Skip to main content

NUEVOS ALGORITMOS NEURO-EVOLUTIVOS PARA CLASIFICACION ORDINAL: APLICACIONES EN CLIMA, ENERGIAS LIMPIAS Y MEDIO AMBIENTE

PID2020-115454GB-C21

•
Funding agency name Agencia Estatal de Investigación
Funding agency acronym AEI
Program Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Call Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Call year 2020
State Plan 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Beneficiary centre UNIVERSIDAD DE ALCALA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 18
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Analysis, Characterization, Prediction and Attribution of Extreme Atmospheric Events with Machine Learning: a Review

  • Salcedo-Sanz S
  • Ascenso G.
  • Del Ser J.
  • Casillas-Pérez D.
  • Kadow C.
  • Fister D.
  • Barriopedro, David
  • García-Herrera, Ricardo
  • Giuliani M.
  • Castelletti A.
Atmospheric Extreme Events (EEs) cause severe damages to human societies and ecosystems. The frequency and intensity of EEs and other associated events are increasing in the current climate change and global warming risk. The accurate prediction, characterization, and attribution of atmospheric EEs is therefore a key research field, in which many groups are currently working by applying different methodologies and computational tools. Machine Learning (ML) methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric EEs. This paper reviews the ML algorithms applied to the analysis, characterization, prediction, and attribution of the most important atmospheric EEs. A summary of the most used ML techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. A number of examples is discussed and perspectives and outlooks on the field are provided., This research has been partially supported by the European Union, through H2020 Project “CLIMATE INTELLIGENCE Extreme events detection, attribution and adaptation design using machine learning (CLINT)”, Ref: 101003876-CLINT. This research has also been partially supported by the project PID2020-115454GB-C21 of the Spanish Ministry of Science and Innovation (MICINN). J. Del Ser also acknowledges support by the Basque Government through EMAITEK and ELKARTEK funds (ref. KK-2020/00049), as well as the consolidated research group MATHMODE (IT1294-19).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Autoencoder-based flow-analogue probabilistic reconstruction of heat waves from pressure fields

  • Pérez-Aracil, Jorge
  • Marina, Cosmin M
  • Zorita, Eduardo
  • Barriopedro, David
  • Zaninelli, Pablo
  • Giuliani, Matteo
  • Castelletti, Andrea
  • Gutiérrez, Pedro A
  • Salcedo-Sanz, Sancho
This paper presents a novel hybrid approach for the probabilistic reconstruction of meteorological fields based on the combined use of the analogue method (AM) and deep autoencoders (AEs). The AE-AM algorithm trains a deep AE in the predictor fields, which the encoder filters towards a compressed space of reduced dimensionality. The AM is then applied in this latent space to find similar situations (analogues) in the historical record, from which the target field can be reconstructed. The AE-AM is compared to the classical AM, in which flow analogues are explicitly searched in the fully resolved field of the predictor, which may contain useless information for the reconstruction. We evaluate the performance of these two approaches in reconstructing the daily maximum temperature (target) from sea-level pressure fields (predictor) recorded during eight major European heat waves of the 1950-2010 period. We show that the proposed AE-AM approach outperforms the standard AM algorithm in reconstructing the magnitude and spatial pattern of the considered heat wave events. The improvement ranges from 7% to 22% in skill score, depending on the heat wave analyzed, demonstrating the potential added value of the hybrid method., This research has been partially supported by the European Union, through the H2020 Project “CLIMATE INTELLIGENCE Extreme events detection, attribution and adaptation design using machine learning (CLINT)”, Ref: 101003876-CLINT. This work has also been partially supported by “Agencia Española de Investigación (España)” (grant references: PID2020-115454GB-C21 and PID2020-115454GB-C22/AEI/10.13039/501100011033)., Peer reviewed
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

A flexible architecture using temporal, spatial and semantic correlation-based algorithms for story segmentation of broadcast news

  • Palomo Alonso, Alberto
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Jiménez Fernández, Silvia|||0000-0002-2065-1754
  • Portilla Figueras, José Antonio|||0000-0001-6569-6780
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
In this article, we propose a novel flexible architec- ture, 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 com- putational 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., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

A fuzzy-based cascade ensemble model for improving extreme wind speeds prediction

  • Peláez Rodríguez, César|||0000-0003-1260-8112
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Prieto Godino, Luis
  • Ghimire, S.
  • Deo, Ravinesh C.
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Se propone un novedoso conjunto en cascada de modelos de regresión basado en lógica difusa para abordar el problema de la predicción de eventos de velocidad extrema del viento, utilizando datos procedentes de modelos de reanálisis atmosférico. Aunque este problema ha sido abordado mayoritariamente en el contexto de tareas de clasificación, la innovación de este trabajo radica en tratarlo dentro de un dominio predictivo continuo, con el objetivo de proporcionar una estimación precisa de los valores extremos de velocidad del viento. El marco propuesto, de tipo jerárquico, implica una partición sucesiva de los datos de entrenamiento en clústeres difusos (fuzzy-soft) en función del valor de la variable objetivo, seguida del entrenamiento de un modelo de regresión específico dentro de cada clúster designado, de modo que cada modelo pueda analizar una parte concreta del dominio objetivo. Finalmente, las predicciones realizadas por los modelos individuales se integran en un conjunto basado en lógica difusa, en el que se asigna a cada modelo un valor de pertinencia en función de la predicción obtenida en la capa previa y de las funciones de pertenencia definidas para cada clúster. Para encontrar la forma óptima de realizar la partición de los datos, se adopta un algoritmo de optimización por Evolución Diferencial (DE). Como esquemas finales de regresión se emplean métodos de redes neuronales aleatorizadas de entrenamiento rápido. El rendimiento de la metodología propuesta se ha evaluado mediante su comparación con métodos del estado del arte, utilizando datos reales de tres parques eólicos en España., A novel fuzzy-based cascade ensemble of regression models is proposed to address a problem of extreme wind speed events forecasting, using data from atmospheric reanalysis models. Although this problem has been mostly explored in the context of classification tasks, the innovation of this paper arises from tackling a continuous predictive domain, aiming at providing an accurate estimation of the extreme wind speed values. The proposed layered framework involves a successive partition of the training data into fuzzy-soft clusters according to the target variable value, and further training a specific regression model within each designated cluster, so that each model can analyze a particular part of the target domain. Finally, predictions made by individual models are integrated into a fuzzy-based ensemble, where a pertinence value is designated to each model based on the previous layer’s prediction, and on the defined membership functions for each cluster. A Differential Evolution (DE) optimization algorithm is adopted to find the optimal way to perform data partitioning. Fast training randomized neural networks methods are used as final regression schemes. The performance of the proposed methodology has been assessed by comparison against state-of-the-art methods in real data from three wind farms in Spain., European Commission, Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Machine learning regression and classification methods for fog events prediction

  • Castillo Botón, Carlos
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Ghimire, Sujan
  • Cerro Prada, E.
  • Gutiérrez Peña, Pedro Antonio
  • Deo, Ravinesh C.
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Los eventos atmosféricos de baja visibilidad suelen estar asociados con la formación de niebla. De hecho, los eventos de baja visibilidad extrema afectan profundamente el transporte aéreo y terrestre, los aeropuertos y las carreteras y autovías, causando accidentes y problemas de tráfico cada año. Los algoritmos de aprendizaje automático (ML) se han aplicado con éxito a muchos problemas de predicción de formación de niebla y baja visibilidad. El problema asociado a la predicción de este tipo de eventos se puede formular como una regresión o como una tarea de clasificación, lo que tiene un impacto en el tipo de enfoque de ML que se utilizará, así como en la calidad de las predicciones obtenidas. En este artículo, realizamos un análisis completo de los problemas de predicción en eventos de baja visibilidad, formulados como problemas de regresión y clasificación. Discutimos el funcionamiento de una gran cantidad de algoritmos de ML en cada tipo de problema y evaluamos su funcionamiento en término de calidad de las soluciones obtenidas, bajo un marco de comparación común. De acuerdo con los resultados obtenidos, se proporcionan indicaciones sobre cuál es la formulación más eficiente para abordar las predicciones de baja visibilidad, y los enfoques de ML que obtienen mejores soluciones para la predicción de este tipo de eventos., Atmospheric low-visibility events are usually associated with fog formation. Extreme low-visibility events deeply affect the air and ground transportation, airports and motor-road facilities causing accidents and traffic problems every year. Machine Learning (ML) algorithms have been successfully applied to many fog formation and low-visibility prediction problems. The associated problem can be formulated either as a regression or as a classification task, which has an impact on the type of ML approach to be used and on the quality of the predictions obtained. In this paper we carry out a complete analysis of low-visibility events prediction problems, formulated as both regression and classification problems. We discuss the performance of a large number of ML approaches in each type of problem, and evaluate their performance under a common comparison framework. According to the obtained results, we will provide indications on what the most efficient formulation is to tackle low-visibility predictions and the best performing ML approaches for low-visibility events prediction., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Hybrid deep CNN-SVR algorithm for solar radiation prediction problems in Queensland, Australia

  • Ghimire, Sujan
  • Bhandari, Binayak
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Deo, Ravinesh C.
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Este estudio propone un nuevo modelo híbrido de aprendizaje profundo (DL), llamado CSVR, para realizar la predicción de radiación solar global (GSR) mediante la integración de una red neuronal convolucional (CNN) y una máquina de vectores soporte para regresión (SVR). Primero, se utiliza un algoritmo basado en una red Convolucional (CNN) para extraer patrones locales, así como características comunes que ocurren recurrentemente en datos de series de tiempo en diferentes intervalos. Posteriormente, se aplica un algoritmo de SVR para reemplazar las capas de CNN completamente conectadas, y predecir así los datos de series de tiempo GSR diarios en seis parques solares en Queensland, Australia. Para desarrollar el modelo híbrido CSVR, se han considerado considerado las variables meteorológicas más adecuadas de la base de datos de modelos climáticos globales, e información científica para propietarios de tierras en Australia. De esta manera, a partir de un conjunto de variables de modelos climáticos globales y observaciones terrestres, se seleccionan las características óptimas a través de un algoritmo metaheurístico para selección de características. Los hiperparámetros del modelo CSVR propuesto se optimizan mediante el método HyperOpt, y el rendimiento general del algoritmo propuesto se ha comparado con ocho métodos de DL alternativos y algunos de los otros métodos de aprendizaje automático (LSTM, DBN, RBF, BRF, MARS, WKNNR, GPML y M5TREE). Los resultados obtenidos muestran que el modelo CSVR propuesto puede ofrecer varias ventajas predictivas sobre los modelos DL alternativos, así como sobre los modelos ML convencionales. Específicamente, observamos que el modelo CSVR registró un error cuadrático medio/error absoluto medio que oscila entre 2,172?3,305 MJ m2/1,624?2,370 MJ m2 en los seis parques solares probados en comparación con 2,514?3,879 MJ m2/1,939?2,866 MJ m2 de los algoritmos ML y DL alternativos. En consonancia con este error obtenido, la correlación entre el GSR medido y el previsto, incluidos los coeficientes de Willmott & Nash-Sutcliffe así como el índice de Legates & McCabe, fue relativamente mayor para el modelo CSVR propuesto en comparación con otros métodos de aprendizaje automático y de DL para todos los sitios de estudio. En consecuencia, este estudio muestra el buen funcionamiento del modelo CSVR para proporcionar una alternativa viable en la predicción de la GSR en la explotación de energía solar, la demanda de energía u otras aplicaciones basadas en pronósticos de tipo meteorológico., This study proposes a new hybrid deep learning (DL) model, the called CSVR, for Global Solar Radiation (GSR) predictions by integrating Convolutional Neural Network (CNN) with Support Vector Regression (SVR) approach. First, the CNN algorithm is used to extract local patterns as well as common features that occur recurrently in time series data at different intervals. Then, the SVR is subsequently adopted to replace the fully connected CNN layers to predict the daily GSR time series data at six solar farms in Queensland, Australia. To develop the hybrid CSVR model, we adopt the most pertinent meteorological variables from Global Climate Model and Scientific Information for Landowners database. From a pool of Global Climate Models variables and ground-based observations, the optimal features are selected through a metaheuristic Feature Selection algorithm, an Atom Search Optimization method. The hyperparameters of the proposed CSVR are optimized by mean of the HyperOpt method, and the overall performance of the objective algorithm is benchmarked against eight alternative DL methods, and some of the other Machine Learning approaches (LSTM, DBN, RBF, BRF, MARS, WKNNR, GPML and M5TREE) methods. The results obtained shows that the proposed CSVR model can offer several predictive advantages over the alternative DL models, as well as the conventional ML models. Specifically, we note that the CSVR model recorded a root mean square error/mean absolute error ranging between 2.172–3.305 MJ m2/1.624–2.370 MJ m2 over the six tested solar farms compared to 2.514–3.879 MJ m2/1.939–2.866 MJ m2 from alternative ML and DL algorithms. Consistent with this predicted error, the correlation between the measured and the predicted GSR, including the Willmott’s, Nash-Sutcliffe’s coefficient and Legates & McCabe’s Index was relatively higher for the proposed CSVR model compared to other DL and Machine Learning methods for all of the study sites. Accordingly, this study advocates the merits of CSVR model to provide a viable alternative to accurately predict GSR for renewable energy exploitation, energy demand or other forecasting-based applications., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Solving an energy resource management problem with a novel multi-objective evolutionary reinforcement learning method

  • Matos Cardoso Leite, Gabriel|||0000-0002-1486-346X
  • Jiménez Fernández, Silvia|||0000-0002-2065-1754
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
  • Gil Marcelino, Carolina
  • Pedreira, C.E.
Microgrids have become popular candidates for integrating diverse energy sources into the power grid as means of reducing fossil fuel usage. Energy Resource Management (ERM) is a type of Unit Commitment problem, where a player operates a microgrid with diverse renewable generators integrated with an external supplier. Calculating the economic dispatch of each committed unit on a planning horizon is an NP-hard problem, and therefore, finding an exact solution is difficult. This paper presents a multi-objective solution to the ERM problem from the perspective of battery operation and external supplier dispatch. First, a novel multi-objective decision problem modeling is proposed that considers three objectives: cost, greenhouse gas emissions, and battery degradation. This framework involves a learning agent that controls the depth of discharge of a Lithium-Ion battery. To address the proposed problem, a new multi-objective algorithm called Multi-Objective Evolutionary Policy Search (MEPS) is introduced. The proposed algorithm uses NeuroEvolution of Augmenting Topologies structure to evolve artificial neural networks for estimating action-preference values considering multi-objective rewards. The MEPS performance is evaluated on both standard and newly-proposed benchmark problems, using the hypervolume as the evaluation metric. When compared to standard deep reinforcement learning, results showed that MEPS provides cost-effective, environmentally friendly, and efficient energy storage management solutions. Furthermore, MEPS effectively solves the proposed ERM problem by finding neural networks with a small number of nodes and connections, which are suitable for use in embedded control systems. Overall, MEPS proved to be a promising multi-objective approach in the transition to clean energy resources., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Deep learning ensembles for accurate fog-related low-visibility events forecasting

  • Peláez Rodríguez, César|||0000-0003-1260-8112
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • López Diz, Alba de
  • Casanova Mateo, C.
  • Fister, Dusan
  • Jiménez Fernández, Silvia|||0000-0002-2065-1754
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
In this paper we propose and discuss different Deep Learning-based ensemble algorithms for a problem of low-visibility events prediction due to fog. Specifically, seven different Deep Learning (DL) architectures have been considered, from which multiple individual learners are generated. Hyperparameters of the models, including parameters concerning data preprocessing, models architecture and training procedure, are randomly selected for each model within a pre-defined discrete range. Also, every model is trained with slightly different data sampled randomly, assuring that every models introduce variety in the ensemble. Then, three different information fusion techniques are employed to build the ensemble models. The influence of the filtering process and the elitism level (the percentage of the individual mod- els entering the ensemble) is also assessed. The performance of the proposed methodology have been tested in two real problems of low-visibility events prediction due to orographical and radiation fog, at the north of Spain. Comparison with different Machine Learning, alternative DL algorithms and meteorological-based methods show the good performance of the proposed deep learning ensembles in this problem., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

A versatile multi-method ensemble for wind farm layout optimization

  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Jiménez Fernández, Silvia|||0000-0002-2065-1754
  • Prieto Godino, Luis
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
In this paper, the performance of a multi-method ensemble for optimization in a turbine layout problem for wind farms is discussed. Specifically, the Coral Reefs Optimization algorithm with Substrate Layer (CRO-SL) is evaluated. The CRO-SL is an evolutionary-type meta-heuristic, able to combine different search procedures or methods within a single population, leading to a multi-method ensemble. The CRO-SL has been adapted to the wind turbines layout problem by including a discrete encoding, based on sparse matrices (though it can also work over continuous encodings) and a convex geometric procedure to speed-up the wake effect calculation for each turbine. The performance of the proposed CRO-SL algorithm in the turbine layout problem is evaluated by means of different experiments, considering simulated and real data from a wind farm in Spain. The performance of the CRO-SL is also evaluated in a different wind turbine layout problem from a NREL-IEA case competition. The contribution of each substrate of the CRO-SL algorithm (search procedure) is discussed in terms of the global search ability of the algorithm in the different problems considered. The global performance of the proposed approach has been finally compared to that of existing approaches in the literature, obtaining excellent results in all scenarios considered., Agencia Estatal de Investigación, Comunidad de Madrid
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Integrated Multi-Head Self-Attention Transformer model for electricity demand prediction incorporating local climate variables

  • Ghimire, Sujan
  • Nguyen-Huy, T.
  • Al-Musaylh, M. S.
  • Deo, Ravinesh C.
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Este artículo desarrolla un modelo de aprendizaje profundo para la predicción de demanda de electricidad a partir de datos y variables climáticas locales. El modelo utiliza un algoritmo conocido como Multi-Head Self-Attention Transformer (TNET) para capturar información crítica de la demanda de electricidad, y lograr así predicciones fiables con datos de variables atmosféricas locales como lluvia, radiación solar, humedad, evaporación y temperaturas máximas y mínimas de las subestaciones de Energex en Queensland, Australia. Posteriormente, el modelo TNET se evalúa con modelos de aprendizaje profundo (LSTM, LSTM Bidireccional, y redesd GRU, Redes Neuronales Convolucionales CNN y Redes Neuronales Profundas DNN) basados en métricas de evaluación de modelos robustos. El método de Estimación de Densidad Kernel se utiliza asimismo para generar el intervalo de predicción (PI) de los pronósticos de demanda de electricidad y derivar métricas de probabilidad y resultados para demostrar que el modelo TNET desarrollado es preciso para todas las subestaciones. El estudio concluye que el modelo TNET propuesto es una herramienta muy fiable para predecir la demanda de electricidad, con alta precisión y bajos errores de predicción, y podría ser empleado como estrategia por gestores de demanda eléctrica, así como gestores de políticas energéticas que deseen incorporar factores climáticos en los patrones de demanda de electricidad, y desarrollar sistemas de análisis e información del mercado energético nacional., This paper develops a trustworthy deep learning model that considers electricity demand ( ) and local climate conditions. The model utilises Multi-Head Self-Attention Transformer (TNET) to capture critical information from , 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., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Machine learning as alternative strategy for the numerical prediction of the shear response in reinforced and prestressed concrete beams

  • Hernández Díaz, Alejandro Mateo
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Lorente Ramos, Eugenio
  • Marina, Cosmin Madalin|||0000-0002-5849-6673
  • Peláez Rodríguez, César|||0000-0003-1260-8112
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Algunos materiales, como el hormigón armado y preesforzado, implican relaciones constitutivas no lineales en los problemas de elasticidad definidos en ellos. En particular, la resistencia de cizalla de una viga de hormigón armado puede calcularse considerando un campo de puntales diagonales en el contexto de las denominadas ?Teorías de Campos de Compresión? (CFT). Este trabajo presenta un método eficiente de Machine Learning (ML), alternativo a los métodos numéricos, para obtener la respuesta completa a la resistencia de cizalla en vigas de hormigón armado y preesforzado basado en CFT, con respecto a tensiones, deformaciones y ángulos de fisuración. Para ello, se desarrolla una tarea de regresión utilizando modelos de ML de última generación. Se entrena un modelo ML por variable de salida con la base de datos de soluciones Newton-Raphson existente. También se considera la región de solubilidad del modelo constitutivo de acero embebido, lo que demuestra el carácter integral del método propuesto. El modelo se valida en dos respuestas de vigas reales, donde los resultados obtenidos demuestran que este método alternativo basado en algoritmos ML aborda de manera efectiva el problema de predicción de la resistencia de cizalla en vigas de hormigón armado y preesforzado. La aproximación de ML propuesta funciona razonablemente bien, sin requerir ninguna aproximación inicial. Además, el regresor ML desarrollado muestra una baja dependencia del área de rigidización por tracción del hormigón que rodea el refuerzo de acero, lo que mejora significativamente el rendimiento de esta metodología en un mayor número de casos de diseño. Por lo tanto, en el contexto práctico de la ingeniería estructural, este último enfoque establece un procedimiento alternativo eficiente para obtener soluciones numéricas de modelos mecánicos basados en el marco de las CFT., Some materials, such as reinforced and prestressed concrete, involve non-linear constitutive relationships in elasticity problems defined on them. In particular, the shear strength of a reinforced concrete beam may be calculated by considering a diagonal struts field in the context of the so-called “Compression Field Theories” (CFTs). This work presents an efficient Machine Learning method alternative to numerical methods for obtaining the full shear response of reinforced and prestressed concrete beams based on CFT regarding stresses, strains, and crack angles. For that, a regression task is developed using state-of-the-art Machine Learning (ML) models. A ML model per output variable is trained with the existing Newton-Raphson solutions database. The solvability region of the embedded steel constitutive model is also considered, demonstrating the comprehensive character of the proposed method. The model is validated on two real beam responses, where the results obtained demonstrate that this alternative method based on ML algorithms effectively addresses the problem of prediction of the shear response in reinforced and prestressed concrete beams. The proposed ML approximation performs reasonably well without requiring any initial approximations. Moreover, the ML regressor here developed shows a low dependence on the concrete tension stiffening area surrounding the steel reinforcement, which significantly improves the performance of this methodology in a higher number of design cases. Thus, in the practical context of structural engineering, this last approach establishes an efficient alternative procedure for obtaining numerical solutions of mechanical models based on the CFTs framework., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Deep learning CNN-LSTM-MLP hybrid fusion model for feature optimizations and daily solar radiation prediction

  • Ghimire, Sujan
  • Deo, Ravinesh C.
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
  • Sharma, Ekta
  • Ali, Mohamed
La predicción de la radiación solar global (GSR) desempeña un papel esencial en la planificación, el control y la monitorización de los sistemas de energía solar. Sin embargo, su comportamiento estocástico es un reto importante para conseguir resultados de predicción satisfactorios. Este artículo tiene como objetivo diseñar un modelo de predicción híbrido innovador, que integra un mecanismo de selección de características utilizando un algoritmo metaheurístico de tipo Slime-Mould, una red neuronal convolucional (CNN), una red neuronal de memoria a corto y largo plazo (LSTM) y una CNN final, con salida de perceptrón multicapa (algoritmo SCLC). El modelo SCLC propuesto se aplicó a seis parques solares en Queensland (Australia) en horizontes temporales diarios y considerando seis horizontes de predicción diferentes. La evaluación comparativa exhaustiva de los resultados obtenidos con los de dos modelos de aprendizaje profundo (CNN-LSTM, red neuronal profunda) y tres modelos de aprendizaje automático (red neuronal artificial, bosque aleatorio, máquinas de aprendizaje extremo diferencial-evolutivo autoadaptativo) destacó un mayor rendimiento del modelo de predicción propuesto en los seis parques solares seleccionados. A partir de los resultados obtenidos, este trabajo establece que el algoritmo SCLC diseñado podría tener una utilidad práctica para aplicaciones en la gestión de recursos energéticos renovables y sostenibles., Global solar radiation (GSR) prediction plays an essential role in planning, controlling and monitoring solar power systems. However, its stochastic behaviour is a significant challenge in achieving satisfactory prediction results. This study aims to design an innovative hybrid prediction model that integrates a feature selection mechanism using a Slime-Mould algorithm, a Convolutional-Neural-Network (CNN), a Long–Short-Term-Memory Neural Network (LSTM) and a final CNN with Multilayer-Perceptron output (SCLC algorithm hereafter). The proposed model was applied to six solar farms in Queensland (Australia) at daily temporal horizons in six different time steps. The comprehensive benchmarking of the obtained results with those from two Deep-Learning (CNN-LSTM, Deep-Neural-Network) and three Machine-Learning (Artificial-Neural-Network, Random-Forest, Self-Adaptive Differential-Evolutionary Extreme-Learning-Machines) models highlighted a higher performance of the proposed prediction model in all the six selected solar farms. From the results obtained, this work establishes that the designed SCLC algorithm could have a practical utility for applications in renewable and sustainable energy resource management., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

A general explicable forecasting framework for weather events based on ordinal classification and inductive rules combined with fuzzy logic

  • Peláez Rodríguez, César|||0000-0003-1260-8112
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Marina, Cosmin Madalin|||0000-0002-5849-6673
  • Prieto Godino, Luis
  • Casanova Mateo, Carlos
  • Gutiérrez, Pedro Antonio
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Este artículo presenta un método para proporcionar explicabilidad en la integración de técnicas de inteligencia artificial (IA) y minería de datos aplicadas a la predicción meteorológica. La inteligencia artificial explicable (XAI) se refiere a la transparencia de los sistemas de IA al ofrecer explicaciones sobre sus predicciones y procesos de toma de decisiones, y contribuye a mejorar la precisión de las predicciones y a aumentar la confianza en los sistemas de IA. El enfoque de este trabajo se centra en los desafíos de interpretabilidad en problemas de clasificación ordinal dentro de la predicción del tiempo. La clasificación ordinal implica predecir fenómenos meteorológicos con clases ordenadas, como rangos de temperatura, velocidad del viento, niveles de precipitación, entre otros. Para abordar este reto, se propone un marco de predicción explicable, novedoso y general, que combina reglas inductivas y lógica difusa. Las reglas inductivas, derivadas de datos meteorológicos históricos, proporcionan una base lógica e interpretable para la predicción, mientras que la lógica difusa gestiona la incertidumbre y la imprecisión inherentes a los datos meteorológicos. El sistema predice un conjunto de probabilidades de pertenencia de la muestra entrante a cada una de las clases consideradas. Además, permite reforzar el proceso de toma de decisiones experto al apoyarse en la transparencia y la explicabilidad física del modelo, y no únicamente en la salida de un algoritmo de caja negra. El marco propuesto se evalúa utilizando dos bases de datos meteorológicas reales relacionadas con la velocidad del viento y con eventos de baja visibilidad debidos a la niebla. Los resultados se comparan tanto con clasificadores de aprendizaje automático como con métodos específicos para problemas de clasificación ordinal, obteniéndose resultados muy competitivos en términos de métricas de rendimiento ordinal, a la vez que se ofrece un mayor nivel de explicabilidad y transparencia en comparación con los enfoques existentes., This paper presents a method for providing explainability in the integration of artificial intelligence (AI) and data mining techniques when dealing with meteorological prediction. Explainable artificial intelligence (XAI) refers to the transparency of AI systems in providing explanations for their predictions and decision-making processes, and contribute to improve prediction accuracy and enhance trust in AI systems. The focus of this paper relies on the interpretability challenges in ordinal classification problems within weather forecasting. Ordinal classification involves predicting weather phenomena with ordered classes, such as temperature ranges, wind speed, precipitation levels, and others. To address this challenge, a novel and general explicable forecasting framework, that combines inductive rules and fuzzy logic, is proposed in this work. Inductive rules, derived from historical weather data, provide a logical and interpretable basis for forecasting; while fuzzy logic handles the uncertainty and imprecision in the weather data. The system predicts a set of probabilities that the incoming sample belongs to each considered class. Moreover, it allows the expert decision-making process to be strengthened by relying on the transparency and physical explainability of the model, and not only on the output of a black-box algorithm. The proposed framework is evaluated using two real-world weather databases related to wind speed and low-visibility events due to fog. The results are compared to both ML classifiers and specific methods for ordinal classification problems, achieving very competitive results in terms of ordinal performance metrics while offering a higher level of explainability and transparency compared to existing approaches., European Commission, Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Efficient prediction of fog-related low-visibility events with Machine Learning and evolutionary algorithms

  • Peláez Rodríguez, César|||0000-0003-1260-8112
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Casanova Mateo, C.
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Low visibility events are a severe problem for road transport, causing accidents and major economic losses. Their accurate prediction may help prevent these problems. For that purpose, machine and deep learning techniques have been applied for fog prediction using in situ meteorological data and persistence variables as baseline predictors. These techniques have been evaluated for different prediction time-horizons: 1 h, 3 h and 6 h. The effect of including data extracted from ERA5 Reanalysis as predictive variables has been studied. A database, covering 23 months, has been used, which contains visibility and other meteorogical variables measured in Mondon similar to edo, Galicia, Spain. A 222000 km2 region around Mondon similar to edo has been delimited. Thus, a proposed iterative forward selection algorithm based on evolutionary algorithms has been applied to determine the optimal variables and nodes in the region for each regressor model. Both Differential Evolution and Particle Swarm Optimization have been used as optimization algorithms, and an improvement of up to 17.3% with respect to the baseline databases have been obtained. Finally, an analysis of the most frequently selected variables by the evolutionary algorithms has been conducted, leading us to conclude which variables and geographical nodes provide better information to the prediction models., Ministerio de Ciencia e Innovación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Hybridizing machine learning algorithms with numerical models for accurate wind power forecasting

  • Abad Santiago, Álvaro
  • Peláez Rodríguez, César|||0000-0003-1260-8112
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Sanz Justo, J.
  • Casanova Mateo, C.
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Una predicción precisa de la generación de energía eólica es fundamental para optimizar la integración de esta fuente en la red eléctrica y garantizar la fiabilidad del suministro energético. Esta investigación se centra en mejorar la precisión de las predicciones de generación eólica mediante la combinación de datos procedentes de modelos de mesoescala y de reanálisis con enfoques de Aprendizaje Automático (ML). Se utilizaron datos de predicción del modelo WRF junto con datos de reanálisis ERA5 para estimar la generación de energía eólica de un parque eólico situado en Valladolid, España. El estudio evaluó el rendimiento de modelos de ML basados individualmente en datos WRF y ERA5, así como un modelo combinado que emplea entradas de ambos conjuntos de datos. El modelo híbrido que integra datos WRF y ERA5 con ML obtuvo una mejora del 15 % en el error cuadrático medio (RMSE) y un incremento del 10 % en el coeficiente de determinación ? 2 R 2 en comparación con los modelos independientes, proporcionando una predicción de la generación eólica a 1 hora más fiable. Además, se abordó la disponibilidad temporal de los datos: WRF ofrece la ventaja de proyectar información hacia el futuro, mientras que ERA5 proporciona datos retrospectivos., An accurate prediction of wind power generation is crucial for optimizing the integration of wind energy into the power grid, ensuring energy reliability. This research focuses on enhancing the accuracy of wind power generation forecasts by combining data from mesoscale and reanalysis models with Machine Learning (ML) approaches. We utilized WRF forecast data alongside ERA5 reanalysis data to estimate wind power generation for a wind farm located at Valladolid, Spain. The study evaluated the performance of ML models based on WRF and ERA5 data individually, as well as a combined model using inputs from both datasets. The hybrid model combining WRF and ERA5 data with ML resulted in a 15% improvement in root mean square error (RMSE) and a 10% increase in R2 compared with standalone models, providing a more reliable 1-h forecast of wind power generation. Additionally, the availability of data over time was addressed: WRF provides the advantage of projecting data into the future, whereas ERA5 offers retrospective data., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review

  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
  • Pérez Aracil, Jorge|||0000-0002-4456-9886
  • Ascenso, Guido
  • Ser Lorente, Javier del
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Kadow, Christopher
  • Fister, Dusan
  • Barriopedro, David
  • García Herrera, David
  • Giuliani, Matteo
  • Castelletti, Andrea
Atmospheric extreme events cause severe damage to human societies and ecosystems. The frequency and intensity of extremes and other associated events are continuously increasing due to climate change and global warming. The accurate prediction, characterization, and attribution of atmospheric extreme events is, therefore, a key research field in which many groups are currently working by applying different methodologies and computational tools. Machine learning and deep learning methods have arisen in the last years as powerful techniques to tackle many of the problems related to atmospheric extreme events. This paper reviews machine learning and deep learning approaches applied to the analysis, characterization, prediction, and attribution of the most important atmospheric extremes. A summary of the most used machine learning and deep learning techniques in this area, and a comprehensive critical review of literature related to ML in EEs, are provided. The critical literature review has been extended to extreme events related to rainfall and floods, heatwaves and extreme temperatures, droughts, severe weather events and fog, and low-visibility episodes. A case study focused on the analysis of extreme atmospheric temperature prediction with ML and DL techniques is also presented in the paper. Conclusions, perspectives, and outlooks on the field are finally drawn., European Commission, Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

CoSeNet: a novel approach for optimal segmentation of correlation matrices

  • Palomo Alonso, Alberto
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Jiménez Fernández, Silvia|||0000-0002-2065-1754
  • Portilla Figueras, José Antonio|||0000-0001-6569-6780
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
In this paper, we propose a novel approach for the optimal identification of correlated segments in noisy correlation matrices. The proposed model is known as CoSeNet (Correlation Segmentation Network) and is based on a four-layer algorithmic architecture that includes several processing layers: input, formatting, re-scaling, and segmentation layer. The proposed model can effectively identify correlated segments in such matrices, better than previous approaches for similar problems. Internally, the proposed model utilizes an overlapping technique and uses pre-trained Machine Learning (ML) algorithms, which makes it robust and generalizable. CoSeNet approach also includes a method that optimizes the parameters of the re-scaling layer using a heuristic algorithm and fitness based on a Window Difference-based metric. The output of the model is a binary noise-free matrix representing optimal segmentation as well as its segmentation points and can be used in a variety of applications, obtaining compromise solutions between efficiency, memory, and speed of the proposed deployment model., Agencia Estatal de Investigación
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

A multi-algorithm approach for operational human resources workload balancing in a last mile urban delivery system

  • Moreno Saavedra, Luis Miguel|||0009-0004-9211-8491
  • Jiménez Fernández, Silvia|||0000-0002-2065-1754
  • Portilla Figueras, José Antonio|||0000-0001-6569-6780
  • Casillas Pérez, David|||0000-0002-5721-1242
  • Salcedo Sanz, Sancho|||0000-0002-4048-1676
Efficient workload assignment to the workforce is critical in last-mile package delivery systems. The explosive increase of e-commerce and last-mile package logistics after the COVID-19 pandemic has produced, among other issues, difficulties in balancing the daily workload of the workforce in many delivery zones. In this context, traditional methods of assigning package deliveries to workers based on geographical proximity can be inefficient and surely guide to an unbalanced workload distribution among delivery workers. In this paper, we look at the problem of operational human resources workload balancing in last-mile urban package delivery systems. The idea is to consider the effort workload to optimize the system, i.e., the optimization process is now focused on improving the delivery time, so that the workload balancing is complete among all the staff. This process should correct significant decompensations in workload among delivery workers in a given zone. Specifically, we propose a multi-algorithm approach to tackle this problem. The proposed approach takes as input a set of delivery points and a defined number of workers, and then assigns packages to workers, in such a way that it ensures that each worker completes a similar amount of work per day. The proposed algorithms use a combination of distance and workload considerations to optimize the allocation of packages to workers. In this sense, the distance between the delivery points and the location of each worker is also taken into account to minimize the travel time of the workers. The proposed multi-algorithm methodology includes different versions of k-means, evolutionary approaches, recursive assignments based on k-means initialization with different problem encodings, and a hybrid evolutionary ensemble algorithm. We have successfully illustrated the performance of the proposed approach in a real-world problem of human resource balancing in an urban last-mile package delivery workforce operating at Azuqueca de Henares, Guadalajara, Spain., Ministerio de Ciencia e Innovación, Cátedra de Innovación Correos - Universidad de Alcalá