A fast minorization-maximization algorithm for the mixture-of-normals logit model: value of time estimates under crowding conditions in the NYC subwayPP

We conduct a discrete choice experiment to quantify travel time savings under crowding conditions in the NYC subway. We show subway images at varying passenger density levels to explicitly understand the impact of crowding on subway route choice. Preliminary results indicate that our sample of New Yorkers are willing to pay \$3.2 to save an hour of travel in uncrowded conditions, which increases to \$7.2 under the technical capacity of passengers in a subway car.

This is poster number 21 in Poster Session

Prateek Bansal (Cornell University)
Ricardo Hurtubia (Pontificia Universidad Catolica de Chile)
Ricardo Daziano (Cornell University)
attribute visualization, choice modeling, expectation-maximization, image representation