Improving Reward Functions in Robots Playing Capture the Flag Using Q-Learning.
Trevor PowersMichael NovitzkyChristopher KorpelaPublished in: CCWC (2021)
Keyphrases
- reward function
- reinforcement learning
- reinforcement learning algorithms
- state space
- cooperative
- optimal policy
- markov decision processes
- initially unknown
- state action
- mobile robot
- inverse reinforcement learning
- multiple agents
- state variables
- partially observable
- transition probabilities
- policy search
- learning agent
- hierarchical reinforcement learning
- agent learns
- learning algorithm
- dynamic programming
- multi agent
- function approximation
- reinforcement learning methods
- multi robot
- robotic systems
- search algorithm