Multi-agent, reward shaping for RoboCup KeepAway.
Sam DevlinMarek GrzesDaniel KudenkoPublished in: AAMAS (2011)
Keyphrases
- multi agent
- reward shaping
- robocup soccer
- reinforcement learning
- temporal difference
- reinforcement learning algorithms
- complex domains
- state space
- multi agent systems
- intelligent agents
- multiagent systems
- software agents
- learning algorithm
- function approximation
- multiple agents
- autonomous agents
- single agent
- markov decision problems
- evaluation function
- optimal control
- learning process
- machine learning
- markov decision processes
- transfer learning
- steady state
- markov chain
- dynamic programming
- learning agent