An automated signalized junction controller that learns strategies by temporal difference reinforcement learning.
Simon BoxBen WatersonPublished in: Eng. Appl. Artif. Intell. (2013)
Keyphrases
- temporal difference
- reinforcement learning
- actor critic
- function approximation
- td learning
- model free
- control strategies
- temporal difference learning
- evaluation function
- reinforcement learning algorithms
- policy evaluation
- step size
- action selection
- monte carlo
- temporal difference methods
- markov decision processes
- function approximators
- optimal control
- control policy
- policy iteration
- optimal policy
- control system
- state space
- supervised learning
- state action
- learning algorithm
- neural network
- cost function
- linear combination
- learning process
- genetic algorithm
- machine learning