Traffic Light Control with Policy Gradient-Based Reinforcement Learning.
Mehmet Bilge Han TasKemal ÖzkanInci SariçiçekAhmet YaziciPublished in: SIU (2024)
Keyphrases
- reinforcement learning
- optimal policy
- policy search
- markov decision process
- model free
- action selection
- state space
- function approximators
- function approximation
- partially observable environments
- reinforcement learning problems
- policy evaluation
- markov decision processes
- reinforcement learning algorithms
- state and action spaces
- control policies
- policy gradient
- actor critic
- approximate dynamic programming
- action space
- partially observable
- partially observable markov decision processes
- policy iteration
- temporal difference
- decision problems
- control policy
- average reward
- markov decision problems
- reward function
- state action
- rl algorithms
- transition model
- finite state
- dynamic programming
- multi agent
- partially observable domains
- infinite horizon
- learning algorithm
- partially observable markov decision process
- learning problems
- optimal control
- least squares
- agent learns
- policy gradient methods
- continuous state
- reinforcement learning methods