CaRL: Cascade Reinforcement Learning with State Space Splitting for O-RAN based Traffic Steering.
Chuanneng SunYu ZhouGueyoung JungTuyen Xuan TranDario PompiliPublished in: CoRR (2023)
Keyphrases
- state space
- reinforcement learning
- reinforcement learning algorithms
- optimal policy
- markov decision processes
- heuristic search
- markov chain
- traffic flow
- network traffic
- state abstraction
- partially observable
- action space
- dynamic programming
- state variables
- function approximation
- initial state
- belief state
- markov decision process
- traffic control
- real time
- planning problems
- internet traffic
- state and action spaces
- learning algorithm
- multi agent
- model free
- dynamical systems
- reward function
- reinforcement learning methods
- continuous time markov process
- stochastic domains
- continuous state spaces
- traffic management
- traffic data
- state transition
- control problems
- temporal difference
- control policy
- road network
- particle filter
- continuous state
- learning agent
- learning process
- reward shaping