Meta-heurística aplicada ao planejamento de atividades de lotes de vagões em terminais ferroviários

• Autor(a) principal: Pinto, Gustavo de Luna
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
• Texto Completo: http://repositorio.ufes.br/handle/10/11297
• topic: Railway Terminal; Multiple modes; Setup time; Blocking; Activity Schedule; Ant Colony Optimization; Meta-heuristic; Múltiplos modos; Planejamento de atividades; Colônia de formigas; Tempo de setup; Retenção; Terminal ferroviário; Múltiplos modos; Tempo de setup; Retenção; Planejamento de atividades; Colônia de Formigas; Meta-heurística; Engenharia Civil; Otimização combinatória; Alocação de recursos; Ferrovias; Terminais ferroviários; Estações - Estações; 624

Railroad is very important for the Brazilian economy, especially in the transport of grains and ores. One area of study for railway operation is concerning rail yards and terminals, and a railway terminal is a yard specialized in products loading and unloading. In these terminals, railway wagons spend half of their service lives and carry out several activities, usually made by sets of wagons called lots. An activity in the terminal may have a predecessor, i.e., it requires the conclusion of another activity to start. In addition, it can be executed in more than one mode (Multi-mode), demanding different resources, which are expensive and scarce, therefore, shared. As delays in operation can generate fines, the goal is to minimize the total time that batches of wagons remain in the terminals (staying time), considering the limitation of terminal resources. The operation at terminals has other characteristics that increase its complexity, such as: setup time between specific activities; variation of resource availability over time (scheduled maintenance); and the blocking characteristic, which keeps a resource occupied by the lot even after the activity's end, due to the movement constraint of the batch imposed by rails and / or other facilities. Only one work was found in the literature considering the characteristics of the problem. It proposed a mathematical model based on the Resource Constrained Project Scheduling Problem (RCPSP). However, the model takes a long time to resolve instances, making it impractical for the dynamic of a rail terminal operation. Thus, a meta-heuristic based on the Ant Colony Optimization meta-heuristic (ACO) was developed in this research for the scheduling of activities of wagons' batches in rail terminals, aiming to minimize the staying time of the lots in the terminal and to do it with a processing time consistent with response time of the railway operation. Although there are many meta-heuristics in the literature addressing the RCPSP, including ACO, none were found, considering all the characteristics of the problem, neither addressing the concept of resource blocking with multiple resources per activity. The meta-heuristic generated results that were equal to or close to optimal, and for larger instances, it got results better than the mathematical model found in the literature. In addition, it took less than 2 minutes to do it, an operationally acceptable processing time, and far below the model execution time, which for the largest instances ran for more than 20 hours.