SACHA: Soft Actor-Critic with Heuristic-Based Attention for Partially Observable Multi-Agent Path Finding.
Qiushi LinHang MaPublished in: CoRR (2023)
Keyphrases
- partially observable
- path finding
- reinforcement learning
- multi agent
- single agent
- partially observable markov decision processes
- policy gradient
- state space
- markov decision processes
- heuristic search
- dynamical systems
- decision problems
- reinforcement learning algorithms
- temporal difference
- function approximation
- search algorithm
- path planning
- average reward
- infinite horizon
- belief state
- policy iteration
- rule learning
- optimal control
- reward function
- multi agent systems
- hill climbing
- multiple agents
- data mining
- learning algorithm
- optimal path
- transfer learning
- supervised learning
- action space
- genetic algorithm