Constructive Policy: Reinforcement Learning Approach for Connected Multi-Agent Systems.
Sayyed Jaffar Ali RazaMingjie LinPublished in: CASE (2019)
Keyphrases
- multi agent systems
- reinforcement learning
- optimal policy
- policy search
- multi agent reinforcement learning
- multi agent
- markov decision process
- action selection
- policy gradient
- partially observable
- partially observable environments
- markov decision processes
- state space
- reinforcement learning algorithms
- actor critic
- control policy
- reinforcement learning problems
- state and action spaces
- action space
- cooperative multi agent systems
- multi agent environments
- policy iteration
- game theory
- markov decision problems
- single agent
- model free
- decision problems
- policy evaluation
- agent architecture
- function approximators
- reward function
- control policies
- dynamic programming
- function approximation
- approximate dynamic programming
- average reward
- agent systems
- policy gradient methods
- state dependent
- state action
- continuous state spaces
- autonomous agents
- learning algorithm
- cooperative
- agent learns
- inverse reinforcement learning
- intelligent agents
- coalition formation
- machine learning
- reinforcement learning methods
- learning agents
- partially observable markov decision processes
- control problems
- temporal difference
- transition model
- rl algorithms
- multi agent learning
- continuous state
- multiagent systems
- dynamical systems
- finite state
- software agents
- sufficient conditions
- supply chain
- average cost
- mobile robot
- agent platform
- partially observable domains
- agent technology