A transportation choice model on the commuter railroads using inverse reinforcement learning

Tomohiro Okubo, Naohiro Kitano, Akinori Morimoto*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


Conventional transportation policies for railroads have primarily focused on minimizing the negative utility, such as shortening the travel time and reducing congestion. However, with the recent introduction of trains with extra fares for greater comfort and changes in work styles, there is an increasing need to focus on the positive utility of travel itself. Moreover, advances in machine learning and artificial intelligence research have facilitated highly accurate and objective analysis from vast amounts of data. The purpose of this research is to construct a new transportation choice model using inverse reinforcement learning, which is a machine learning method, and to quantify the positive utility of commuter railroads. The results of a comparison of the proposed model with conventional methods indicate the advantages and disadvantages of the model. Further, a transportation choice model for railroads was created to understand the tendency of each selected train type.

Original languageEnglish
Article number100072
JournalAsian Transport Studies
Publication statusPublished - 2022 Jan


  • Inverse reinforcement learning
  • Machine learning
  • Positive utility of travel
  • Transportation choice model

ASJC Scopus subject areas

  • Transportation


Dive into the research topics of 'A transportation choice model on the commuter railroads using inverse reinforcement learning'. Together they form a unique fingerprint.

Cite this