RAPTOR Reimplementation

The main branch contains all optimizations. To run each of the optimizations, checkout to its branch!

Optimization Branches

Branch	Description
baseline	The original RAPTOR algorithm, as specified in the research paper.
opt-1	Optimizing GTFS preprocessing
opt-2	Parallelizing earliest trip selection
opt-3	Parallelizing Q-Route scanning
opt-4	Changing vector-based membership to set-based membership

Command Line Instructions

Run the following commands to build and run the full RAPTOR algorithm (check out the notes section for information on custom parameters and the unit-test suite).

make clean
make -j
./main.exe <num_iters> --dataset <dataset_name> --source <source_stop_id> --dest <dest_stop_id> --departure <departure_time>`

Notes:

• <num_iters>: Defaults to 500 iterations
• <dataset_name>: Defaults to gtfs-data, which represents the Chicago GTFS data. Alternative is the gtfs-data-newyork2 dataset, which represents the New York GTFS data
• <source_stop_id>: Defaults to random source stop, specify source stop id from the respective id in the stops.txt file
• <dest_stop_id>: Defaults to random destination stop, specify destination stop id from the respective id in the stops.txt file
• <departure_time>: Defaults to random time between 10AM-6PM, specify time based on seconds past midnight