Decision making strategies for real time trains movement planning / Estrategias de decisão para o planejamento de circulação de trens em tempo real

AUTOR(ES)
DATA DE PUBLICAÇÃO

2007

RESUMO

Railways plays a major role in freight and passenger transportation in the whole world. The Brazilian railway system has suffered a process of abandon and deterioration from 1960 to 1990. Since 1990 the privatization of the national railways brought new investments and in the last years the demand for railway transportation has increased significantly. Railway resources must be efficiently managed to match the increasingly transportation demand. This requires efficient strategic, tactical and operational planning. One of the main tasks at the operational planning level concerns train circulation and associated tools. Railway operation is a very dynamic process because trains are subject to many unexpected interferences. Computational tools to help trains circulation planning must provide solutions in a time range consistent with real-time needs. This work suggests algorithms for real-time train movement planning, using computational intelligence and fuzzy set theory methodology. One of the algorithms decides the preference between trains competing for a single line track at the same moment. The aim is to drive train circulation as dose as possible to reference trajectories supplied by human experts, global optimization algorithms or both. Other algorithm decides preference between trains and chose the velocity with which trains must travel to remain as dose as possible to its references. The third algorithm uses depth search algorithm to obtain a solution for train circulation problems. A comparative study considering the algorithms proposed herein and algorithms suggested in the literature. The comparative study is done using small railway system instances. Data of a major Brazilian railway is adopted to illustrate how the algorithms behave to solve larger instances. Results show that the algorithms here proposed obtain near optimal solutions for small instances and satisfactory solutions for the real case.

ASSUNTO(S)

train dispatch heuristics conjuntos difusos fuzzy control heuristica ferrovias - trafego trains movement planning heuristic search

Documentos Relacionados