ALGORITMO GENÉTICO APLICADO AO PLANEJAMENTO DE REDES DE TELECOMUNICAÇÕES

Abstract

Telecommunication systems are in constant development and the increasing demand of users and new services have enabled the emergence of new technologies. Planning has become indispensable due to the competitiveness and the large amount of financial resources involved. This work aims to propose and evaluate a genetic optimization algorithm for the planning of telecommunications networks. Because it is a combinatorial problem, the objective is to evaluate the advantages and disadvantages of the model based on the genetic algorithm. The graphs representing the networks were encoded in incidence matrices and the genetic operators of crossing and mutation were designed to act on matrices. MATLAB® software was used as a computational tool to implement the algorithms. The proposed model minimizes cost, considering the constraints of demand and technical capacity. The results found are compared to the published results in the SNDlib network instance library. The evaluation of the first version of the algorithm was based on a small PDH (Plesiochronous Digital Hierarchy) instance. The gain obtained in the cost of this network, compared to the solution presented in the library using linear programming with an arc-path approach, is 15.15%. In the second step, the algorithm for the optimization of a larger SDH (Synchronous Digital Hierarchy) network was applied. In this case, the need to hybridize the initial algorithm with a postoptimization algorithm was identified. The results obtained for the larger network were close to that of the SNDlib network library, although they were not better. The results found are promising because they approach similar solutions at a substantially shorter execution time than the SNDlib reference time. New research must be done so that the proposed algorithm can give good answers to large networks due to this being the reality of this area of research.