//Flow Tutorials

Tutorials for getting started with deep reinforcement learning (RL) and transportation.

Flow Tutorials on Python Jupyter Notebooks

Tutorial Title Jupyter Notebook
Initalization Setup for the tutorials Read Setup
Tutorial 1 Running SUMO simulations in Flow Open Notebook
Tutorial 2 Running Aimsun simulations in Flow Open Notebook
Tutorial 3 Running RLlib experiments for mixed-autonomy traffic Open Notebook
Tutorial 4 Running rllab experiments for mixed-autonomy traffic Open Notebook
Tutorial 5 Saving and visualizing resuls from non-RL simulations and testing simulations in the presence of an RLlib/rllab agent Open Notebook
Tutorial 6 Creating custom scenarios Open Notebook
Tutorial 7 Creating custom environments Open Notebook
Tutorial 8 Creating custom controllers Open Notebook
Tutorial 9 Traffic lights Open Notebook
Tutorial 10 Running simulations with inflows of vehicles Open Notebook
Tutorial 11 Running rllab experiments on Amazon EC2 instances Open Notebook

ITSC 2018 Tutorial on Deep Reinforcement Learning and Transportation

Sunday, November 4th, 2018 | Maui, Hawaii

Session Title Slides Description
1 Welcome, opening remarks Download Why deep RL and transportation?
2 Reinforcement learning and approximate dynamic programming Foundations of RL
3 Policy optimization methods (policy gradient methods and non-policy gradient methods) Foundations of RL
4 Deep RL from a transportation lens (model-based RL and inverse RL) Download Prior research at the intersection of RL and transportation
5 Tools of the trade (SUMO, Flow, Ray RLlib)
Software and simulation tools for conducting research in deep RL and transportation
6 Hands-on tutorial on //Flow Download Hands-on exercises with //Flow for getting started with empirical deep RL and transportation
7 Advanced topics in deep reinforcement learning (multi-agent RL, representation learning) Download Compelling topics for further exploration in deep RL and transportation

Please check this page again! New tutorials will be added in the naear future.

Still Have questions?
Get in touch with us using our email or mailing list