Transit travel strategy as solution of a Markov decision problem: Theory and applications

The search of optimal travel strategy on unreliable transit network is analyzed as solution of a Markov decision problem and two uses of this approach, recently presented by the authors, are recalled and compared. The first application is relative to normative strategy search, such as for path recommendation in innovative transit trip planners. The second use concerns subjective optimal strategy search in a dynamic strategy-based path choice modelling, especially suitable for real-time run-oriented simulation-based mesoscopic assignment models. The paper concludes with an overview of the benefits of the approach and outlines scopes for further research.