A framework where a deep Q-Learning Reinforcement Learning agent tries to choose the correct traffic light phase at an intersection to maximize traffic efficiency.
Just stumbled upon your work yesterday, it provides a very nice starting point for my DQN experiments with SUMO.
A small suggestion (if you care about maintaining this project) would be to use a collections.deque data structure for your memory samples. By specifying a size_max on instantiation, you don't have to manually pop old observations from your memory DB.
Hi,
Just stumbled upon your work yesterday, it provides a very nice starting point for my DQN experiments with SUMO.
A small suggestion (if you care about maintaining this project) would be to use a
collections.deque
data structure for your memory samples. By specifying asize_max
on instantiation, you don't have to manually pop old observations from your memory DB.