Heurísticas para o problema do caixeiro viajante branco e preto

This work describes a generalization of the Traveling Salesman Problem - TSP called Black and White Traveling Salesman Problem - BWTSP. The BWTSP is defined in a graph G (oriented or not), in a such way that the associated vertex set is partitioned into black and white vertices. The objective is to...

Descripción completa

Detalles Bibliográficos
Autor: Maciel, André Cordeiro Macedo
Tipo de recurso: tesis de maestría
Estado:Versión publicada
Fecha de publicación:2005
País:Brasil
Institución:Universidade Federal Fluminense (UFF)
Repositorio:Repositório Institucional da Universidade Federal Fluminense (RIUFF)
Idioma:portugués
OAI Identifier:oai:app.uff.br:1/17815
Acceso en línea:https://app.uff.br/riuff/handle/1/17815
Access Level:acceso abierto
Palabra clave:Metaheurística GRASP
Heurística
Caixeiro-viajante
GRASP
Regras de Redução
V ND
V NS
Heuristics
Metaheuristics
The traveling salesman problem
Reduction rules
CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO::COMPUTABILIDADE E MODELOS DE COMPUTACAO
Descripción
Sumario:This work describes a generalization of the Traveling Salesman Problem - TSP called Black and White Traveling Salesman Problem - BWTSP. The BWTSP is defined in a graph G (oriented or not), in a such way that the associated vertex set is partitioned into black and white vertices. The objective is to find a shortest hamiltonian tour subject to Cardinality and Length constraints. These constraints are related to the number of white vertices (cardinality) and the total distance (length) between two consecutive black vertices in a feasible solution. The main applications of this problem are found in the telecomunication area and the scheduling of airline operations that incorporate maintenance conections. In this work we introduce mathematical formulations for asymmetrical and symmetrical BWTSP, we propose news construction and local search algorithms that are used in the metaheuristics GRASP, V NS and V ND. Computational results show the viability of the proposed methods.