Two-stage reward allocation with decay for multi-agent coordinated behavior for sequential cooperative task by using deep reinforcement learning.
Yuki MiyashitaToshiharu SugawaraPublished in: Auton. Intell. Syst. (2022)
Keyphrases
- multi agent
- reinforcement learning
- cooperative
- multi agent environments
- cooperative behavior
- multi agent systems
- function approximation
- model free
- state space
- agent behavior
- agent based models
- multiagent systems
- cooperating agents
- intelligent agents
- optimal policy
- eligibility traces
- multi agent reinforcement learning
- real robot
- machine learning
- single agent
- temporal difference
- reward function
- learning process
- multiple agents
- resource allocation
- evolutionary learning
- reinforcement learning algorithms
- dynamic programming
- autonomous agents
- cooperative agents
- human behavior
- cognitive agents
- coalition formation
- partial observability
- policy evaluation
- reinforcement learning methods
- optimal control
- state action
- multiagent reinforcement learning
- average reward
- learning capabilities