Thesis Modelo de localización de sensores para la agricultura de precisión.
Loading...
Date
2023-11-15
Authors
Journal Title
Journal ISSN
Volume Title
Program
Magíster en Ciencias de la Ingeniería Industrial
Campus
Campus Santiago Vitacura
Abstract
La agricultura de precisión es una disciplina que ha logrado aumentar la eficiencia, calidad y productividad de los cultivos agrícolas, así como también, disminuir el impacto ambiental de los mismos a través de la aplicación de tecnologías y análisis de datos. Por esta razón, la recopilación de la data es un paso clave en la implementación de estas tecnologías. Para ello, en el mercado existe una amplia variedad de sensores tanto fijos como móviles, cada uno con objetivos específicos y un amplio rango de precios.
Usualmente, los datos se recopilan de manera puntual y se utilizan técnicas de interpolación para poder tener una muestra en términos continuos del ´índice de interés, es por ello, que en este trabajo se busca determinar dónde localizar un número limitado de sensores de manera de aumentar la representatividad de la muestra obtenida.
A través de dos modelos de optimización lineal y en variables enteras binarias se busca encontrar las ubicaciones ´optimas para un número limitado de sensores y además determinar cuántos de ellos es necesario instalar para obtener un muestreo eficiente.
Ambos modelos son comparados en un caso de estudio resuelto a través del Solver Gurobi Optimizer en Julia JuMP.
Due to the benefits of using precision agriculture, many farmers want to install sensors in their fields. However, the high cost that a sensor could achieve limit the number of these to be acquired and to interpolate the information obtained to the whole extension of the field. For this reason it is interesting to determine where to locate a limited number of sensors, increasing the representativeness of the sample and, consequently, the reliability of the interpolation. Therefore, this paper addresses an optimization–based approach that integrates the management zone problem and a discrete location problem. The input data of the proposed model assumes a set of sample points of a soil or crop property to find an optimal location of a given number of sensors to obtain many others soil or crop properties in the resulting locations, and by interpolating the information obtained, reliable data about new or future properties can be obtained. The proposed model is solved using the integer programming solver Gurobi Optimizer in Julia JuMP and computational results from a set of instances are presented to show the impact of the adopted methodology.
Due to the benefits of using precision agriculture, many farmers want to install sensors in their fields. However, the high cost that a sensor could achieve limit the number of these to be acquired and to interpolate the information obtained to the whole extension of the field. For this reason it is interesting to determine where to locate a limited number of sensors, increasing the representativeness of the sample and, consequently, the reliability of the interpolation. Therefore, this paper addresses an optimization–based approach that integrates the management zone problem and a discrete location problem. The input data of the proposed model assumes a set of sample points of a soil or crop property to find an optimal location of a given number of sensors to obtain many others soil or crop properties in the resulting locations, and by interpolating the information obtained, reliable data about new or future properties can be obtained. The proposed model is solved using the integer programming solver Gurobi Optimizer in Julia JuMP and computational results from a set of instances are presented to show the impact of the adopted methodology.
Description
Keywords
Investigación de operaciones, Modelo de localización, Modelos de optimización