GRASP algorithm for the unrelated parallel machine scheduling problem with setup times and additional resources

[EN] This paper provides practitioners with new approaches for solving realistic scheduling problems that consider additional resources, which can be implemented on expert and intelligent systems and help decision making in realistic settings. More specifically, we study the unrelated parallel machi...

Descripción completa

Detalles Bibliográficos
Autores: Yepes-Borrero, Juan C., Perea Rojas Marcos, Federico, Caballero-Villalobos, Juan Pablo, Villa Juliá, Mª Fulgencia|||0000-0003-0019-8777
Tipo de recurso: artículo
Fecha de publicación:2020
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/158835
Acceso en línea:https://riunet.upv.es/handle/10251/158835
Access Level:acceso abierto
Palabra clave:Unrelated parallel machines
Scheduling
Sequence dependent setup times
Makespan
Additional resources
GRASP
ESTADISTICA E INVESTIGACION OPERATIVA
Descripción
Sumario:[EN] This paper provides practitioners with new approaches for solving realistic scheduling problems that consider additional resources, which can be implemented on expert and intelligent systems and help decision making in realistic settings. More specifically, we study the unrelated parallel machine scheduling problem with setup times and additional limited resources in the setups (UPMSR-S), with makespan minimization criterion. This is a more realistic extension of the traditional problem, in which the setups are assumed to be done without using additional resources (e.g. workers). We propose three metaheuristics following two approaches: a first approach that ignores the information about additional resources in the constructive phase, and a second approach that takes into account this information about the resources. Computational experiments are carried out over a benchmark of small and large instances. After the computational analysis we conclude that the second approach shows an excellent performance, overcoming the first approach.