Catálogo de publicaciones - libros

Sandwich type piezoelectric ceramic transducers are the most frequently applied source of ultrasounds in technical cleaning system. They have the ability to radiate in an ultrasonic medium, e.g. water, with maximum acoustic power when the vibration is activated by a current whose frequency equals the mechanical resonance frequency of the transducer. In resonant inverters the transducer units are part of the oscillating circuit, for which equivalent electrical circuit consist of connection in parallel: Co end RLC. The resonant frequency of the real circuit varies during the operation in function of many parameters, among others, the most important are temperature, time, the column of cleaning factor, and the surface of the cleaned elements. In this situation, to obtain the maximum value of the converter efficiency, its important role of control system to assure the optimal mechanical resonant frequencies of converter.

One of the most important challenges of collaborative learning systems is to offer mechanisms to facilitate the study of the relationships between the collaboration process and the characteristics of the solution (product) built by the learners in this work process. In this article, a machine learning algorithm that generates a set of rules to classify the different forms of collaboration within a group of learners with respect to the quality of the solution built is presented. The algorithm, based on a fuzzy model, is put into practice using data registered in a collaborative learning environment.

This paper describes the development of a clustering based fuzzy controller of an electromagnetic suspension vehicle using gain scheduling method and Kalman filter for a simplified single magnet system. Electromagnetic suspension vehicle systems are highly nonlinear and essentially unstable systems. For achieving the levitation control, we considered a fuzzy system modeling method based on clustering algorithm which a set of input/output data is collected from the well defined Linear Quadratic Gaussian (LQG) controller. Simulation results show that the proposed clustering based fuzzy controller robustly yields uniform performance over the mass variation range.

In this paper, we propose a new design methodology of fuzzy-neural networks – Fuzzy Relation–based Polynomial Neural Networks (FRPNN) with symbolic genetic algorithms and Information Granules (IG). We have developed a design methodology based on symbolic genetic algorithms to find the optimal structure for fuzzy-neural networks that expanded from Group Method of Data Handling (GMDH). Such parameters as the number of input variables, the order of the polynomial, the number of membership functions, and a collection of the specific subset of input variables are optimized for topology of FRPNN with the aid of symbolic genetic optimization that has search capability to find the optimal solution on the solution space. The augmented and genetically developed FRPNN (gFRPNN) results in a structurally optimized structure and comes with a higher level of flexibility in comparison to the one we encounter in the conventional FRPNNs. The GA-based design procedure being applied at each layer of FRPNN leads to the selection of the most suitable nodes (or FRPNs) available within the FRPNN. The performance of genetically optimized FRPNN (gFRPNN) is quantified through experimentation where we use a number of modeling benchmarks data which are already experimented with in fuzzy or neurofuzzy modeling.

This paper starts from the fuzzy environments of foreign currency options markets, introduces fuzzy sets theory, and gives a fuzzy version of Garman-Kohlhagen currency options pricing model. By taking exchange rate, domestic interest rate, foreign interest rate, and volatility as triangular fuzzy numbers, the currency option price will turn into a fuzzy number. This makes the financial investors who can pick any currency option price with an acceptable belief degree for the later use. In order to obtain the belief degree, an optimization procedure has been applied. An empirical study is performed based on market data. The study result indicates the fuzzy currency options pricing method is a useful tool for modeling the imprecise problem in the real world.

The use of rough set’s theoretic concepts has permitted in this work to make the mathematical model on mode decision in deinterlacing system. In this paper, a rough set approach based decision making problem is proposed. In the literature, some conventional deinterlacing methods provide high performance with higher computational burden. On the other hand, some other methods give low performance with lower computational burden. Those all methods have been reported that interpolate missing pixels indiscriminately in the same way. Our algorithm chooses the most suitable method adaptively based on rough set theory using four parameters. This deinterlacing approach employs a size reduction of the database system, keeping only the essential information for the process, especially in the representation of and reasoning with vague and/or imprecise knowledge. Decision making and interpolation results are presented. The results of computer simulations show that the proposed method outperforms a number of methods presented in the literature.

A clonal selection algorithm (Clonalg) inspires from Clonal Selection Principle used to explain the basic features of an adaptive immune response to an antigenic stimulus. It takes place in various scientific applications and it can be also used to determine the membership functions in a fuzzy system. The aim of the study is to adjust the shape of membership functions and a novice aspect of the study is to determine the membership functions. Proposed method has been implemented using a developed Clonalg program for a single input and output fuzzy system. In the previous work [1], using genetic algorithm (GA) is proposed to it. In this study they are compared, too and it has been shown that using clonal selection algorithm is advantageous than using GA for finding optimum values of fuzzy membership functions.

This paper introduces novel methodology for the clustering of data represented as leaf-labelled trees on the same leaf-set. We define an abstract term - the representative tree, which can be represented with a variety of trees, depending on applications. The quality of tree-clustering is based on Information Gain, which measures the increase of information contained by representative trees of the resulting clusters compared to a single representative tree of the whole dataset. Finally, we propose the k-best algorithm the objective function of which is to maximize the information gain. We show how it can be constructed for two different representative trees, well- known in phylogenetic analysis. Developed algorithms yield very promissing results.

We model populations of classifiers which are aimed to function in permanently varying environments, adapt to unexpected changes, to comply fitness function and survive. A failure to fulfill survivability condition is resulting in unsuccessful agents being removed from the agent society and be replaced by newborns which inherit some upbringing learning information from parent agents. We split the agent population into groups and suggest storing agent’s gains accumulated during most recent periods, distort randomly training signals and a level of survival threshold. A presence of optimal number of groups and a necessity of small groups with mutually collaborating agents is demonstrated.

Predictions computed by a classification tree are usually constant on axis-parallel hyperrectangles corresponding to the leaves and have strict jumps on their boundaries. Frequently a better approximation may be expected, if the prediction function of the original tree is replaced by a continuous approximation. The approximation is constructed using the same training data on which the original tree was grown and the structure of the tree is preserved.

The current paper uses the model of trees with soft splits suggested by Quinlan and implemented in C4.5, however, the training algorithm is substantially different. The method uses simulated annealing, so it is quite computationally expensive. However, numerical test with data derived from an experiment in particle physics shows that besides the expected better approximation of the training data, also smaller generalization error is achieved.