Q-Learning Algorithm with Double-Agent Reinforcement Learning for Smart Traffic Controller.
Jalu ReswaraNana SutisnaInfall SyafalniTrio AdionoPublished in: MWSCAS (2023)
Keyphrases
- reinforcement learning
- learning algorithm
- learning agent
- traffic signal
- multi agent
- action selection
- reinforcement learning algorithms
- optimal control
- reinforcement learning agents
- real time
- control policy
- partially observable
- single agent
- exploration strategy
- reward function
- traffic light
- actor critic
- learning capabilities
- control strategies
- state space
- multiagent systems
- network traffic
- control system
- reward shaping
- intelligent agents
- adaptive control
- function approximation
- multi agent environments
- multi agent systems
- state action
- learning agents
- state abstraction
- traffic management
- markov decision processes
- control method
- traffic control
- markov decision process
- reward signal
- learning process
- decision making
- machine learning
- active learning
- traffic congestion
- temporal difference
- agent learns
- dynamic environments
- learning problems
- fitted q iteration
- multiple agents
- supervised learning
- software agents
- learning tasks
- training data
- agent receives
- dynamic programming
- mobile agents
- autonomous agents
- control algorithm
- learning rate
- traffic flow
- model free
- rl algorithms
- control scheme
- optimal policy
- machine learning algorithms
- mobile robot
- learning mechanism