Search

In this article, we propose a novel flexible architecture, with different algorithmic procedures, for effective story segmentation of broadcast news from subtitle files. The proposed system exploits spatial and temporal distance, as well as sentence similarity, to classify different stories in news broadcasts. The computational algorithms which form the architecture mainly focus on each sentence's features (temporal distance, spatial distance, and semantic similarity), and are combined to build an overall classifier. The first algorithm in the architecture focuses on the segmentation task, detecting boundaries between news. The second and third algorithms identify high semantic correlation between pieces of text, whether they are consecutive in space or not. Video Text Track (VTT) subtitle files are used to evaluate the performance of the proposed approach, although any file format that includes temporal information could also be considered. These VTT files may contain text errors and inaccuracies, and the proposed algorithms have been designed to deal with noisy content.

Numerical association rule mining offers a very efficient way of mining association rules, where algorithms can operate directly with categorical and numerical attributes. These methods are suitable for mining different transaction databases, where data are entered sequentially. However, little attention has been paid to the time series numerical association rule mining, which offers a new technique for extracting association rules from time series data. This paper presents a new algorithmic method for time series numerical association rule mining and its application in smart agriculture. We offer a concept of a hardware environment for monitoring plant parameters and a novel data mining method with practical experiments. The practical experiments showed the method’s potential and opened the door for further extension.

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.cor.2008.08.012 Byline: Sancho Salcedo-Sanz, Emilio G. Ortiz-Garcia, Angel M. Perez-Bellido, Antonio Portilla-Figueras, Francisco Lopez-Ferreras Abstract: This paper proposes a linear programming (LP)-guided Hopfield-genetic algorithm for a class of combinatorial optimization problems which admit a 0-1 integer linear programming. The algorithm modifies the updating order of the binary Hopfield network in order to obtain better performance of the complete hybrid approach. We theoretically analyze several different updating orders proposed. We also include in the paper a novel proposal to guide the Hopfield network using the crossover and mutation operators of the genetic algorithm. Experimental evidences that show the good performance of the proposed approach in two different combinatorial optimization problems are also included in the paper. Author Affiliation: Department of Signal Theory and Communications, Universidad de Alcala de Henares, Campus Universitario, 28871 Alcala de Henares, Madrid, Spain Article Note: (footnote) [star] This work has been partially supported by Universidad de Alcala and Comunidad de Madrid with a Grant no. CCG07-UAH/TIC-1894, and by Ministerio de Educacion with a Grant no. TEC 2006-07010/TCM. A.M. Perez-Bellido and E.G. Ortiz-Garcia are supported by Universidad de Alcala, under the University F.P.I. grants program.

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.cor.2005.05.001 Byline: Sancho Salcedo-Sanz (a), Jing Su (b) Abstract: In this paper we present two strategies for reducing the time of convergence of two metaheuristics (genetic algorithms and simulated annealing) in inductive query by example (IQBE), which is a process for assisting the users of a given information retrieval system in the formulation of queries. Both strategies are based on a reduction of the search space size of the metaheuristics. The first strategy that we introduce is a compression-expansion strategy, where the terms are arranged into sets of a given size. The second strategy we consider is the so-called restricted search, where the number of terms in the genetic algorithm and simulated annealing encodings is fixed to be a given number m. We describe the implementation of the strategies and analyze when they can be successful, and the main drawbacks associated with them. Author Affiliation: (a) Sancho Salcedo-Sanz, Departamento de Teoria de la SeA[+ or -]al y Comunicaciones,Universidad de Alcala, Campus Universitario Carretera Madrid-Barcelona, km. 33, 28871 Alcala de Henares, Madrid, Spain (b) School of Computer Science, The University of Birmingham, B15 2TT, Edgbaston, Birmingham, UK

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.cor.2004.08.010 Byline: Sancho Salcedo-Sanz (a), Yong Xu (b), Xin Yao (b) Abstract: In this paper we tackle the task assignment problem (TSAP) in heterogeneous computer systems. The TSAP consists of assigning a given distributed computer program formed by a number of tasks to a number of processors, subject to a set of constraints, and in such a way a given cost function to be minimized. We introduce a novel formulation of the problem, in which each processor is limited in the number of task it can handle, due to the so called resource constraint. We propose two hybrid meta-heuristic approaches for solving this problem. Both hybrid approaches use a Hopfield neural network to solve the problem's constraints, mixed with a genetic algorithm (GA) and a simulated annealing for improving the quality of the solutions found. We test the performance of the proposed algorithms in several computational TSAP instances, using a GA with a penalty function and a GA with a repairing heuristic for comparison purposes. We will show that both meta-heuristics approaches are very good approaches for solving the TSAP. Author Affiliation: (a) Department of Signal Theory and Communications, Universidad Carlos III de, 28911 Leganes, Madrid, Spain (b) Centre for Research in Computational Intelligence and Applications (CERCIA), School of Computer Science, The University of Birmingham

The huge amount of data chronologically collected in short periods of time by different devices and technologies is an important challenge in the analysis of times series. This problem has produced the development of new automatic techniques to reduce the number of points in the resulting time series, in order to facilitate their processing and analysis. In this paper, we propose a new modification of a coral reefs optimization algorithm (CRO) to tackle the problem of reducing the size of the time series minimizing the approximation error. The modification includes a memetization procedure (hybridization with a local search procedure) of the standard algorithm to improve its quality when finding a promising solution in a given searching area. The memetization process is applied to the worse individuals of the algorithm at the beginning, and only to the best ones at the end of the algorithm’s convergence, resulting in a dynamical search approach called dynamic memetic CRO (DMCRO). The proposed DMCRO performance is compared in this paper against other state-of-the-art CRO algorithms, such as the standard one, its statistically driven version (SCRO) and two different hybrid versions (HCRO and HSCRO, respectively), and the standard memetic version (MCRO). All the algorithms compared have been tested in 15 time series approximation, collected from different sources, including financial problems, oceanography data, and cardiology signals, among others, showing that the best results are obtained by DMCRO.

Nowadays, one of the priorities of the European Commission is to reduce the environmental impact of aviation through the advanced design of novel aircraft configurations. This is of utmost importance in order to decrease the environmental footprint of aviation and to reduce fuel consumption and make airlines more profitable. This implies that new methods and tools for aerodynamic shape optimization will have to be developed, allowing aircraft configurations that cannot be obtained with traditional strategies. This paper focuses on the application of enhanced methods in aerodynamic shape design optimization to enable advanced aircraft configurations. In particular, this work aims to demonstrate the feasibility of the proposed strategy to reach optimal configurations that are far away from its baseline geometry. For this purpose, evolutionary algorithms are combined with support vector machines and applied to the optimization of a baseline geometry for different flow conditions. In particular, the selected application is based on the shape optimization problem of the landing gear master cylinder. Results pointed out the feasibility of the mentioned strategy to enable novel configurations within an aerodynamic shape optimization process.

The Community Detection Problem (CDP) in Social Networks has been widely studied from different areas such as Data Mining, Graph Theory Physics, or Social Network Analysis, among others. This problem tries to divide a graph into different groups of nodes (communities), according to the graph topology. A partition is a division of the graph where each node belongs to only one community. However, a common feature observed in real-world networks is the existence of overlapping communities , where a given node can belong to more than one community. This paper presents a new Multi-Objective Genetic Algorithm (MOGA-OCD) designed to detect overlapping communities, by using measures related to the network connectivity. For this purpose, the proposed algorithm uses a phenotype-type encoding based on the edge information, and a new fitness function focused on optimizing two classical objectives in CDP: the first one is used to maximize the internal connectivity of the communities, whereas the second one is used to minimize the external connections to the rest of the graph. To select the most appropriate metrics for these objectives, a comparative assessment of several connectivity metrics has been carried out using real-world networks. Finally, the algorithm has been evaluated against other well-known algorithms from the state of the art in CDP. The experimental results show that the proposed approach improves overall the accuracy and quality of alternative methods in CDP, showing its effectiveness as a new powerful algorithm for detecting structured overlapping communities. [ABSTRACT FROM AUTHOR]

In this paper, we describe two new methods to address the model type selection problem (MTSP) based on modifications of the coral reefs optimisation algorithm (CRO). The effectiveness of these novel approaches is subsequently illustrated in a problem of energy demand estimation in Spain. First, we describe how coral species can be defined in the CRO algorithm, so each specie defines a competing model for the MTSP. Second, we propose another method to solve MTSPs by modifying the original CRO with a substrate layer, so that the different models considered can be encoded similarly. This second method to solve the MTSP simplifies the application of the CRO operators. Finally, we evaluate the performance of the two CRO-based algorithms by solving a MTSP consisting of the prediction of future energy demand from macro-economic data in Spain as a case study.

Cyber attacks have recently gained momentum in the research community as a sharply concerning phenomenon further ignited by the proliferation of social networks, which unfold a variety of ways for cybercriminals to access compromised information of their users. This paper gravitates on impersonation attacks, whose motivation may go beyond economic interests of the attacker towards getting unauthorised access to information and contacts, as often occurs between teenagers and early users of social platforms. This manuscript proposes a meta-heuristically optimised learning model as the algorithmic core of a non-intrusive detection system that relies exclusively on connection time features to detect evidences of an impersonation attack. The proposed scheme hinges on the K-Means clustering approach applied to a set of time features specially tailored to characterise the usage of users, which are weighted prior to the clustering under detection performance maximisation criteria. The obtained results shed light on the potentiality of the proposed methodology for its practical application to real social networks.

Se propone un nuevo método de pronóstico para la columna total de ozono (CTO) basado en la combinación de algoritmos de vectores de soporte para regresión (VSR) y variables de predicción provenientes del satélite de colaboración nacional en órbita polar Suomi, así como de modelos numéricos del Sistema Global de Predicción (SGP) y mediciones directas. Los datos de satélite incluyen perfiles de temperatura y humedad a diferentes alturas, y mediciones de CTO realizadas en los días anteriores al pronóstico. El modelo SGP proporciona datos de temperatura y humedad para el día del pronóstico. El sistema también considera los datos alternos de mediciones in situ, p. ej. de la profundidad óptica de aerosoles a diferentes longitudes de onda. Mediante la metodología VSR se puede obtener un pronóstico exacto de la CTO a partir de estas variables de predicción, con mejores resultados que los obtenidos con otros métodos de regresión, p. ej. redes neuronales. También se efectúa un análisis del mejor subconjunto de características del pronóstico de CTO. La parte experimental de la investigación consiste en la aplicación de VSR a datos de observación directa obtenidos en el laboratorio radiométrico de Madrid, España, donde están disponibles mediciones de ozono adquiridas por medio de un espectrofotómetro Brewer, lo que posibilita el entrenamiento del sistema y la evaluación de sus resultados.

This paper develops a trustworthy deep learning model that considers electricity demand (G) and local climate conditions. The model utilises Multi-Head Self-Attention Transformer (TNET) to capture critical information from G, to attain reliable predictions with local climate (rainfall, radiation, humidity, evaporation, and maximum and minimum temperatures) data from Energex substations in Queensland, Australia. The TNET model is then evaluated with deep learning models (Long-Short Term Memory LSTM, Bidirectional LSTM BILSTM, Gated Recurrent Unit GRU, Convolutional Neural Networks CNN, and Deep Neural Network DNN) based on robust model assessment metrics. The Kernel Density Estimation method is used to generate the prediction interval (PI) of electricity demand forecasts and derive probability metrics and results to show the developed TNET model is accurate for all the substations. The study concludes that the proposed TNET model is a reliable electricity demand predictive tool that has high accuracy and low predictive errors and could be employed as a stratagem by demand modellers and energy policy-makers who wish to incorporate climatic factors into electricity demand patterns and develop national energy market insights and analysis systems.

This manuscript focuses on the increasing frequency and scales of worldwide wildfires and the need for enhancing the effectiveness of firefighting resources. The scope is focused on optimally deploying firefighting aircrafts on aerodromes and airports existing over an area based on fire risk predictions. This scenario is formulated as a capacity-constrained multi-objective optimisation problem where the utility of the deployed resources with respect to fire forest risk predictions is to be maximised, and expenditures associated with the reallocation of aircrafts must be minimised. This formulation is further complemented by including the impact of the distance from the wildfire to water sources in the firefighting utility function. To efficiently tackle this problem a multi-objective harmony search solver is designed and tested in synthetically generated and real scenarios for the Iberian Peninsula. The results obtained pave the way towards the utilisation of this tool by decision makers when outlining their firefighting logistics.

Abstract: This paper introduces the oriented-tree network design problem (OTNDP), a general problem of tree network design with several applications in different fields. We also present several adaptations needed by evolutionary algorithms with Cayley-type encodings to tackle the OTNDP. In particular, we present these adaptations in two Cayley-encodings known as Prüfer and Dandelion codes. We include changes in Cayley-encodings to consider rooted trees. We also show how to use a fixed-length encoding for Cayley codes in evolutionary algorithms, and how to guarantee that the optimal solution is included in the search space. Finally, we present several adaptations of the evolutionary algorithm’s operators to deal with Cayley-encodings for the OTNDP. In the experimental part of the paper, we compare the performance of an evolutionary algorithm (implementing the two Cayley-encodings considered) in several OTNDP instances: first, we test the proposed techniques in randomly generated instances, and second, we tackle a real application of the OTNDP: the optimal design of an interactive voice response system (IVR) in a call center. [Copyright &y& Elsevier]

Mobile telecommunications is nowadays one of the most successful technological fields nowadays. Since its introduction in the global market in the last decade of the previous century, it became a very important contributor to the Gross Domestic Product (GDP) of developed and developing countries. In this market, a large number of powerful international telecommunications companies, often in harsh competition among them, are making huge investments, with the aim of achieving large revenues accordingly. The search for profit maximization and the legacy position of the former state-owned monopolistic companies may generate conflicts and unfair practices (e.g. abuse of monopolistic power, anticompetition strategies etc.). The National Regulatory Authorities (NRA) of the different countries have to arbitrate and establish market rules through regulatory processes. In this paper, the foundations of telecommunications regulation are reviewed, and the evolution of mobile call termination charges is studied for three countries with very different socioeconomic and politic situations. The Long Run Incremental Cost (LRIC) approach emerges as the method of choice to evaluate costs and appears to lead to quite similar results as markets enter the maturity stage.

In this paper we present three meta-heuristic approaches for FPGA segmented channel routing problems (FSCRPs) with a new cost function in which the cost of each assignment is not known in advance, and the cost of a solution only can be obtained from entire feasible assignments. Previous approaches to FSCPs cannot be applied to this kind of cost functions, and meta-heuristics are a good option to tackle the problem. We present two hybrid algorithms which use a Hopfield neural network to solve the problem's constraints, mixed with a Genetic Algorithm (GA) and a Simulated Annealing (SA). The third approach is a GA which manages the problem's constraints with a penalty function. We provide a complete analysis of the three metaheuristics, by tested them in several FSCRP instances, and comparing their performance and suitability to solve the FSCRP.In this paper we present three meta-heuristic approaches for FPGA segmented channel routing problems (FSCRPs) with a new cost function in which the cost of each assignment is not known in advance, and the cost of a solution only can be obtained from entire feasible assignments. Previous approaches to FSCPs cannot be applied to this kind of cost functions, and meta-heuristics are a good option to tackle the problem. We present two hybrid algorithms which use a Hopfield neural network to solve the problem's constraints, mixed with a Genetic Algorithm (GA) and a Simulated Annealing (SA). The third approach is a GA which manages the problem's constraints with a penalty function. We provide a complete analysis of the three metaheuristics, by tested them in several FSCRP instances, and comparing their performance and suitability to solve the FSCRP.

A discrete-time quantized-state Hopfield neural network is analyzed with special emphasis in its convergence, complexity and scalability properties. This network can be considered as a generalization of the Hopfield neural network by Shrivastava et al. [27] into the interior of the unit hypercube. This extension allows its use in a larger set of combinatorial optimization problems and its properties make of it a good candidate to build hybrid algorithms along with other heuristics such as the evolutive algorithms. Finally, the network is illustrated in some instances of the linear assignment problem and the frequency assignment problem.

We propose two novel approaches for feature selection and ranking tasks based on simulated annealing (SA) and Walsh analysis, which use a support vector machine as an underlying classifier. These approaches are inspired by one of the key problems in the insurance sector: predicting the insolvency of a non‐life insurance company. This prediction is based on accounting ratios, which measure the health of the companies. The approaches proposed provide a set of ratios (the SA approach) and a ranking of the ratios (the Walsh analysis ranking) that would allow a decision about the financial state of each company studied. The proposed feature selection methods are applied to the prediction the insolvency of several Spanish non‐life insurance companies, yielding state‐of‐the‐art results in the tests performed. Copyright © 2005 John Wiley & Sons, Ltd.

In mask-alignment systems a reference signal is needed to align the mask with the silicon wafers. The optical reference signal is the autocorrelation of two two-dimensional (2D) codes with binary transmittance. For a long time, one-dimensional codes have been used in grating-measurement systems to obtain a reference signal. The design of this type of code has needed a great computational effort, which limits the size of the code to about 100 elements. Recently, we have applied genetic algorithms to design codes with arbitrary length. We propose the application of these algorithms to design 2D codes to generate 2D optical signals used in mask-alignment systems.

A new technique for the generation of optical reference signals with optimal properties is presented. In grating measurement systems a reference signal is needed to achieve an absolute measurement of the position. The optical signal is the autocorrelation of two codes with binary transmittance. For a long time, the design of this type of code has required great computational effort, which limits the size of the code to ?30 elements. Recently, the application of the dividing rectangles (DIRECT) algorithm has allowed the automatic design of codes up to 100 elements. Because of the binary nature of the problem and the parallel processing of the genetic algorithms, these algorithms are efficient tools for obtaining codes with particular autocorrelation properties. We design optimum zero reference codes with arbitrary length by means of a genetic algorithm enhanced with a restricted search operator.