Developing End-to-End Control Policies for Robotic Swarms Using Deep Q-learning.
Yufei WeiXiaotong NieMotoaki HiragaKazuhiro OhkuraZlatan CarPublished in: J. Adv. Comput. Intell. Intell. Informatics (2019)
Keyphrases
- end to end
- control policies
- optimal policy
- reinforcement learning
- cooperative
- finite horizon
- control strategies
- markov decision processes
- state space
- congestion control
- particle swarm optimization
- continuous state
- control system
- admission control
- control policy
- sufficient conditions
- reward function
- infinite horizon
- finite state
- decision problems
- mobile robot
- learning algorithm
- robotic systems
- real time
- action space
- motion control
- transport layer
- application layer
- linear programming
- dynamic programming
- machine learning