Transit Networks

We developed more general analytical models for optimizing a number of transit systems (corridors and networks) that do not rely on mountains of input data.

Optimal transit service atop ring-radial and grid street networks

Which of the following two street networks is more suitable for designing efficient transit network? And can you tell why?

Hints:

Which network exhibits more degrees of freedom for design? Which network is more adaptable to the spatial distribution of passenger flows? Note that the passenger flows are concentrated near the city center. Which network is more efficient in routing (i.e., has a shorter average trip distance)?

For more details, including the methodology used (Continuum Approximation (CA), a parsimonious modeling technique that is suitable for optimizing spatial designs of transit systems), please see http://dx.doi.org/10.1016/j.trb.2015.06.012)

Presentation video

Alternative transit service schemes

Most transit systems operate in the all-stop fashion, where buses or trains visit each and every stop along the way.

We examine a transit corridor served by two alternative service schemes that are simple and rotationally symmetric:

We develop continuum models that account for: (i) service coordination at transfer stations via bus convoying; and (ii) simple route choice between express and local routes.

Key Findings:

Optimally designed alternative schemes, especially the skip-stop scheme with service coordination, often outperform the traditional all-stop systems. Converting existing all-stop systems to alternative-scheme systems can be beneficial too. The much cheaper bus rapid transit (BRT) systems operated in the skip-stop fashion can substitute for the very expensive rail systems in many cities.

Presentation slides; paper

Ongoing work: We are currently exploring optimal hierarchical networks that combine differentiated access/feeder modes (e.g., bicycle sharing) and trunk modes (e.g., rail) under various demand patterns.

Question: Given spatially uniform demand and the density of bike-sharing stations, what geometric layout of the stations will minimize users’ average walking distance? Suppose people walk along a dense grid street network.