Uma meta-heurística baseada em Algoritmo Genético para Seleção de Oficiais de Justiça por Distritos

Abstract

This work proposes a solution to the problem of selecting bailiffs by districts (PSB/D) at the central of warrants in Maranhão, Brazil. The solution, named AGSOJ/D, is based on a meta-heuristic called Genetic Algorithms (GA). The complete solution to the PSB/D problem is a Web-based software called GAPSB/D. Were conducted various experiments to prove its applicability considering two scenarios. The first one represents the central of warrants in the city of Imperatriz devised by six districts and 27 bailiffs. The second one depicts the central of warrants in the city of São Luís composed by 16 districts and 84 bailiffs. The results of the GAPSB/D are compared against three approaches: the manual selection currently in use, GAPSB/D-X with crossover operator, and Simulated Annealing (SA). Results have shown that GAPSB/D could create solutions as good as the manual process, and it is more stable than GAPSB/D-X and SA, especially in the second scenario, in which the size of the search space is 2 1344 containing ≈ 1, 10 ∗ 1089 possibilities of feasible solutions. Moreover, the Appendix A of this work includes the software design of the GAPSB/D, which contains both functional and non-functional requirements, use cases, and class diagrams. Further, presents the technologies and tools used in the development of the software GAPSB/D