Avaliação da Meta-heurística VNS para um problema de planejamento operacional do transporte público

• Ano de defesa: 2014
• Autor(a) principal: Sakiyama, Rubens Zenko
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual de Maringá; Brasil; Departamento de Informática; Programa de Pós-Graduação em Ciência da Computação; UEM; Maringá, PR; Centro de Tecnologia
• Palavras-chave em Português: Transporte público; Escalonamento; Algoritmos; Algoritmos heurísticos; Meta-heurística VNS; Problema de escalonamento de motoristas (PEM); Brasil.; Public transport; Bus driver; Scheduling problem; Large real-world; Heuristic; VNS Metaheuristic; Brazil.; Ciências Exatas e da Terra; Ciência da Computação
• Link de acesso: http://repositorio.uem.br:8080/jspui/handle/1/2498

The Bus Driver Scheduling Problem (BDSP) consists to generate a set of drivers schedule to cover a set of vehicles schedule at the lowest cost, satisfying constraints imposed by labor laws, trade union agreements and company standards. This process is vital to the operational planning of public transportation companies since the drivers cost afIects a significant portion ofthe overall cost ofthe company. Considered NP-Hard, several works address the resolution of PEM through heuristic algorithms due to the limitations of the exact algorithms to work with large instances. The present work propose a approach for solving the BDSP involving two local search procedures in a neighborhood structure, called PCR and k-swap, in a deterministic way and in conjunction with VNS meta-heuristic. To validate the work is proposed real instances with over 2300 travei and random instances extracted of real instances. The experiments demonstrated the efficacy of VNS meta-heuristic for large instances, where the present results are compared with results reported by other studies that used PCR and k-swap procedures without the use ofVNS meta-heuristic.