2D-learning: A collaborative distributed strategy for multi-agent reinforcement learning through consensus + innovations

Computer Science

• Communication Network 100%
• Multi-Agent Reinforcement Learning 100%
• Function Value 50%
• And-States 50%
• Mutual Information 50%
• Markov Decision Process 50%
• multi agent 50%
• State Transition 50%
• Distributed Version 50%
• Control Action 50%

Keyphrases

• Consensus + Innovations 100%
• Multi-Agent Markov Decision Processes 50%
• Stochastic Communication 50%
• Infinite Horizon Discounted Cost 50%
• Global State 50%
• Remote Controller 50%