Aplicação de algoritmos bioinspirados e técnicas de geoprocessamento em python no posicionamento de rádio enlaces ponto a ponto

Abstract

This work aims to develop and compare optimization models for the placement of antennas in multiple radio links, considering geographical factors, coverage, interference, and other aspects relevant to the performance of telecommunication networks. For implementation, the region between the cities of Sobral and Massapê was selected to represent, respectively, the transmitting and receiving areas. Geospatial analysis was conducted using geoprocessing techniques implemented in Python to extract the necessary data for evaluating link configurations. The proposed methodology is based on the application and parameterization of bioinspired algorithms—specifically, Particle Swarm Optimization (PSO), Firefly Algorithm (FA), Biogeography-Based Optimization (BBO), and Non-dominated Sorting Genetic Algorithm II (NSGA-II)—which were compared with a random approach implemented using Matlab software, with the aim of highlighting their respective efficiencies. To ensure the robustness of the results, a Monte Carlo (MC) procedure was adopted, allowing for a statistically consistent evaluation of computational costs and solution quality. The results indicate that although the bioinspired algorithms may exhibit higher execution times in certain cases, they yield significantly lower average cost solutions compared to the random technique. Furthermore, a comparative analysis between swarm intelligence-based me- taheuristics (PSO and FA) and evolutionary computation-based approaches (BBO and NSGA-II) was conducted. The findings suggest that the choice of approach should be guided by the specific application: while swarm intelligence methods tend to produce more robust solutions despite higher computational costs, evolutionary computation techniques—particularly BBO—offer advantages when parameterization time is a critical constraint, even if the overall performance is relatively modest. Thus, the results of this study confirm that the use of bioinspired algorithms for optimizing antenna placement constitutes a promising alternative for enhancing the efficiency and robustness of telecommunication networks, thereby contributing to overcoming the limitations of traditional methods and fostering significant advancements in the field.