Thesis Modelo de predicción de consumo de recursos computacionales
Loading...
Date
2024-08
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Informática
Departament
Campus
Campus Casa Central Valparaíso
Abstract
Este trabajo aborda la problemática de predecir el consumo de recursos computacionales en una máquina objetivo a partir del consumo de recursos y características una máquina base. En específico, el objetivo es predecir una de las métricas principales, el tiempo total de ejecución de un programa en la máquina objetivo, mediante la creación de un conjunto de datos diverso y representativo y la implementación de cuatro modelos de aprendizaje automático. Al compararlos con un estudio del estado del arte, se obtuvo un desempeño competitivo, donde el mejor modelo fue XGBoost, y la métrica utilizada para compararlos fue el Mean Absolute Percentage Error (MAPE).
This work addresses the problem of predicting the consumption of computational resources on a target machine based on the resource consumption and characteristics of a base machine. Specifically, the objective is to predict one of the main metrics, the total execution time of a program on the target machine, by creating a diverse and representative dataset and implementing four machine learning models. When compared to a state-of-the-art study, a competitive performance was achieved, where the best model was XGBoost, and the metric used for comparison was the Mean Absolute Percentage Error (MAPE).
This work addresses the problem of predicting the consumption of computational resources on a target machine based on the resource consumption and characteristics of a base machine. Specifically, the objective is to predict one of the main metrics, the total execution time of a program on the target machine, by creating a diverse and representative dataset and implementing four machine learning models. When compared to a state-of-the-art study, a competitive performance was achieved, where the best model was XGBoost, and the metric used for comparison was the Mean Absolute Percentage Error (MAPE).
Description
Keywords
Redes neuronales, Procesamiento de datos en tiempo real, Computación evolutiva
