This repository contains code to replicate experiments in: Real-Time Approximate Routing for Smart Transit Systems. URL: https://arxiv.org/abs/2103.06212.Usage instructions: to replicate experiments for the Manhattan network (table 2 in the paper):Download all files and unzip files manhattan_dist_1.txt, manhattan_dist_2.txt and manhattan_dist_3.txtRun line_planning.py.The experiments run by default using trip request from April 3rd 2018. To run the experiments for the fhv data from Feb 3 or March 6, uncomment the parameter 'month' line #407 in the file line_instance.py.The code allows to test

This is the software code for the rolling horizon framework suggested in the paper below: Offline Pickup and Delivery Problem with Time Windows via Rolling Horizon Trip-Vehicle Assignment, Y Kim, D Edirimanna, M Wilbur, P Pugliese, A Laszka, A Dubey, S Samaranayake, accepted in The 37th AAAI Conference on Artificial Intelligence.Data includes:RollingHorizon└── data   ├── map   │   ├── edges.csv   │   ├── nodes.csv   │   └── times.csv   ├── requests   │   └── request

Extract the datasets to the data folder and into their respective folders. apc: cleaned-wego-daily.apc.parquet weather: darksky_nashville_20220406.csv and weatherbit_weather_2010_2022.parquet gtfs: alltrips_mta_wego.parquet traffic: inrix data, can download separatelyAlso, you could refer the code repository at: https://github.com/smarttransit-ai/mta_occupancy_prediction. The related paper is: J. P. Talusan, A. Mukhopadhyay, D. Freudberg and A. Dubey, "On Designing Day Ahead and Same Day Ridership Level Prediction Models for City-Scale Transit Networks Using Noisy APC