Parts | Days | Selection | Search | Updates | Downloads | Help
SOE: Fachverband Physik sozio-ökonomischer Systeme
SOE 11: Physics of Collective Mobility (joint session SOE / DY / BP / jDPG, accompanying the symposium)
SOE 11.2: Talk
Tuesday, March 21, 2017, 15:00–15:15, GÖR 226
New York Taxi Rides as Point Processes — •Philip Marszal1,2, Debsankha Manik1, Andreas Sorge1, and Marc Timme1,2,3 — 1Network Dynamics, Max Planck Institute for Dynamics and Self-Organization (MPIDS), 37077 Gottingen, Germany — 2Faculty of Physics, Georg August University Gottingen, 37073 Göttingen, Germany — 3Department of Physics, University of Darmstadt, 64289 Darmstadt, Germany
Optimized ride sharing promises the flexibility of taxis with an efficiency close to that of public transportation. To provide a test environment for ride sharing algorithms, we here propose stochastic models characterizing the spatio-temporal request dynamics of taxis in New York City. On short time scales (48 minutes) and accumulating across spatial variation, we found the number of taxi pick-up events to be Poisson distributed, suggesting that the pick-up events can be modeled by a Poisson process. On long time scales (24 hours), the ride pick-up events can be modeled by an inhomogeneous Poisson process in space and time. Interestingly, the ride requests follow definitive patterns in time, yet vary strongly in space. Furthermore, a Gaussian mixture model of pick-up locations provides a possibility of inferring likely drop-off locations from initial conditions of a ride.