Relax and cut: limitantes duais para o problema do caixeiro viajante

• Ano de defesa: 2014
• Autor(a) principal: Kawashima, Makswell Seyiti [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Otimização matematica; Programação (Matemática); Problema do caixeiro viajante; Programming (Mathematics)
• Link de acesso: http://hdl.handle.net/11449/110602

The Traveling Salesman Problem (TSP) is a classical Combinatorial Optimization problem. Given a set of cities and travel costs between each pair of them, the objective is to find a tour through all the cities, visiting each city once, and returning to the city of origin with minimum total cost. The simple enunciate and non-trivial resolution enchanted many people through the years. In the literature various formulations for the Traveling Salesman Problem are presented, and the quality of the linear relaxation of such formulations is compared. The classical TSP formulation is strong, but have an exponencial number of constraints, and is equivalent to the multi-commodity formulation, of polinomial order. The computational cost to solve the linear relaxation of the multi-commodity formulation is high, stimulating the search of new ways of obtaining dual bounds. In the literature, procedures to obtain dual bounds to the TSP using the relax and cut technique are proposed, starting from the assignment problem (AP) and dualizing violated valid inequalities by the AP’s optimal solution. In this work, we propose an application of the relax and cut technique to the multi-commodity formulation for the TSP. The results obtained by the computational study are encouraging, with the implementation of an algorithm that generates good dual bounds in low running time