problems. However, according to the current literature, it may not work properly with nonconvex problems. A modified version of this method, called multiplicative Promethee, is proposed in this paper. Both versions are applied to some analytical problems, previously optimized by an evolutionary algorithm. The multiplicative Promethee got much better results than the original Promethee II, being capable of solving convex and nonconvex problems, with continuous and discontinuous Pareto fronts.

-Norm Complexity Measure and the Minimum Gradient Method: A Novel Approach to the Machine Learning Structural Risk Minimization Problem

Abstract In this paper, both the architecture and learning procedure underlying the parallel layer perceptron is presented. This topology, different to the previous ones, uses parallel layers of perceptrons to map nonlinear input–output relationships. Comparisons between the parallel layer perceptron, multi-layer perceptron and ANFIS are included and show the effectiveness of the proposed topology.

A multiplicative version of Promethee II applied to multiobjective optimization problems

A procedure to optimize the specific absorption rate deposed in the patient during oncology hyperthermia treatment is presented. It is based on a genetic algorithm coupled to a finite-element formulation. The optimization procedure is applied to a real human body obtained from computerized tomography scans.

Parallel layer perceptron

This paper presents a multiobjective shape optimization of an offset reflector antenna using the Cone of Efficient Directions Algorithm (CEDA), that includes a multiobjective line search. This algorithm features a monotone dominance convergence: the sequence of solution antennas occur in such a way that all the objectives are improved simultaneously. Given an area to be covered, the desired radiation pattern is the one with the maximum mean gain and uniformity, possibly weighted, inside it. These features are achieved by the definition of a single objective function, which must held in three different frequencies (a total of three conflicting objectives), in order to provide broad-band characteristics to the designed antenna