Smart Edge-Enabled Traffic Light Control: Improving Reward-Communication Trade-offs with Federated Reinforcement Learning.
Nathaniel HudsonPratham OzaHana KhamfroushThidapat ChantemPublished in: SMARTCOMP (2022)
Keyphrases
- reinforcement learning
- trade off
- state space
- edge information
- smart objects
- function approximation
- edge detection
- reinforcement learning algorithms
- eligibility traces
- learning algorithm
- dynamic programming
- communication systems
- multi agent
- policy gradient
- reinforcement learning methods
- average reward
- control policy
- partially observable environments
- partially observable
- reward function
- action selection
- model free
- smart grid
- optimal control
- communication networks
- markov decision processes
- total reward
- autonomous agents