Search

The paper shows the application of the multidimensional scaling to discover the intrinsic dimensionality of the substitution matrices. These matrices are used in Bioinformatics to compare amino acids in the alignment procedures. However, the methodology can be used in other applications to discover the intrinsic dimensionality of a wide class of symmetrical matrices. The discovery of the intrinsic dimensionality of substitutions matrices is a data processing problem with applications in chemical evolution. The problem is related with the number of relevant physical, chemical and structural characteristic involved in these matrices. Many studies have dealt with the identification of relevant characteristic sets for these matrices, but few have concerned with establishing an upper bound of their cardinality. The methodology of multidimensional scaling is used to map the substitution matrix information in a virtual low dimensional space. The relationship between the quality of this process and the dimensionality of the mapping provides clues about the number of characteristics which better represents the matrix. To avoid the local minima problem, a genetic algorithm is used to minimize the objective function of the multidimensional scaling procedure. The main conclusion is that the number of effective characteristics involved in substitution matrices is small. [ABSTRACT FROM AUTHOR]

Evolutionary computation is a generic term used to make reference to the solution of computational problems planned and implemented based on models of an evolutionary process. Most of the evolutionary algorithms propose biological paradigms, and the concepts of natural selection, mutation and reproduction. However, other paradigms that can be adopted in the creation of evolutionary algorithms exist. Many problems involve not structured environments that can be considered from the perspective of cultural paradigms; the cultural paradigms offer a wide range of categorized models that ignore the possible solutions to the problem, -a common situation in the real life-. The purpose of the present work is to apply the computational properties of the cultural technology; on this case, to corroborate them by means of data mining to propose the solution to a specific problem. The above mentioned, carrying out an adaptation from the perspective of the societies modeling. An environment to carry out tests of this type was developed to allow the learning on the not very conventional characteristics of a cultural technology. This environment is called Baharastar. [ABSTRACT FROM AUTHOR]

The weblog medium while fundamentally an innovation in personal publishing has also come to engender a new form of social interaction on the web: a massively distributed but completely connected conversation covering every imaginable topic of interest. A by product of this ongoing communication is the set of hyperlinks made between weblogs in the exchange of dialog, a form of social acknowledgement on the part of authors. This paper seeks to understand the social implications of linking in the community, drawing from the hyperlink citations collected by the Blogdex project over the past three years. Social network analysis is employed to describe the resulting social structure, and two measures of authority are explored: popularity, as measured by webloggers'public affiliations and influence measured by citation of each others writing. These metrics are evaluated with respect to each other and with the authority conferred by references in the popular press. [ABSTRACT FROM AUTHOR]

Programming learning is an important subject for the students of computer science. These students must acquire knowledge and abilities which will deal with their future programming work for solving problems. In this sense, the discipline of programming constitutes a framework where Problem Based Learning (PBL) is the base used for acquiring the knowledge and abilities needed. Computer programming is a good research field where students should be assisted by an Intelligent Tutoring System (ITS) that guides them in their learning process. Furthermore, the complexity of these eLearning environments makes indispensable the necessity of the reuse and interoperability principles among eLearning tools. In this paper we will present an architectural approach that enables PBL for programming learning, merging several techniques: from Artificial Intelligence (AI) disciplines such as Bayesian Networks (BN) and Fuzzy Logic (FL); and from eLearning standards such as IMS Learning Design (IMS-LD). [ABSTRACT FROM AUTHOR]

With the changing demands of the markets and the developments of grid technology, grid computing is beginning to spill out over the boundaries of individual organizations. Partner grids create a collaboration environment to share computing resources, data and applications. OPGSA is a grid-based implementation of ebusiness solutions built upon commercial enterprise grid systems, and it makes use of Globus Toolkits 4 and DRMAA open standards. With this architecture, it is possible for the firms participating in a corporate supply chain to incorporate greater management capabilities and gain greater control over complex business processes. [ABSTRACT FROM AUTHOR]

In this paper a statistical selection of relevant features is presented. An experiment was designed to select relevant and not redundant features or characterization functions, which allow quantitatively discriminating among different types of complex networks. As well there exist researchers given to the task of classifying some networks of the real world through characterization functions inside a type of complex network, they do not give enough evidences of detailed analysis of the functions that allow to determine if all are necessary to carry out an efficient discrimination or which are better functions for discriminating. Our results show that with a reduced number of characterization functions such as the shortest path length, standard deviation of the degree, and local efficiency of the network can discriminate efficiently among the types of complex networks treated here. [ABSTRACT FROM AUTHOR]

The main objective of this work is to apply more effective methods than the traditional supervised techniques in the implementation of personalized recommender systems, which improve the accuracy of the predictions in classification tasks. Different model-based classification algorithms based on association rules and others that combine the induction of decision trees with this type of rule were studied. Data from the MovieLens recommender system was used in the analysis and comparison of the different algorithms. [ABSTRACT FROM AUTHOR]

The exponential growth of resources available in the Web has made it important to develop instruments to perform search efficiently. This paper proposes an approach for chemical information discovery by using focused crawling. The comparison of combination using various feature representations and classifier algorithms to implement focused crawlers was carried out. Latent Semantic Indexing (LSI) and Mutual Information (MI) were used to extract features from documents, while Naive Bayes (NB) and Support Vector Machines (SVM) were the selected algorithms to compute content relevance score. It was found that the combination of LSI and SVM provided the best solution. [ABSTRACT FROM AUTHOR]

This paper considers a continuous-time optimal portfolio problem, where the price processes of assets depend on the state of the stochastic market, which is assumed to follow a diffusion process. And trading in the risky asset is stopped if the market state hits a predefined threshold. The problem is formulated as a utility maximization with random horizon. Using the techniques of dynamic programming and Feynman-Kac representation theorem, we obtain a stochastic representation of optimal portfolio. Furthermore, in some special case, the closed-form of optimal portfolio is derived. Finally, we present computational results that show the differentiation between this proposed model and classical Merton model. [ABSTRACT FROM AUTHOR]

This paper investigates the use of the method of dimensional reduction Cascaded Nonlinear Components Analysis (C-NLPCA) in the protein secondary structure prediction problem. The use of the C-NLPCA is justified by the fact that this method manage to obtain a dimensional reduction that considers the nonlinearity of the data. In order to prove the effectiveness of the C-NLPCA, this paper presents comparisons of methods of components extraction, as well as, of existing predictors. The C-NLPCA revealed to be efficient, propelling a new field of research. [ABSTRACT FROM AUTHOR]

For the diagnosis of glaucoma, we propose a system of Artificial Intelligence that employs Artificial Neural Networks (ANN) and integrates the analysis of the nerve fibres of the retina from the study with scanning laser polarimetry (NFAII;GDx), perimetry and clinical data. The present work shows an analysis of 106 eyes of 53 patients, in accordance with the stage of glaucomatous illness in which each eye was found. The groups defined include stage 0, which corresponds to normal eyes; stage 1, for ocular hypertension; 2, for early glaucoma; 3, for established glaucoma; 4, for advanced glaucoma and 5, for terminal glaucoma. The developed ANN is a multilayer perceptron provided with the Levenberg-Marquardt method. The learning was carried out with half of the data and with the training function of gradient descent w/momentum backpropagation and was checked by the diagnosis of a glaucoma expert ophthalmologist. The other half of the data served to evaluate the model of the neuronal network. A 100% correct classification of each eye in the corresponding stage of glaucoma has been achieved. Specificity and sensitivity are 100%. This method provides an efficient and accurate tool for the diagnosis of glaucoma in the stages of glaucomatous illness by means of AI techniques. [ABSTRACT FROM AUTHOR]

Virtual screening (VS) experiments were applied to rank more than 700000 candidate lead-like virtual molecules in order of likelihood of binding to the colchicine site of tubulin, which is an important antitumor target. The best ranked compounds were clustered and classified by means of "ad hoc" semiautomatic chemoinformatic tools. The results obtained in this way were compared with those achieved by visual inspection protocols and the best were selected for synthesis and screening stages. [ABSTRACT FROM AUTHOR]

This paper describes the characteristics and the improvements of the free web-based spectral database NATPRO-C13, containing 13C NMR spectra data from more than 5.000 natural compounds and related derivates. It provides tools that facilitate the structural identification of natural compounds even before their purification. This database allows for searches by chemical structure, substructure, name, family compounds, and by spectral features i.e. chemical shifts and multiplicities. These capabilities are used together with visual interactive tools, which enable the structural elucidation of known and unknown compounds by comparison of their 13C NMR data. [ABSTRACT FROM AUTHOR]

DNA microarrays provide rich profiles that are used in cancer prediction considering the gene expression levels across a collection of samples.Support Vector Machines (SVM), have been applied to the classification of cancer samples with encouraging results. However, they are usually based on Euclidean distances that fail to reflect accurately the sample proximities. Besides, SVM classifiers based on non-Euclidean dissimilarities fail to reduce significantly the errors. In this paper, we propose an ensemble of SVM classifiers in order to reduce the errors. The diversity among classifiers is induced considering a set of complementary dissimilarities and kernels. The experimental results suggest that that our algorithm improves classifiers based on a single dissimilarity and a combination strategy such as Bagging. [ABSTRACT FROM AUTHOR]

This paper presents an autonomous intelligent agent developed for health care in geriatric residences. The paper focuses on the construction of an autonomous agent which incorporates a model of human thinking, such as reasoning based on past experiences. The work here presented focuses in the development of the CBP internal structure. The planning mechanism has been implemented by means of a novel FYDPS neural network. The system has been tested and this paper presents the results obtained. [ABSTRACT FROM AUTHOR]

An approach for noise filtering based on anisotropic nonlinear diffusion is presented. The method combines edge-preserving noise reduction with a strategy to enhance local structures and a mechanism to further smooth the background. The performance is illustrated with its application to electron cryotomography, a leading imaging technique for visualizing the molecular architecture of complex biological specimens. A challenging task in this discipline is to increase the extremely low signal-to-noise ratio to allow visualization and interpretation of the three-dimensional structures. The filtering method presented here succeeds in substantially reducing the noise with excellent preservation of the structures. [ABSTRACT FROM AUTHOR]

Many Expressed Sequence Tags (EST) sequencing projects produce thousands of sequences that must be cleaned and annotated. This research presents the so-called Full-Lengther, an algorithm that can find out full-length cDNA sequences from EST data. To accomplish this task, Full-Lenther is based on a BLAST report using a protein database such as UniProt. Blast alignments will guide to locate protein coding regions, mainly the start codon. Full-Lengther contains an ORF prediction algorithm for those cases which is not homologous to any sequence. The algorithm is implemented as a web tool to simplify its use and portability. This can be worldwide accessible via http://castanea.ac.uma.es/genuma/full-lengther/ [ABSTRACT FROM AUTHOR]

Bioinformatics tools are required to produce reliable, high quality data devoid of unwanted sequences in the preprocessing stage of current sequencing and EST projects. In this paper we describe SeqTrim, an algorithm designed to extract the insert sequence from any sequence read devoid of any foreign, contaminant or unwanted sequence, whatever the experimental process was. SeqTrim is easy to install and able to identify the sequence insert by removing low quality sequences, cloning vector, poly A or T tails, adaptors, and sequences that can be considered contaminants. It is easy to use and can be used as stand-alone application or as web page. The default parameters of the algorithm are best suited for most cases but a configuration file can be provided along with input sequences. SeqTrim admits several input and output formats (with and without quality values), which enables its inclusion in already or newly defined sequence processing workflows. SeqTrim is under continuous refinement due to collaboration between biologists and computer scientists which has succeed in correct dealing with most sequence cases and opens the possibility to include new capabilities to manage new kinds of bad sequences. [ABSTRACT FROM AUTHOR]

Electron cryotomography (cryoET) has the potential to elucidate the structure of complex biological specimens at molecular resolution but technical and computational improvements are still needed. This work addresses the determination and correction of the contrast transfer function (CTF) of the electron microscope in cryoET. Our approach to CTF detection and defocus determination depends on strip-based periodogram averaging, extended throughout the tilt series to overcome the low contrast conditions in cryoET. A method for CTF correction that deals with the defocus gradient in images of tilted specimens is also proposed. These approaches to CTF determination and correction have been applied here to several examples of cryoET of pleomorphic specimens and of single particles. CTF correction is essential for improving the resolution, particularly in those studies that combine cryoET with single particle averaging techniques. [ABSTRACT FROM AUTHOR]

Learning is a never ending activity for humans; it takes place everywhere and even when we do not realize. However, current learning environments make students deal with lectures, mostly associated with low control of the situation and implicit motivation. In contrast, previous researches have shown that sports, games or hobbies are activities that make people reach optimal experiences where self-motivation, control of the situation, high level of concentration and enjoyment are present. Some current efforts to design next generation of learning environments make use of ubiquitous systems to encourage students to perform learning activities everywhere and at anytime. However, those approaches lack the affective factor related to optimal experiences. To address this problem we present eCoology, an edutainment application that creates a ubiquitous learning environment with emotional features, and discuss some experimental results. [ABSTRACT FROM AUTHOR]

Wireless Sensor Networks represent a novel technology which is expected to experience a dramatic diffusion thanks to the promise to be a pervasive sensory means; however, one of the issues limiting their potential growth relies in the difficulty of managing and interpreting huge amounts of collected data. This paper proposes a cognitive architecture for the extraction of high-level knowledge from raw data through the representation of processed data in opportune conceptual spaces. The presented framework interposes a conceptual layer between the subsymbolic one, devoted to sensory data processing, and the symbolic one, aimed at describing the environment by means of a high level language. The features of the proposed approach are illustrated through the description of a sample application for wildfire detection. [ABSTRACT FROM AUTHOR]

A multiagent system that incorporates an Artificial Neural Networks based Intrusion Detection System (IDS) has been defined to guaranty an efficient computer network security architecture. The proposed system facilitates the intrusion detection in dynamic networks. This paper presents the structure of the Mobile Visualization Connectionist Agent-Based IDS, more flexible and adaptable. The proposed improvement of the system in this paper includes deliberative agents that use the artificial neural network to identify intrusions in computer networks. The agent based system has been probed through anomalous situations related to the Simple Network Management Protocol. [ABSTRACT FROM AUTHOR]

Case-Based Reasoning (CBR) systems solve new problems using others which have been previously resolved in a case memory, where each case represents a solved situation. Therefore, the case memory size and its organization influences on the computational time needed to solve new situations. For this reason, we organize the memory using a Self-Organization Map for defining patterns to allow system to do a selective retrieval using only the cases of the most suitable pattern. This works presents a case-based maintenance to incrementally introduce knowledge in SOM without retraining it because this process is very expensive in terms of computational time. The strategy is semi-supervised because we use the feedback provided by the expert and, at the same time, the self-organization of cases when clusters are readjusted. Results show a successful case-based maintenance. [ABSTRACT FROM AUTHOR]

Despite the wide variety of algorithms that exist to build predictive models, it can still be difficult to make accurate predictions for unknown values for certain types of data. New and innovative techniques are needed to overcome the problems underlying these difficulties for poor quality data, or data with a lack of available training cases. In this paper the authors propose a technique for integrating data from related datasets with the aim of improving the accuracy of predictions using Artificial Neural Networks. An overall improvement in the prediction power of models was shown when using the integration algorithm, when compared to models constructed using non-integrated data. [ABSTRACT FROM AUTHOR]

On discussing the necessary features for a group of agents to argue about their positions and intentions over a specific issue some questions arise: Why do agents interact? Is saving the knowledge generated in the interaction useful? How do agents manage arguments? How do agents dialogue? CBR is an adequate way to tackle such argumentation issues in MAS. This paper clarifies the advances achieved by applying CBR to argumentation in MAS, identifies open issues and proposes new ideas to face future challenges. [ABSTRACT FROM AUTHOR]

This paper presents a hybrid deliberative architecture based on the concept of CBPBDI agent. A CBP-BDI agent is a BDI agent that incorporates a CBP reasoning engine. The work here presented focuses in the development of the CBP internal structure. The planning mechanism has been implemented by means of a novel FYDPS neural network. The system has been tested and this paper presents the results obtained. [ABSTRACT FROM AUTHOR]

One objective for classifying textures in natural images is to achieve the best performance possible. Unsupervised techniques are suitable when no prior knowledge about the image content is available. The main drawback of unsupervised approaches is its worst performance as compared against supervised ones. We propose a new unsupervised hybrid approach based on two well-tested classifiers: Vector Quantization (VQ) and Fuzzy k-Means (FkM). The VQ unsupervised methods establishes an initial partition which is validated and improved through the supervised FkM. A comparative analysis is carried out against classical classifiers, verifying its performance. [ABSTRACT FROM AUTHOR]

When performing quality inspection of recycled paper one phenomenon of concern is the appearance of macroscopic undulations on the paper sheet surface that may emerge shortly or some time after its production. In this paper we explore the detection and measurement of this defect by means of computer vision and statistical pattern recognition techniques that may allow early detection at the production site. We propose features computed from Gabor Filter Banks (GFB) and Discrete Wavelet Transforms (DWT) for the characterization of paper sheet surface bumpiness in recycled paper images. The lack of a precise definition of the defect and the great variability of the sheet deformation shapes and scales, both within each image and between images, introduce additional difficulties to the problem. We obtain, with both proposed modeling approaches (GFB and DWT), classification accuracies are comparable to the agreement between human observers. The best performance is obtained using DWT features. [ABSTRACT FROM AUTHOR]

Image retrieval is generally implemented by image matching or based-regions retrieval, but it's difficult to balance retrieval performance and complexity. Query images may appear with different scales and rotations in different images, so a hierarchical image segmentation is proposed to partition the retrieved images into equal blocks with different sizes at different levels. Then, the similar metrics of these sub-blocks to query image, are evaluated to retrieve those sub-blocks with contents in query images. Meanwhile, information about scales and locations of query objects in retrieved images can also be returned. The hierarchical block image retrieval schemes with geometric invariants, normalized histograms and their combinations are tested by experiments via a database with 500 images, respectively. The retrieval accuracy with geometric invariants as invariant features can achieve 78% for the optimal similar metric threshold. Furthermore, the scheme can also work with different size images. [ABSTRACT FROM AUTHOR]

The paper proposes a Bayes-based relevance feedback approach integrating visual features and semantics for content-based image retrieval systems. The data of the image database are divided into small clusters by semantic supervised clustering algorithm. The cluster here is called as index cluster. So the data of each index cluster are similar both in visual features and in semanitcs. During relevance feedback process, users sign the positive and negative examples regarded a cluster as unit rather than a single image, and each feedback cluster construct a Bayessian classifier on visual features; and the semantic classes of the feedback examples construct the Bayessian classifier on semantics. At last, we use Bayesian classifiers on visual features and semantics respectively to adjust retrieval similarity distance. Our experiments on an image database show that a few cycles of relevance feedback by the proposed approach can significantly improve the retrieval precision. [ABSTRACT FROM AUTHOR]

Social foraging behavior of Escherichia coli bacteria has recently been explored to develop a novel algorithm for distributed optimization and control. The Bacterial Foraging Optimization Algorithm (BFOA), as it is called now, is currently gaining popularity in the community of researchers, for its effectiveness in solving certain difficult real-world optimization problems. Until now, very little research work has been undertaken to improve the convergence speed and accuracy of the basic BFOA over multi-modal fitness landscapes. This article comes up with a hybrid approach involving Particle Swarm Optimization (PSO) and BFOA algorithm for optimizing multi-modal and high dimensional functions. The proposed hybrid algorithm has been extensively compared with the original BFOA algorithm, the classical g_best PSO algorithm and a state of the art version of the PSO. The new method is shown to be statistically significantly better on a five-function test-bed and one difficult engineering optimization problem of spread spectrum radar poly-phase code design. [ABSTRACT FROM AUTHOR]

This paper presents an application of a genetic algorithm (GA) to the scheduling of hot rolling mills. The objective function used is based on earlier developments on flow stress modeling of steels. A hybrid two-phase procedure was applied in order to calculate the optimal pass reductions, in terms of minimum total rolling time. In the first phase, a non-linear optimization function was applied to evaluate the computational cost to the problem solution. For the second phase, a GA was applied. A comparison with two-point and simulated binary (SBX) crossover operators was established. The results were validated with data of industrial schedules. A GA with SBX crossover operator is shown to be an efficient method to calculate the multi-pass schedules at reduced processing time. [ABSTRACT FROM AUTHOR]

In this paper the problem of determining the atomic cluster configurations that minimize the Lennard-Jones potential energy is approached. Traditional studies are oriented to improve the quality of the solution and practically do not present statistical information to support the efficiency of the reported solution methods. Without this type of evidence the effectiveness of these methods might highly be dependent only on the capacity of the available computing resources. In this work it is proposed to incorporate statistical information on the performance of the solution methods. An advantage of this approach is that when the performance tests are standardized and statistically supported, we can take advantage of efficient solution methods that have been tested only in conditions of modest computing resources. An experimental study of the problem is presented in which the generated statistical information is used to identify two potential areas to improve the performance of the evaluated method. [ABSTRACT FROM AUTHOR]

In this work an initial approach to design Artificial Neural Networks to forecast time series is tackle, and the automatic process to design is carried out by a Genetic Algorithm. A key issue for these kinds of approaches is what information is included in the chromosome that represents an Artificial Neural Network. There are two principal ideas about this question: first, the chromosome contains information about parameters of the topology, architecture, learning parameters, etc. of the Artificial Neural Network, i.e. Direct Encoding Scheme; second, the chromosome contains the necessary information so that a constructive method gives rise to an Artificial Neural Network topology (or architecture), i.e. Indirect Encoding Scheme. The results for a Direct Encoding Scheme (in order to compare with Indirect Encoding Schemes developed in future works) to design Artificial Neural Networks for NN3 Forecasting Time Series Competition are shown. [ABSTRACT FROM AUTHOR]

This paper presents a new hybrid intelligent system that solves the Bin Packing Problem. The methodology involves the fusion of Soft Computing by means a genetic algorithm and Hard Computing using limits criterion and deterministic strategies. The innovative proposal inverts minimum computational resources expressed in generations with a high level quality solution and shows the algorithm performance with statistical methods. The average theoretical ratio for 1370 standard instances was 1.002 and the best known solution was achieved in 83.72% of the cases. As future work, an exhaustive analysis of characteristics of the hardest instances is proposed; the purpose is to find new hybrid methods. [ABSTRACT FROM AUTHOR]

This paper presents a new variant of Particle Swarm Optimization algorithm named QPSO for solving global optimization problems. QPSO is an integrated algorithm making use of a newly defined, multiparent, quadratic crossover operator in the Basic Particle Swarm Optimization (BPSO) algorithm. The comparisons of numerical results show that QPSO outperforms BPSO algorithm in all the twelve cases taken in this study. [ABSTRACT FROM AUTHOR]

A new hybrid approach to optimization in dynamical environments called Collaborative Evolutionary-Swarm Optimization (CESO) is presented. CESO tracks moving optima in a dynamical environment by combining the search abilities of an evolutionary algorithm for multimodal optimization and a particle swarm optimization algorithm. A collaborative mechanism between the two methods is proposed by which the diversity provided by the multimodal technique is transmitted to the particle swarm in order to prevent its premature convergence. The effect of changing the random number generator used for selection and for variation operators within CESO with a chaotic sequence generator is tested. [ABSTRACT FROM AUTHOR]

Parallel genetic algorithms (PGAs) are a powerful tool to deal with complex optimization problems. Nevertheless, the task of selecting its parameters accurately is an optimization problem by itself. Any additional help or hints to adjust the configuration parameters will lead both towards a more efficient PGA application and to a better comprehension on how these parameters affect optimization behavior and performance. This contribution offers an analysis on certain PGA parameters such as migration frequency, topology, connectivity and number of islands. The study has been carried out on an intensive set of experiments that collect PGA performance on several representative problems. The results have been analyzed using machine learning methods to identify behavioral patterns that are labeled as "good" PGA configurations. This study is a first step to generalize relevant patterns from the problems analyzed that identify better configurations in PGAs. [ABSTRACT FROM AUTHOR]

In present, data processing is an important process in many organizations. Classification trees are used to assign a classification to unknown data and can be also used for data partitioning (data clustering). The classification tree must be able to cope with outliers and have acceptably simple structure. An important advantage is the white-box structure. This paper presents a novel method called ACO-DTree for classification tree generation and their evolution inspired by natural processes. It uses a hybrid metaheuristics combining evolutionary strategies and ant colony optimization. Proposed method benefits from the stochastic process and population approach, which allows the algorithm to evolve more efficiently than the methods alone. The paper also consults the parameter estimation for the method. Tests on real data (UCI and MIT-BIH database) have been performed and evaluated. [ABSTRACT FROM AUTHOR]

In this work, it is shown that linear Cellular Automata based on rules 90/150 generate all the solutions of linear difference equations with binary constant coefficients. Some of these solutions are binary sequences with application in cryptography. In this sense, we propose CA-based linear models that realize the solutions of difference equations as well as behave as cryptographic keystream generators. Due to the simple transition rules that govern these CA, the implementation of such models is quite easy. Some illustrative examples complete the work. [ABSTRACT FROM AUTHOR]

A general method to learn the inverse kinematics of multi-link robots by means of neuro-controllers is presented. We can find analytical solutions for the most used and known robots in the bibliography. However, these solutions are specific to a particular robot configuration and are not generally applicable to other robot morphologies. The proposed method is general in the sense that it is not dependant on the robot morphology. We base our method in the Evolutionary Computation paradigm for obtaining incrementally better neuro-controllers. Furthermore, the proposed method solves some very specific issues in robotic neuro-controller learning. (1) It allows to escape from any neural network learning algorithm which relies on the classical supervised input-target learning scheme and hence it lets to obtain neuro-controllers without providing targets or correct answers which -in this case- are un known in prior. (2) It can converge beyond local optimal solutions which is one of the main drawbacks of some neural-network training algorithms based on gradient descent when applied to highly redundant robot morphologies. (3) Using learning algorithms such as the Neuro-Evolution of Augmenting Topologies (NEAT) it is also possible learning the neural network topology on-the-fly which is a common source of empirical testing in neuro-controllers design. Finally, experimental results are provided by applying the method in two multi-link robot learning tasks with a comparison between fixed and learnable topologies. [ABSTRACT FROM AUTHOR]

A method to generate movements in a snake robot using proportional-integral-derivate controllers (PID) and adjust the constants values to be natural is proposed. Specifically, the method is applied to adjust the movement of a snake robot to natural postures defining a simplify PID controller and adjusting the constants values of the controller. Our approach is based on proportional-integral-derivate controllers, using genetics algorithms to solve the problem. In this paper we explain how adjust the restrictions that must be accomplished for generate a natural movement and make an exhaustive study about snake robots, proportional-integral-derivate controllers and genetics algorithms. [ABSTRACT FROM AUTHOR]

Robot controller is one of the key facts affecting performance of robot. To meet the requirements of theoretical and application research, a mobile robot controller based on CAN Bus and Real-time Operating System (RTOS) is presented. In hardware aspect, a distributed architecture is achieved using CAN Bus and Bluetooth technology. Each module of the controller is integrated into a multi-agent based architecture, and is implemented in RTOS in software aspect. A comprehensive illustration of ATU-II mobile robot platform based on the proposed controller is presented and two experiments of ATU-II verify its performance. [ABSTRACT FROM AUTHOR]

In this study ensemble techniques have been applied in the frame of topology preserving mappings with visualization purposes. A novel extension of the ViSOM (Visualization Induced SOM) is obtained by the use of the ensemble meta-algorithm and a later fusion process. This main fusion algorithm has two different variants, considering two different criteria for the similarity of nodes. These criteria are Euclidean distance and similarity on Voronoi polygons. The goal of this upgrade is to improve the quality and robustness of the single model. Some experiments performed over different datasets applying the two variants of the fusion and other simpler models are included for comparison purposes. [ABSTRACT FROM AUTHOR]

The Multiple Classifier Systems are nowadays one of the most promising directions in pattern recognition. There are many methods of decision making by the ensemble of classifiers. The most popular are methods that have their origin in vote method, where the decision of the common classifier is a combination of individual classifiers decisions. This work presents method of weighted classifiers combination and experimental results of proposed algorithm. [ABSTRACT FROM AUTHOR]

In this paper we present a hybrid heuristic for the capacitated p-median problem (CPMP). This problem considers a set of n points, each of them with a known demand, the objective consists of finding p medians and assign each point to exactly one median such that the total distance of assigned points to their corresponding medians is minimized, and the a capacity limit on the medians may not be exceeded. The purpose of this paper is to present a new hybrid heuristic to solve the CPMP, called Clustering Search (CS), which consists in detecting promising search areas based on clustering. Computational results show that the CS found the best known solutions in all most instances. [ABSTRACT FROM AUTHOR]

There exist several approaches for solving clustering problems. Hopfield networks and self-organizing maps are the main neural approaches studied for solving clustering problems. Criticism of these approaches includes the tendency of the Hopfield network to produce infeasible solutions and the lack of generalization of the self-organizing approaches. Genetic algorithms are the other most studied bio-inspired approaches for solving optimization problems as the clustering problems. However, the requirement of tuning many internal parameters and operators is the main disadvantage of the genetic algorithms. This paper proposes a new technique which enables feasible solutions, removes the tuning phase, and improves solutions quality of clustering problems. Moreover, several biology inspired approaches are analyzed for solving traditional benchmarks. [ABSTRACT FROM AUTHOR]