Description
This is NYC taxi trip records used as the real-world dataset for the comprehensive ride-sharing simulation available at: https://github.itap.purdue.edu/Clan-labs/Dynamic_Matching_RS. We use 2016-05 trips for training and 2016-06 trips for evaluation.
To run this simulator, the user can choose to either go through the pre-processing steps explained in README in the above Github link (also available under "Supporting Docs" here) to generate the data under the title "Data Generation", or just fetch the pre-processed files directly from here, load them into a directory, and set the DATA_DIR variable in config/settings.py.
Description of the provided files:
1. The "osrm" folder (available in the Github repository) contains the OSM data downloaded from https://download.bbbike.org/osm/bbbike/NewYork/NewYork.osm.pbf and preprocessed using the OSRM backend engine to extract the road network and geometry. (Data Generation steps #1 & #2). This is, then, used if the user decides to use the OSRM engine for real-time routing.
2. db.sqlite3 is the trip database used for simulation after pre-processing the trip records downloaded from https://s3.amazonaws.com/nyc-tlc/trip+data/.
The columns in this database include ['request_datetime', 'trip_time', 'origin_lon', 'origin_lat', 'destination_lon', 'destination_lat', 'fare']. (Data Generation steps #3 through #7). In addition, (in Data Generation step #8), prepares statistical demand profile using the training dataset and stores that in this database as well.
3. routes.pkl and tt_map.npy contains the pre-computed trajectories and trip times respectively. reachable_map.npy contains the discretized map of NewYork city according to the maps obtained from OSRM. These are pre-computed and generated using the OSRM engine to be used in real-time during simulation as the Fast Routing option provided to the user. (Data Generation steps #9).
Cite this work
Researchers should cite this work as follows:
- Haliem, M. W.; Aggarwal, V.; Bhargava, B. (2021). DRSP-Sim: A Simulator for Ride-sharing with Pooling: Joint Matching,Pricing, Route Planning, and Dispatching. Purdue University Research Repository. doi:10.4231/GDFY-5A56