Xiangxiangzhu/proc-tensorflow-tls

distributed learning and simulation using tensorflow for general traffic light agent

sumolights

SUMO adaptive traffic signal control - DQN, DDPG, Webster's, Max-pressure, Self-Organizing Traffic Lights

Technical details available at An Open-Source Framework for Adaptive Traffic Signal Control

Setup

Install SUMO traffic microsimulator by following instructions here (v1.2).

Using Python 3, create a virtual environment and then install dependancies with:

pip install -r requirements.txt

Comparing adaptive traffic signal controllers

First train reinforcement learning controllers:

./train_dqn.sh
./train_ddpg.sh

Then execute simulations to generate performance results for all controllers:

./gen_results.sh

Visualize results with:

python graph_results.py

Optimizing hyperparameters

Search for optimal hyperparameters for each controller:

./hp_optimization

Warning, search for reinforcement learning can require significant compute time!

Visualize hyperparameters with:

python graph_results.py -type hp