Catálogo de publicaciones - libros

Efficiency of two mutation operators applied in a clonal selection based optimization algorithm AIIA for non-stationary tasks is investigated. In both operators traditional Gaussian random number generator was exchanged by -stable random number generator and thus became one of the parameters of the algorithm. Obtained results showed that appropriate tuning of the parameter allows to outperform the results of algorithms with the traditional operators.

This paper deals with the flow-shop scheduling problem with no-store policy and minimal cycle time criterion. The model and some properties of the problem have been presented. To solve the problem, we propose new genetic algorithm equipped with auxiliary gene expression mechanism, which creates offspring using genetic information from both parents as well as asleep information from ancestors (grand- father, grand grandfather). The presented computational tests proved superiority of the proposed approach over traditional, basic GA scheme.

In this paper we propose a genetic and greedy algorithm combination for the optimization of the Topological Active Nets (TAN) model. This is a deformable model used for image segmentation that integrates features of region-based and edge-based segmentation techniques, being able to fit the edges of the objects and model their inner topology. The hybrid approach we propose can optimize the active nets through the minimization of the model energy functions and, moreover, it can provide some segmentation results unreachable by the GA method alone such as changes in the net topology.

Chromatic number, chromatic sum and chromatic sum number are important graph coloring characteristics. The paper proves that a parallel metaheuristic like the parallel genetic algorithm (PGA) can be efficiently used for computing approximate sum colorings and finding upper bounds for chromatic sums and chromatic sum numbers for hard–to–color graphs. Suboptimal sum coloring with PGA gives usually much closer upper bounds then theoretical formulas known from the literature.

A new Genetic Programming variant called Liquid State Genetic Programming (LSGP) is proposed in this paper. LSGP is a hybrid method combining a dynamic memory for storing the inputs (the liquid) and a Genetic Programming technique used for the problem solving part. Several numerical experiments with LSGP are performed by using several benchmarking problems. Numerical experiments show that LSGP performs similarly and sometimes even better than standard Genetic Programming for the considered test problems.

This paper presents optimal planning of tie-switch operation in an electric power distribution system under an emergency feed condition, i.e. operation during a post-fault condition. A heuristic fault isolation algorithm and a genetic-based service restoration algorithm are proposed and compared. With the proposed restoration algorithm, high reliable service of electric distribution systems is expected. To ensure a small number of customer interruption, average energy not supplied (AENS) is used as the objective function to be minimized. 25-node and 118-node distribution test feeders were employed for test. Satisfactory results show that the genetic approach is appropriate to a kind of tie-switch operation planning in order to minimize effects of a permanent fault on customer service interruption.

This paper deals with the multi-objective definition of the feature selection problem for different pattern recognition domains. We use NSGA II the latest multi-objective algorithm developed for resolving problems of multi-objective aspects with more accuracy and a high convergence speed. We define the feature selection as a problem including two competing objectives and we try to find a set of optimal solutions so called Pareto-optimal solutions instead of a single optimal solution. The two competing objectives are the minimization of both the number of used features and the classification error using 1-NN classifier. We apply our method to five databases selected from the UCI repository and we report the results on these databases. We present the convergence of the NSGA II on different problems and discuss the behavior of NSGA II on these different contexts.

The paper presents a method for designing 2-D IIR filters with a quadrantally symmetric magnitude response. The method is based on two error criteria, i.e., equiripple error criterion in the passband and least-squared error criterion in the stopband. Two objective functions are introduced and the filter design problem is transformed into an equivalent bicriterion optimization problem. The stability of the filter is ensured by explicitly including stability constraints in the considered optimization problem. A two-step solution procedure of the considered problem is proposed. In the first step, a genetic algorithm is applied. The final point from the genetic algorithm is used as the starting point for a local optimization method. Two design examples are given to illustrate the proposed technique. A comparison with a 2-D IIR filter designed using LS approach is also presented.

In this study, a hybrid genetic algorithm, which merges a genetic algorithm with simulated annealing, is derived for nonlinear channel blind equalization using RBF networks. The proposed hybrid genetic algorithm is used to estimate the output states of a nonlinear channel, based on the Bayesian likelihood fitness function, instead of the channel parameters. From these estimated output states, the desired channel states of the nonlinear channel are derived and placed at the center of a RBF equalizer to reconstruct transmitted symbols. In the simulations, binary signals are generated at random with Gaussian noise. The performance of the proposed method is compared with those of a conventional genetic algorithm (GA) and a simplex GA. It is shown that the relatively high accuracy and fast convergence speed have been achieved.

This paper presents Genetic Algorithms and Simulated Annealing (GASA) based on feature extraction of speech signal and comparison with traditional Linear Predictive Coding (LPC) methods. The performance of each method is analyzed for ten speakers with independent text speaker verification database from Center for Spoken Language Understanding (CSLU) which was developed by Oregon Graduate Institute (OGI). The GASA algorithm is also analyzed with constant population size for different generation numbers, crossover and mutation probabilities. When compared with the Mean Squared Error (MSE) of the each speech signal for each method, all simulation results of the GASA algorithm are more effective than LPC methods.