Distributed Fleet Control Simulator

Reference

• MOVI: A Model-Free Approach to Dynamic Fleet Management
• Distributed Fleet Control with Maximum Entropy Deep Reinforcement Learning

Setup

Below you will find step-by-step instructions to set up the NYC taxi simulation using 2016-05 trips for training and 2016-06 trips for evaluation. Please make more than 10 GB memory resource available to Docker Engine.

1. Download OSM Data

wget https://download.bbbike.org/osm/bbbike/NewYork/NewYork.osm.pbf -P osrm

2. Preprocess OSM Data

cd osrm
docker run -t -v $(pwd):/data osrm/osrm-backend osrm-extract -p /opt/car.lua /data/NewYork.osm.pbf
docker run -t -v $(pwd):/data osrm/osrm-backend osrm-partition /data/NewYork.osrm
docker run -t -v $(pwd):/data osrm/osrm-backend osrm-customize /data/NewYork.osrm

3. Download Trip Data

mkdir data
wget https://s3.amazonaws.com/nyc-tlc/trip+data/yellow_tripdata_2016-05.csv -P data/trip_records
wget https://s3.amazonaws.com/nyc-tlc/trip+data/green_tripdata_2016-05.csv -P data/trip_records
wget https://s3.amazonaws.com/nyc-tlc/trip+data/yellow_tripdata_2016-06.csv -P data/trip_records
wget https://s3.amazonaws.com/nyc-tlc/trip+data/green_tripdata_2016-06.csv -P data/trip_records

4. Build Docker image

docker-compose build sim

5. Preprocess Trip Records

docker-compose run --no-deps sim python src/preprocessing/preprocess_nyc_dataset.py ./data/trip_records/ --month 2016-05
docker-compose run --no-deps sim python src/preprocessing/preprocess_nyc_dataset.py ./data/trip_records/ --month 2016-06

6. Snap origins and destinations of all trips to OSM

docker-compose run sim python src/preprocessing/snap_to_road.py ./data/trip_records/trips_2016-05.csv ./data/trip_records/mm_trips_2016-05.csv
docker-compose run sim python src/preprocessing/snap_to_road.py ./data/trip_records/trips_2016-06.csv ./data/trip_records/mm_trips_2016-06.csv

7. Create trip database for Simulation

docker-compose run --no-deps sim python src/preprocessing/create_db.py ./data/trip_records/mm_trips_2016-06.csv

8. Prepare statistical demand profile using training dataset

docker-compose run --no-deps sim python src/preprocessing/create_profile.py ./data/trip_records/mm_trips_2016-05.csv

9. Precompute trip time and trajectories by OSRM

docker-compose run sim python src/preprocessing/create_tt_map.py ./data

The tt_map needs to be recreated when you change simulation settings such as MAX_MOVE.

10. Change simulation settings

You can find simulation setting files in src/config/settings and src/dqn/settings.

Quick Start

1. Run Simulation using OSRM

This mode

docker-compose up

sim container is created and runs bin/run.sh.

2. Run Simulation using precomputed OSRM routing data

This mode uses precomputed ETA and trajectories by OSRM, which is much faster than above.

docker-compose run --no-deps sim python src/run.py --train --tag test