Andres Mendez-Vazquez

A Self-Organization Algorithm for Robust Networking of Wireless Devices

This paper deals with distributed creation and maintaining of ad hoc networks including diverse wireless devices, namely sensors, PDA, cell phones, and video games consoles. A cluster-based self-organizing strategy is proposed for building a backbone among the mobile devices, detecting segmentation, and recovery. In this approach, each mobile device is controlled by a multi-role agent, which performs these tasks efficiently based only on local interactions; role management allows the backbone reconfiguration when the nodes leave or arrive to the network yielding a complex global emergent behavior. Energy saving is achieved by adapting the time interval and power of transmission after the network formation. Simulations showed that the distributed algorithms performance is close to that obtained by an optimizing global procedure.

Discovery of Possible Gene Relationships through the Application of Self-Organizing Maps to DNA Microarray Databases

DNA microarrays and cell cycle synchronization experiments have made possible the study of the mechanisms of cell cycle regulation of Saccharomyces cerevisiae by simultaneously monitoring the expression levels of thousands of genes at specific time points. On the other hand, pattern recognition techniques can contribute to the analysis of such massive measurements, providing a model of gene expression level evolution through the cell cycle process. In this paper, we propose the use of one of such techniques –an unsupervised artificial neural network called a Self-Organizing Map (SOM)–which has been successfully applied to processes involving very noisy signals, classifying and organizing them, and assisting in the discovery of behavior patterns without requiring prior knowledge about the process under analysis. As a test bed for the use of SOMs in finding possible relationships among genes and their possible contribution in some biological processes, we selected 282 S. cerevisiae genes that have been shown through biological experiments to have an activity during the cell cycle. The expression level of these genes was analyzed in five of the most cited time series DNA microarray databases used in the study of the cell cycle of this organism. With the use of SOM, it was possible to find clusters of genes with similar behavior in the five databases along two cell cycles. This result suggested that some of these genes might be biologically related or might have a regulatory relationship, as was corroborated by comparing some of the clusters obtained with SOMs against a previously reported regulatory network that was generated using biological knowledge, such as protein-protein interactions, gene expression levels, metabolism dynamics, promoter binding, and modification, regulation and transport of proteins. The methodology described in this paper could be applied to the study of gene relationships of other biological processes in different organisms.

Learning fuzzy rules through ant optimization, LASSO and Dirichlet Mixtures

In the area of fuzzy systems, one of the main problems is finding the set of rules that can give us the best results in specific problems. Further, the finding of this set is a combinatorial problem. There are several techniques for building these sets, but it is possible to group them in two main classes: The bottom-up approaches and the top-down approaches. This work proposes a new top-down approach to the fuzzy systems learning based in clustering and optimization techniques. The algorithm is split in two stages: First, it determines the fuzzy sets of each input and output linguistic variable, and second, it calculates the fuzzy rules from the obtained fuzzy sets. For the first part, a Dirichlet Mixture (DM) is used to cluster data to assign a fuzzy sets to each new cluster, since a fuzzy set can be seen as a generalized probability function, and hence the fuzzy sets of a given linguistic variable can be seen as a mixture of probabilities (a Gaussian Mixture). Then, an optimization problem is solved by using Ant Colony Optimization (ACO) to generate the minimum set of possible rules for classification by using a version of the Least Absolute Shrinkage and Selection Operator(LASSO) for the fitness function. This ACO was implemented in a CUDA GPU to deal with the combinatorial problem of rule generation. Finally, this new algorithm is used to attack the problem of color image segmentation.

Design and implementation of a robust wireless sensor network

This paper presents the design and implementation of a novel wireless sensor network system using the device Freescale MC1321X. The proposed network management system consists of two different modules: the first one performs network formation and maintenance under a policy of power consumption reduction, and the second one measures and collects sensory data over the built backbone. In the case of the network formation, a distributed algorithm based on a self-organization strategy is used. Then, the second module algorithm implements a sensor management information system to collect information about minimal, maximal, and average temperatures. For this, the system module first builds a tree using a given sink over the formed network. Then, it performs the tasks of measurement, processing, and transmission to the sink node. Finally, the experiments, which have been performed under different scenarios in open and closed environments, have provided enough information to demonstrate the advantages regarding low energy consumption, reorganization and reliability of the proposed wireless sensor network system.

A multi-objective PSO strategy for energy-efficient ad-hoc networking

In this work, virtual backbone generation in ad-hoc networks under constraints of limited energy resources is addressed through a novel global optimization method. It is based on the maximal independent set approach which is stated as a multi-objective optimization problem to represent the different functional constraints of the backbone generation. A discrete version of a Particle Swarm Optimization strategy is proposed for searching Pareto optimal solutions under the multi-objective cost function. Finally, a diversity-based indicator is applied to prove that the Pareto frontier has a uniform distribution in the non-dominated solutions.

Similarity-based method for reduction of fuzzy rules

Fuzzy Similarity Measures (FSMs) are widely used for comparison of fuzzy sets, as well as fuzzy rules. A multitude of different FSMs have been proposed so far. It is not straightforward to identify a single FSM that is the most suitable for a given task. In this paper, we investigate suitability of a few FSMs for the problem of reduction of number of rules for an image segmentation process. We use Dirichlet-based approach to generate fuzzy sets that are further used for construction of fuzzy if-then rules. We analyze similarity of these rules and select a specified number of rules for image segmentation purposes. We applied this approach to two different images.

A Novel Distributed Energy-Efficient Self-Organized Algorithm for Wireless Ad Hoc Networks

Wireless ad-hoc networks require a special management because of their hardware and energy limitations compared with wired networks. The problem of constructing a backbone structure over wireless ad-hoc networks has been widely researched. The basic problem is to minimize the wireless backbone size by taking into consideration the nodes capabilities. Therefore, an efficient, self-organized, scalable, and fault-tolerant algorithm is proposed, where node connection in the backbone is minimal. The proposed algorithm groups the node elements in clusters by means of a self-organization strategy based in four possible roles for each node in the network: leader, gateway, member and bridge. In order to show the performance of the algorithm, an implementation in the NS2 simulator is used. This simulation allows evaluating the structure built by the proposed algorithm. This allows showing that the proposed algorithm can obtain better results, in wireless backbone size and energy consumption, when comparing with MWAC algorithm.

Wireless Network Formation and Maintaining for Mobile Devices Based on Self-Organization Strategies

This paper deals with creation and maintenance of mobile devices networks. A multi-agent approach based on self-organization strategy is proposed for building a virtual backbone, segmentation detection, and network recovery. In this approach each mobile device is controlled by a multi-role agent, which performs these tasks efficiently by using local interactions; role management allows backbone reconfiguration when the nodes leave or arrive to the network yielding a complex global emergent behavior. Energy saving is achieved by adapting the time interval and power of transmission after the network formation. Simulations show the algorithm performance regarding energy saving and segmentation recovery.

Generation and Reduction of Fuzzy Sets with PG-Means and Fuzzy Similarity Measures

The probabilistic clustering techniques can be applied to generate fuzzy sets in situations where there is little or no information about data. Quite often, they generate a huge number of clusters. These clusters can be interpreted as fuzzy sets in a process of building a fuzzy system. A large number of fuzzy sets introduce noise to the fuzzy system, hence the need to reduce their number. Fuzzy Similarity Measures (FSMs) are widely used for comparison of fuzzy sets. Multiple FSMs have been proposed so far, but identifying a single FSM that is the most suitable for a given task is not always a straightforward process. On many occasions, FSMs are used to reduce a number of fuzzy sets. In this paper, we present the results of analyzing suitability of FSMs to reduce number of fuzzy sets and fuzzy if-then rules for an image segmentation problem. We use a PG-Means algorithm to generate fuzzy sets on both input and output variables. We propose and apply algorithms utilizing FSMs to reduce the number of fuzzy sets and rules. The paper includes a case study investigating the application of the proposed method on two images.

FuzzyVD: An algorithm that uses fuzzy logic and fuzzy systems to estimate the number of endmembers present in a hyperspectral image

The application fields of Hyperspectral Image (HI) analysis has been increasing in the last years because the availability of new devices and public data-sets. There are many published works demonstrating that it is possible to use hyperspectral imagery in order to detect targets and create material maps. Many of the proposed techniques require to have prior knowledge about the number of different materials present into the HI to be analyzed. This paper proposes a novel algorithm, FuzzyVD, to estimate the number of different materials in a given HI, which does not require parameters. The FuzzyVD algorithm provide a new approach to solve this problem and expands the application field of fuzzy logic into the HI analysis. This algorithm has been applied to real and synthetic images and the results conclude its robustness and dependability.