Search

The electrical energy supply relies on the satisfactory operation of insulators. The ultrasound recorded from insulators in different conditions has a time series output, which can be used to classify faulty insulators. The random convolutional kernel transform (Rocket) algorithms use convolutional filters to extract various features from the time series data. This paper proposes a combination of Rocket algorithms, machine learning classifiers, and empirical mode decomposition (EMD) methods, such as complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), empirical wavelet transform (EWT), and variational mode decomposition (VMD). The results show that the EMD methods, combined with MiniRocket, significantly improve the accuracy of logistic regression in insulator fault diagnosis. The proposed strategy achieves an accuracy of 0.992 using CEEMDAN, 0.995 with EWT, and 0.980 with VMD. These results highlight the potential of incorporating EMD methods in insulator failure detection models to enhance the safety and dependability of power systems.

Electroencephalography (EEG) is an exam widely adopted to monitor cerebral activities regarding external stimuli, and its signals compose a nonlinear dynamical system. There are many difficulties associated with EEG analysis. For example, noise can originate from different disorders, such as muscle or physiological activity. There are also artifacts that are related to undesirable signals during EEG recordings, and finally, nonlinearities can occur due to brain activity and its relationship with different brain regions. All these characteristics make data modeling a difficult task. Therefore, using a combined approach can be the best solution to obtain an efficient model for identifying neural data and developing reliable predictions. This paper proposes a new hybrid framework combining stacked generalization (STACK) ensemble learning and a differential-evolution-based algorithm called Adaptive Differential Evolution with an Optional External Archive (JADE) to perform nonlinear system identification. In the proposed framework, five base learners, namely, eXtreme Gradient Boosting, a Gaussian Process, Least Absolute Shrinkage and Selection Operator, a Multilayer Perceptron Neural Network, and Support Vector Regression with a radial basis function kernel, are trained. The predictions from all these base learners compose STACK’s layer-0 and are adopted as inputs of the Cubist model, whose hyperparameters were obtained by JADE. The model was evaluated for decoding the electroencephalography signal response to wrist joint perturbations. The variance accounted for (VAF), root-mean-squared error (RMSE), and Friedman statistical test were used to validate the performance of the proposed model and compare its results with other methods in the literature, including the base learners. The JADE-STACK model outperforms the other models in terms of accuracy, being able to explain around, as an average of all participants, 94.50% and 67.50% (standard deviations of 1.53 and 7.44, respectively) of the data variability for one step ahead and three steps ahead, which makes it a suitable approach to dealing with nonlinear system identification. Also, the improvement over state-of-the-art methods ranges from 0.6% to 161% and 43.34% for one step ahead and three steps ahead, respectively. Therefore, the developed model can be viewed as an alternative and additional approach to well-established techniques for nonlinear system identification once it can achieve satisfactory results regarding the data variability explanation.

Due to its capacity to gather vast, high-level data about human activity from wearable or stationary sensors, human activity recognition substantially impacts people’s day-to-day lives. Multiple people and things may be seen acting in the video, dispersed throughout the frame in various places. Because of this, modeling the interactions between many entities in spatial dimensions is necessary for visual reasoning in the action recognition task. The main aim of this paper is to evaluate and map the current scenario of human actions in red, green, and blue videos, based on deep learning models. A residual network (ResNet) and a vision transformer architecture (ViT) with a semi-supervised learning approach are evaluated. The DINO (self-DIstillation with NO labels) is used to enhance the potential of the ResNet and ViT. The evaluated benchmark is the human motion database (HMDB51), which tries to better capture the richness and complexity of human actions. The obtained results for video classification with the proposed ViT are promising based on performance metrics and results from the recent literature. The results obtained using a bi-dimensional ViT with long short-term memory demonstrated great performance in human action recognition when applied to the HMDB51 dataset. The mentioned architecture presented 96.7 ± 0.35% and 41.0 ± 0.27% in terms of accuracy (mean ± standard deviation values) in the train and test phases of the HMDB51 dataset, respectively.

Disruptive failures threaten the reliability of electric supply in power branches, often indicated by the rise of leakage current in distribution insulators. This paper presents a novel, hybrid method for fault prediction based on the time series of the leakage current of contaminated insulators. In a controlled high-voltage laboratory simulation, 15 kV-class insulators from an electrical power distribution network were exposed to increasing contamination in a salt chamber. The leakage current was recorded over 28 h of effective exposure, culminating in a flashover in all considered insulators. This flashover event served as the prediction mark that this paper proposes to evaluate. The proposed method applies the Christiano–Fitzgerald random walk (CFRW) filter for trend decomposition and the group data-handling (GMDH) method for time series prediction. The CFRW filter, with its versatility, proved to be more effective than the seasonal decomposition using moving averages in reducing non-linearities. The CFRW-GMDH method, with a root-mean-squared error of 3.44×10−12, outperformed both the standard GMDH and long short-term memory models in fault prediction. This superior performance suggested that the CFRW-GMDH method is a promising tool for predicting faults in power grid insulators based on leakage current data. This approach can provide power utilities with a reliable tool for monitoring insulator health and predicting failures, thereby enhancing the reliability of the power supply.

The predictive maintenance of electrical machines is a critical issue for companies, as it can greatly reduce maintenance costs, increase efficiency, and minimize downtime. In this paper, the issue of predicting electrical machine failures by predicting possible anomalies in the data is addressed through time series analysis. The time series data are from a sensor attached to an electrical machine (motor) measuring vibration variations in three axes: X (axial), Y (radial), and Z (radial X). The dataset is used to train a hybrid convolutional neural network with long short-term memory (CNN-LSTM) architecture. By employing quantile regression at the network output, the proposed approach aims to manage the uncertainties present in the data. The application of the hybrid CNN-LSTM attention-based model, combined with the use of quantile regression to capture uncertainties, yielded superior results compared to traditional reference models. These results can benefit companies by optimizing their maintenance schedules and improving the overall performance of their electric machines.

Contaminated insulators can have higher surface conductivity, which can result in irreversible failures in the electrical power system. In this paper, the ultrasound equipment is used to assist in the prediction of failure identification in porcelain insulators of the 13.8 kV, 60 Hz pin profile. To perform the laboratory analysis, insulators from a problematic branch are removed after an inspection of the electrical system and are evaluated in the laboratory under controlled conditions. To perform the time series predictions, the stacking ensemble learning model is applied with the wavelet transform for signal filtering and noise reduction. For a complete analysis of the model, variations in its configuration were evaluated. The results of root mean square error (RMSE), mean absolute percentage error (MAPE), mean absolute error (MAE), and coefficient of determination (R2) are presented. To validate the result, a benchmarking is presented with well-established models, such as an adaptive neuro-fuzzy inference system (ANFIS) and long-term short-term memory (LSTM).

The energy price influences the interest in investment, which leads to economic development. An estimate of the future energy price can support the planning of industrial expansions and provide information to avoid times of recession. This paper evaluates adaptive boosting (AdaBoost), bootstrap aggregation (bagging), gradient boosting, histogram-based gradient boosting, and random forest ensemble learning models for forecasting energy prices in Latin America, especially in a case study about Mexico. Seasonal decomposition of the time series is used to reduce unrepresentative variations. The Optuna using tree-structured Parzen estimator, optimizes the structure of the ensembles through a voter by combining several ensemble frameworks; thus an optimized hybrid ensemble learning method is proposed. The results show that the proposed method has a higher performance than the state-of-the-art ensemble learning methods, with a mean squared error of 3.37 × 10−9 in the testing phase.

Insulators installed outdoors are vulnerable to the accumulation of contaminants on their surface, which raise their conductivity and increase leakage current until a flashover occurs. To improve the reliability of the electrical power system, it is possible to evaluate the development of the fault in relation to the increase in leakage current and thus predict whether a shutdown may occur. This paper proposes the use of empirical wavelet transform (EWT) to reduce the influence of non-representative variations and combines the attention mechanism with a long short-term memory (LSTM) recurrent network for prediction. The Optuna framework has been applied for hyperparameter optimization, resulting in a method called optimized EWT-Seq2Seq-LSTM with attention. The proposed model had a 10.17% lower mean square error (MSE) than the standard LSTM and a 5.36% lower MSE than the model without optimization, showing that the attention mechanism and hyperparameter optimization is a promising strategy.

The cost of electricity and gas has a direct influence on the everyday routines of people who rely on these resources to keep their businesses running. However, the value of electricity is strongly related to spot market prices, and the arrival of winter and increased energy use owing to the demand for heating can lead to an increase in energy prices. Approaches to forecasting energy costs have been used in recent years; however, existing models are not yet robust enough due to competition, seasonal changes, and other variables. More effective modeling and forecasting approaches are required to assist investors in planning their bidding strategies and regulators in ensuring the security and stability of energy markets. In the literature, there is considerable interest in building better pricing modeling and forecasting frameworks to meet these difficulties. In this context, this work proposes combining seasonal and trend decomposition utilizing LOESS (locally estimated scatterplot smoothing) and Facebook Prophet methodologies to perform a more accurate and resilient time series analysis of Italian electricity spot prices. This can assist in enhancing projections and better understanding the variables driving the data, while also including additional information such as holidays and special events. The combination of approaches improves forecast accuracy while lowering the mean absolute percentage error (MAPE) performance metric by 18% compared to the baseline model.

Support Vector Machines (SVMs) technique for achieving classifiers and regressors. However, to obtain models with high accuracy and low complexity, it is necessary to define the kernel parameters as well as the parameters of the training model, which are called hyperparameters. The challenge of defining the more suitable value to hyperparameters is called the Parameter Selection Problem (PSP). However, minimizing the complexity and maximizing the generalization capacity of the SVMs are conflicting criteria. Therefore, we propose the Nature Inspired Optimization Tools for SVMs (NIOTS) that offers a method to automate the search process for the best possible solution for the PSP, allowing the user to quickly obtain several sets of good solutions and choose the one most appropriate for his specific problem. • The PSP has been modeled as a Multiobjective Optimization Problem (MOP) with two objectives: (1) good precision and (2) low complexity (low number of support vectors). • The user can evaluate multiple solutions included in the Pareto front, in terms of precision and low complexity of the model. • Apart from the Adaptive Parameter with Mutant Tournament Multiobjective Differential Evolution (APMT-MODE), the user can choose other metaheuristics and also among several kernel options.

This work presents a new software, programmed as a Python class, that implements a multi-objective Bayesian optimization algorithm. The proposed method is able to calculate the Pareto front approximation of optimization problems with fewer objective functions evaluations than other methods, which makes it appropriate for costly objectives. The software was extensively tested on benchmark functions for optimization, and it was able to obtain Pareto Function approximations for the benchmarks with as many as 20 objective function evaluations, those results were obtained for problems with different dimensionalities and constraints.

This data in brief presents displacement measurements acquired from a piezoelectric cantilevered actuator when subjected to harmonic excitations. The micro displacements are measured with optical sensors. The dataset has been used recently for the purpose of nonlinear black-box modelling, where the hysteretic behaviour of such devices has been modelled [1,2]. We hope to enable reproducibility by sharing the data used in [1,2], which are previous works by the authors, allowing the comparison of new methods on a common basis. Additionally, researchers interested in piezoelectric actuators for high precision tasks may also benefit on working with the present dataset. Keywords: Micromanipulators, System identification, Piezoelectric actuator, Mechatronics

The Nonlinear AutoRegressive with eXogenous inputs (NARMAX) models are among the most general classes of mathematical abstractions for dynamic systems and has many successful applications os data-driven modeling in different fields. In the present paper we introduce the narmax package in R for nonlinear black-box system identification using power-form polynomials. The goals are to provide the community with software which enables the resolution of nonlinear identification problems effectively, so practitioners can share their code with repeatable results, and to introduce a framework so one can build on to provide other identification methods in the rich R ecosystem. We aim at encapsulating procedures such as generating regression matrices, predicting free-run simulation, estimation of the parameters with standard and extended orthogonal least squares methods, and model validation utilities, so the user can focus most of the time and effort on building and testing different models.

Electric companies face flow control and inventory obstacles such as reliability, outlays, and time-consuming tasks. Convolutional Neural Networks (CNNs) combined with computational vision approaches can process image classification in warehouse management applications to tackle this problem. This study uses synthetic and real images applied to CNNs to deal with classification of inventory items. The results are compared to seek the neural networks that better suit this application. The methodology consists of fine-tuning several CNNs on Red–Green–Blue (RBG) and Red–Green–Blue-Depth (RGB-D) synthetic and real datasets, using the best architecture of each domain in a blended ensemble approach. The proposed blended ensemble approach was not yet explored in such an application, using RGB and RGB-D data, from synthetic and real domains. The use of a synthetic dataset improved accuracy, precision, recall and f1-score in comparison with models trained only on the real domain. Moreover, the use of a blend of DenseNet and Resnet pipelines for colored and depth images proved to outperform accuracy, precision and f1-score performance indicators over single CNNs, achieving an accuracy measurement of 95.23%. The classification task is a real logistics engineering problem handled by computer vision and artificial intelligence, making full use of RGB and RGB-D images of synthetic and real domains, applied in an approach of blended CNN pipelines.

Electricity price forecasting plays a vital role in the financial markets. This paper proposes a self-adaptive, decomposed, heterogeneous, and ensemble learning model for short-term electricity price forecasting one, two, and three-months-ahead in the Brazilian market. Exogenous variables, such as supply, lagged prices and demand are considered as inputs signals of the forecasting model. Firstly, the coyote optimization algorithm is adopted to tune the hyperparameters of complementary ensemble empirical mode decomposition in the pre-processing phase. Next, three machine learning models, including extreme learning machine, gradient boosting machine, and support vector regression models, as well as Gaussian process, are designed with the intent of handling the components obtained through the signal decomposition approach with focus on time series forecasting. The individual forecasting models are directly integrated in order to obtain the final forecasting prices one to three-months-ahead. In this case, a grid of forecasting models is obtained. The best forecasting model is the one that has better generalization out-of-sample. The empirical results show the efficiency of the proposed model. Additionally, it can achieve forecasting errors lower than 4.2% in terms of symmetric mean absolute percentage error. The ranking of importance of the variables, from the smallest to the largest is, lagged prices, demand, and supply. This paper provided useful insights for multi-step-ahead forecasting in the electrical market, once the proposed model can enhance forecasting accuracy and stability.

Load forecasting impacts directly financial returns and information in electrical systems planning. A promising approach to load forecasting is the Echo State Network (ESN), a recurrent neural network for the processing of temporal dependencies. The low computational cost and powerful performance of ESN make it widely used in a range of applications including forecasting tasks and nonlinear modeling. This paper presents a Bayesian optimization algorithm (BOA) of ESN hyperparameters in load forecasting with its main contributions including helping the selection of optimization algorithms for tuning ESN to solve real-world forecasting problems, as well as the evaluation of the performance of Bayesian optimization with different acquisition function settings. For this purpose, the ESN hyperparameters were set as variables to be optimized. Then, the adopted BOA employs a probabilist model using Gaussian process to find the best set of ESN hyperparameters using three different options of acquisition function and a surrogate utility function. Finally, the optimized hyperparameters are used by the ESN for predictions. Two datasets have been used to test the effectiveness of the proposed forecasting ESN model using BOA approaches, one from Poland and another from Brazil. The results of optimization statistics, convergence curves, execution time profile, and the hyperparameters’ best solution frequencies indicate that each problem requires a different setting for the BOA. Simulation results are promising in terms of short-term load forecasting quality and low error predictions may be achieved, given the correct options settings are used. Furthermore, since there is not an optimal global optimization solution known for real-world problems, correlations among certain values of hyperparameters are useful to guide the selection of such a solution.

The surface contamination of electrical insulators can increase the electrical conductivity of these components, which may lead to faults in the electrical power system. During inspections, ultrasound equipment is employed to detect defective insulators or those that may cause failures within a certain period. Assuming that the signal collected by the ultrasound device can be processed and used for both the detection of defective insulators and prediction of failures, this study starts by presenting an experimental procedure considering a contaminated insulator removed from the distribution line for data acquisition. Based on the obtained data set, an offline time series forecasting approach with an Adaptive Neuro-Fuzzy Inference System (ANFIS) was conducted. To improve the time series forecasting performance and to reduce the noise, Wavelet Packets Transform (WPT) was associated to the ANFIS model. Once the ANFIS model associated with WPT has distinct parameters to be adjusted, a complete evaluation concerning different model configurations was conducted. In this case, three inference system structures were evaluated: grid partition, fuzzy c-means clustering, and subtractive clustering. A performance analysis focusing on computational effort and the coefficient of determination provided additional parameter configurations for the model. Taking into account both parametrical and statistical analysis, the Wavelet Neuro-Fuzzy System with fuzzy c-means showed that it is possible to achieve impressive accuracy, even when compared to classical approaches, in the prediction of electrical insulators conditions.

The aim is to develop soft sensors (SSs) to provide an estimation of the laundry moisture of clothes introduced in a household Heat Pump Washer−Dryer (WD-HP) appliance. The developed SS represents a cost-effective alternative to physical sensors, and it aims at improving the WD-HP performance in terms of drying process efficiency of the automatic drying cycle. To this end, we make use of appropriate Machine Learning models, which are derived by means of Regularization and Symbolic Regression methods. These methods connect easy-to-measure variables with the laundry moisture content, which is a difficult and costly to measure variable. Thanks to the use of SSs, the laundry moisture estimation during the drying process is effectively available. The proposed models have been tested by exploiting real data through an experimental test campaign on household drying machines.

As abordagens de inteligência computacional, tais como sistemas nebulosos e redes neurais artificiais, têm-se gradualmente estabelecido como ferramentas robustas para a tarefa de aproximação de sistemas não-lineares complexos e previsão de séries temporais. Em aplicações envolvendo a área de Finanças, evidências empíricas anteriores indicam que modelos de inteligência computacional são mais precisos, dada sua maior capacidade em capturar não-linearidades e outros fatos estilizados presentes em séries financeiras. Nesse sentido, este artigo investiga a hipótese de que os modelos matemáticos de redes neurais perceptron multicamadas, redes neurais função de base radial e o sistema nebuloso Takagi-Sugeno (TAKAGI; SUGENO, 1985) são capazes de fornecer uma previsão fora-da-amostra mais acurada que os modelos auto-regressivos de médias móveis (ARMA) e auto-regressivo de médias móveis supondo heterocedasticidade condicional auto-regressiva (ARMA-GARCH). O desempenho de previsão um-passo-à-frente dos modelos foi comparado utilizando-se séries de retorno da taxa de câmbio real/dólar (R$/US$) com freqüências de 15 minutos, 60 minutos, 120 minutos, diária e semanal. Resultados indicam que o desempenho dos modelos está diretamente relacionado à freqüência observada das séries. Além disso, os modelos de redes neurais obtiveram um desempenho superior em relação aos demais modelos considerados. A avaliação da estratégia de negociação estabelecida com base nas previsões geradas pelos modelos indicou que estratégias baseadas em modelos de redes neurais forneceram retornos superiores em relação àquelas baseadas em modelos ARMA e ARMA-GARCH e também em relação à estratégia buy-and-hold.Computational intelligence approaches, such as artificial neural networks and fuzzy systems, have become popular tools in approximating complicated nonlinear systems and time series forecasting. In Finance applications, there is evidence that these computational intelligence models are able to provide a more accurate forecast given their capacity for capturing nonlinearities and other stylized facts of financial time series. Thus, this paper investigates the hypothesis that the mathematical models of multilayer perception, radial basis function neural networks (NN), and the Takagi-Sugeno (TS) fuzzy systems are able to provide a more accurate out-of-sample forecast than the traditional AutoRegressive Moving Average (ARMA) and ARMA Generalized AutoRegressive Conditional Heteroskedasticity (ARMA-GARCH) models. Using a series of Brazilian exchange rate (R$/US$) returns with 15 minutes, 60 minutes, 120 minutes, daily and weekly basis, the one-step-ahead forecast performance is compared. The results indicate that forecast performance is strongly related to the series' frequency, possibly due to nonlinearities effects. Besides, the forecasting evaluation shows that NN models perform better than the ARMA and ARMA-GARCH ones. In the trade strategy based on forecasts, NN models achieved higher returns when compared to a buy-and-hold strategy and to the other models considered in this study.

As ferramentas de identificação de sistemas e previsão de séries temporais permitem a concepção de modelos matemáticos baseados em dados numéricos. O problema essencial, nestes casos, é determinar o modelo matemático apropriado. Esse artigo apresenta o projeto de uma rede neural função de base radial (RN-RBF) para a previsão de séries temporais. Na utilização da RN-RBF para previsão de sistemas não-lineares é difícil determinar um conjunto apropriado de centros e aberturas para as funções de ativação Gaussianas para obter uma boa estrutura. Neste trabalho, a configuração da RN-RBF é baseada em uma abordagem híbrida baseada em método de agrupamento de dados de Gustafson-Kessel e procedimento de otimização usando evolução diferencial. O projeto de RN-RBF é validado para previsão de um passo à frente dos preços de troncos de eucalipto para celulose e serraria para ilustrar a eficiência da abordagem híbrida proposta. Além disso, o desempenho do projeto de RN-RBF baseado nos resultados de previsão é apresentado e discutido neste artigo.Computational tools of system identification and prediction of time series allows for the conception of mathematical models based on numerical data. The key problem in these cases is to find a suitable mathematical model. This paper presents a radial basis function neural network (RBF-NN) design for forecasting time series. Using the RBF-NN for nonlinear system forecasting is quite difficult as one has to choose an appropriate set of centers and spreads for the Gaussian activation functions to achieve a good network structure. In this work, the setup of RBF-NN is based on a hybrid method based on the Gustafson-Kessel clustering method and optimization procedure by differential evolution. The RBF-NN design is validated for the one-step ahead forecasting of eucalyptus wood prices for cellulose and sawmill to illustrate the effectiveness of this hybrid approach. The performance of the RBF-NN design based on forecasting results is presented and discussed in details.

A prevenção de fraude em cartão de crédito é uma importante aplicação comercial para aplicação de abordagens de métodos de previsão e inteligência computacional. A inteligência computacional é uma associação de metodologias bio-inspiradas que incluem, como principais membros, as redes neurais, sistemas nebulosos, computação evolutiva, inteligência coletiva e computação probabilística. Recentemente, a aplicação de técnicas da inteligência computacional no suporte de tarefas de serviço, tais como detecção e identificação de anomalia, classificação de padrão, diagnóstico, prognóstico, estimação e controle, tem emergido nos ambientes industrial e comercial. Este artigo apresenta um sistema computacional eficiente baseado em metodologias da inteligência computacional para detecção de fraude em operações reais de dados de cartão de crédito em transações de comércio eletrônico. O novo método proposto é denominado FControl® de detecção de fraude e classificação integra conceitos de sistemas inteligentes híbridos baseado em abordagens de redes neurais, sistemas nebulosos e computação evolutiva. O método proposto gera soluções de qualidade em termos de eficiência e sucesso de previsão. O programa computacional do FControl® está atualmente instalado em um Intel Pentium IV 2.4 MHz (bi-processador e RAM de 4 Gbytes) na companhia Ciashop E-Commerce e está correntemente em uso para detecção de fraude para 250 companhias com serviços de comércio eletrônico.The prevention of credit card fraud is an important commercial application for prediction methods and computational intelligence of pattern recognition approaches. Computational intelligence is an association of bio-inspired computational methodologies which are founded principally on neural networks, fuzzy systems, evolutionary computation, swarm intelligence, and probabilistic computation. The application of computational intelligence techniques in the support of service tasks such as anomaly detection and identification, pattern classification, diagnostics, prognostics, estimation and control has recently emerged in industrial and commercial environments. This paper presents an efficient system based on computational intelligence methodologies for the detection of fraud in the operation of real credit card data in E-commerce transactions. This new proposed method, called FControl® for fraud detection and classification, integrates concepts of hybrid intelligent systems based on neural networks, fuzzy systems and evolutionary computation approaches. The proposed method generates quality solutions in terms of prediction efficiency and success. The computational program of the FControl® system has been installed on an Intel Pentium IV 2.4 MHz (bi-processor and 4 Gbytes RAM) at the company Ciashop E-Commerce and is currently in use for fraud detection by 250 companies that offer E-commerce services.

Este artigo aborda uma solução computacional para o problema de layout industrial considerando hard-constraints não tratadas em trabalhos anteriores. O problema é solucionado em duas etapas. Na primeira, uma solução inicial baseada em heurística construtiva é gerada e na segunda, por intermédio da aplicação da meta-heurística busca tabu, a solução inicial é melhorada. Como contribuição, apresenta-se a ferramenta computacional denominada Ambiente Visual para Otimização de Layout Industrial (AVOLI), a fim de gerar layouts viáveis e eficientes capazes de tratar problemas de grande porte sujeitos a um conjunto de restrições realistas.This paper approaches a computational solving of the industrial layout problem considering hardconstraints not handled in previous works. The problem is solved in two steps: first a constructiveheuristic-based initial solution is generated and then tabu search heuristic is used to improve it. As a contribution, it shown a computational tool named AVOLI (Visual Environment for Optimization of Industrial Layout) that will aid planners and production engineers in the generation of layouts that contemplate real constraints.

The hygrothermal analysis of roofs is relevant due to the large areas exposed to a wide range of weather conditions, these directly affecting the energy performance and thermal comfort of buildings. However, after a long life service, the solar absorptivity coatings of roofs can be altered by mould accumulation. Based on two well established mathematical models, one that adopts driving potentials to calculate temperature, moist air pressure and water vapor pressure gradients, and the other to estimate the mould growth risk on surfaces, this research introduces an approach to predict mould growth considering a reduced computational effort and simulation time. By adopting multiple MISO (Multiple-Input, Single-Output) Nonlinear AutoRegressive with eXogenous inputs (NARX) models, a machine learning technique known as Least Squares Support Vector Machines (LS-SVM), a maximum margin model based on structural risk minimization, was used to predict vapor flux, sensible heat flux, latent heat flux, and mould growth risk on roof surfaces. The proposed model was validated in terms of the Multiple Correlation Coefficient (R2R2R2), Mean Square Error (MSE) and Mean Absolute Error (MAE) performance indices considering as input the weather file from Curitiba city—Brazil, showing consistent precision when compared to the results of a validated numerical model.