Deep Q Learning with LSTM for Traffic Light Control.
Chung-Jae ChoeSeungho BaekBongyoung WoonSeung-Hyun KongPublished in: APCC (2018)
Keyphrases
- reinforcement learning
- cooperative
- state space
- function approximation
- multi agent
- recurrent neural networks
- learning algorithm
- model free
- optimal policy
- traffic light
- dynamic programming
- markov decision processes
- action selection
- multi agent reinforcement learning
- td learning
- cellular automata
- deep learning
- temporal difference learning
- potential field
- traffic simulator