Big data is becoming a research focus recently. Urban rail transit systems produce large amounts of data, such as real time train speed and position, passenger origin-destination (OD) information, etc. With the support of big data analytics, the rail transit operators will be able to improve the operation efficiency of rail transit systems. In this paper, we obtain the historical passenger OD data from the automatic fare collection system (AFC), and process these data to get the passenger arrival rate and passenger alighting proportion using Hadoop big data platform. A multi-objective model is proposed to optimize train schedule time table. The model consists of two submodel components, namely, train operation model and passenger demand model. We propose a parallel genetic algorithm (GA) using an adaptive crossover operator and mutation operator to obtain the optimal solution. The proposed model and solution method are evaluated using real-life data. The obtained results demonstrate the efficiency and accuracy of the proposed method.
Leveraging Big Data Analytics for Train Schedule Optimization in Urban Rail Transit Systems
2018-11-01
191817 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch