A general formulation for multi-modal dynamic traffic assignment considering multi-class vehicles, public transit and parking

Implemented by Wei Ma and Xidong Pi (AlanPi1992), advised by Sean Qian, Civil and environmental engineering, Carnegie Mellon University.

Requirements

• cvxopt 1.1.9
• numpy 1.14.2
• MNMAPI: MNMAPI is a traffic simulation library developed by MAC in CMU, please refer to https://github.com/Lemma1/MAC-POSTS and http://mac-posts.com/
• MNM_mcnb: the folder interface to MNMAPI, please refer to https://github.com/Lemma1/MAC-POSTS/tree/master/side_project/network_builder

Instructions

Please clone the whole repo, and run runner.ipynb using jupyter notebook.

Experiments

To check the details of the experiments in exp_config.py, please refer to the paper.

File specifications

• src/exp_config.py: experiment settings in the paper
• src/gp.py: gradient projection method
• src/models.py: implementation of multi-modal DUE
• src/runner.ipynb: script that runs the MMDUE
• img/.: imagines used in the paper
• data/input_files_small_multiclass: experiment network in the paper

Questions

For any question, please contact Wei Ma (Lemma171@gmail.com) and Xidong Pi (xpi@andrew.cmu.edu).