Edwin García
University of Antioquia
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Edwin García.
Información tecnológica | 2011
Fernando Villada; Edwin García; Juan D. Molina
Resumen Se propone un modelo para el pronostico del precio de la energia electrica en Colombia mediante el uso de redes neuro-difusas. Se utilizan dos estructuras de redes incluyendo como entradas la serie de precios diarios en la primera y la serie de precios mas el nivel medio de los embalses en la segunda. Los resultados son comparados con dos estructuras de redes neuronales y con un modelo Autoregresivo Condicional Heterocedastico Generalizado (GARCH). Los datos historicos fueron obtenidos de la Compania XM del Grupo ISA; datos para 120 dias son usados para entrenamiento y los 31 dias siguientes para verificar la capacidad predictiva del modelo. Se encontro ventajas en este ultimo dentro del periodo de muestreo para una variable de entrada, pero un mejor desempeno de las redes neuro-difusas en el periodo fuera de la muestra tanto para una como para dos variables de entrada. Palabras clave: precio de la electricidad, redes neuronales artificiales, redes neuro-difusas, modelos de series de tiempo
Información tecnológica | 2012
Carlos García; Edwin García; Fernando Villada
The optimization of the expansion planning of the real distribution system in the city of Medellin in Colombia, is analyzed in this work. The distribution system has 155 nodes and 154 branches at a voltage level of 13200 volts. Efficient evolutionary algorithms are applied, with the purpose of observing their improvements and performance when compared with conventional evolutionary algorithms. Besides, a mono-stage and mono-objective distribution planning model, which objective function is the optimization of construction costs and technical losses of the system is also studied. Finally, the impact of specially designed efficient operators for optimization is discussed and the estimated costs are analyzed.
Información tecnológica | 2012
Fernando Villada; Nicolás Muñoz; Edwin García
Resumen Este trabajo propone un modelo basado en redes neuronales artificiales para el pronostico de los precios de dos de las principales acciones transadas en mercado de valores colombiano. El modelo propuesto se aplica al estudio de las acciones de Ecopetrol y Preferencial Bancolombia, empresas que negocian en las bolsas de valores de Colombia y Nueva York. Se utilizan dos estructuras de redes incluyendo como entradas la serie de precios diarios en la primera y la serie de precios mas el indice del dolar estadounidense DXY en la segunda. Se prueban diferentes configuraciones de redes neuronales utilizando una serie de seis meses, donde los datos de los primeros cinco se utilizan para entrenamiento dejando el ultimo mes para verificar la capacidad predictiva de la red. Los resultados muestran un buen comportamiento de las redes neuronales con bajos errores en su desempeno tanto en aprendizaje como en prediccion. Palabras clave: mercado de valores, redes neuronales artificiales, pronostico de precios
Información tecnológica | 2015
Carlos García; Edwin García; Fernando Villada
espanolEn este trabajo se presenta la implementacion del algoritmo evolutivo multi-objetivo de frente de Pareto (SPEA) para la planeacion de la expansion de sistemas de distribucion, considerando como funciones objetivos los costos de la inversion y el numero de huecos de voltaje esperado por ano. El algoritmo se implementa aplicandolo a un sistema de distribucion real teniendo en cuenta algunos parametros fijos del sistema, tales como el numero de tramos, el numero de nodos y la poblacion inicial. Tambien se consideran otras variables que son propias de los algoritmos evolutivos como el porcentaje en poblaciones distintas que se cruzan entre si en una generacion (cruce) y la mutacion. Se concluye que la implementacion de SPEA para la planeacion de sistemas de distribucion reales es una buena y eficiente herramienta computacional cuando se requiere considerar varias funciones objetivo. EnglishThis article describes the implementation of the multi-objective evolutionary algorithm Pareto front (SPEA) for the planning of distribution system expansion, taking as target functions the investment costs and the number of Sag voltages expected per year. The algorithm is implemented by applying it to real distribution system, given some fixed system parameters such as the number of network segments, number of nodes and the number of initial population. Other variables that are proper of evolutionary algorithms such as percentage of generation population that crosses another (cross) and mutation are taken into consideration. It is concluded that the implementation of SPEA for planning real distribution systems is a good and efficient computational tool when it is necessary to consider several objective functions.
Poromechanics V: Proceedings of the Fifth Biot Conference on Poromechanics | 2013
Sayuri Kimoto; Edwin García; Fusao Oka
In order to study the effect of the water infiltration due to the rise of the water level of the rivers on the river dikes, a series of two-dimensional numerical analyses of experiments are carried out using a multi-phase seepage-deformation coupled method for unsaturated soil. The mechanism of the surface deformation and the strain localization on these soil structures are discussed mainly with respect to the water permeability of the soils. Results obtained by the simulations show that the deformation of the embankments significantly depends on the water permeability of the soil and it is localized on the slope surface at the river sides. The larger the saturated water permeability of the soil, the velocity of the seepage flow and the deformation on the surface of the river embankments are larger. The method adopted here, can effectively be used to study the practical seepage deformation coupled problems on unsaturated soils.
Archive | 2011
Fusao Oka; Edwin García; Sayuri Kimoto
It is known that the unstable behavior of unsaturated soil includes collapses behavior which is due to the decrease of suction as well as the shear failure. In addition, we often meet the numerical instability for the simulation of unsaturated soil during the wetting process. In the present study, we have performed a one-dimensional linear instability analysis for the air-water-soil coupled three-phase viscoplastic material. From the instability analysis of the viscoplastic unsaturated soil subjected to the water infiltration, we have found that the specific moisture capacity, the matric suction and the hardening parameter affect the instability. Then, we have performed numerical simulations of the behavior of unsaturated soil during the water infiltration using a multi-phase coupled elasto-viscoplastic finite element analysis method. The conditions under which the numerical instability occurs during the simulation agree well with the results by linear instability analysis.
International Journal for Numerical and Analytical Methods in Geomechanics | 2011
Edwin García; Fusao Oka; Sayuri Kimoto
Geosynthetics International | 2007
Edwin García; Chaminda Gallage; Taro Uchimura
Construction and Building Materials | 2016
Henry A. Colorado; Edwin García; M.F. Buchely
International Journal of Solids and Structures | 2010
Edwin García; Fusao Oka; Sayuri Kimoto